-
https://clinton.presidentiallibraries.us/files/original/44b87f3ed59ed9d883686b17c25486ff.pdf
1d7731668c427165ad31016706d9286e
PDF Text
Text
FOIA Number: 2006-0810-F
FOIA
MARKER
This is not a textual record. This is used as an
administrative marker by the William J. Clinton
Presidential Library Staff.
Collection/Record Group:
Clinton Presidential Records
Subgroup/Office of Origin:
Health Care Task Force
Series/Staff Member:
Subseries:
OA/ID Number:
1229
FolderlD:
Folder Title:
Briefing Book on AIDS [2]
Stack:
Row:
Section:
Shelf:
Position:
S
51
6
5
1
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
M
�Tab 4.
This S e c t i o n i n c l u d e s :
Morbidity and M o r t a l i t y Weekly Report,
A p r i l 30, 1993/Vol 42/No. 16
�S N B •C C ODD^HIV - O S E T Y •D
C
X R X 7033EO
: 4- 6-93 :11:40AM
CDC
202 865 6800:? 2/ 5
April 30,1993/Vol. 4 2 / N o . 16
301
301
3M
Clean Air Week — May 1993
Populetiont at Risk from Air Pollution
Nutrition end Mortality Assessment —
—^wuhern Siirim, Slecrh I ' m
308 Impact of the Expended AIDS S u r v e i U a n c e ^ A ^ /
Case Definition on AIDS Ceee Reporting
Updala: Drecunculiaaitf tnfficilton —
Nigerie, 1992
312 Linking Multiple Data Sources in Fatal Alcohol
Syndrome Surveillance — Alaska
315 Notice to Readers
323 Quartarly AIDS Mao
-
MORBIDITY AND MORTALITY WEEKLY REPORT
Clean Air Week — May 1993
The American Lung Association (ALA) sponsors Clean Air Week in May each
year to educate the public about the importance of clean air to lung health. The
theme for this year's campaign is "Clean Air Is Up to Youl", emphasizing the role
for each person in achieving clean air. Through local Clean Air Week activities during 1993, the ALA will educate the public about practical methods to promote clean
air through actions such as decreasing driving and other forms of energy conservation, improving indoor air quality at home and the workplace, and supporting clean
air regulations and enforcement in the community.
Local lung associations designate the week for their respective communities to
observe as Clean Air Week and will disseminate messages through environmental
and health fairs, school presentations, other community events, and the media.
Nearly 50 locations in the United States also will participate in the Clean Air Challenge (a pledge-based bicycling event), to raise funds for local clean air programs
and other efforts to prevent lung disease. In addition, local lung associations will
observe a Clean Commute Day and request that motorists use alternative forms of
transportation (e.g., carpools, mass transit, or bicycles).
Additional information about Clean Air Week and related activities is available
from local offices of the ALA, which are listed in the white pages of the telephone
directory. The address of the national office of the ALA is 1740 Broadway, New York,
NY 10019-4374; telephone (212) 315-8700.
Health
Objectives
for the
Nation
Populations at Risk from Air Pollution — United S t a t e s , 1991
•
The Clean Air Act of 1970 required the U.S. Environmental Protection Agency (EPA)
to establish National Ambient Air Quality Standards (NAAQSs) for six of the most
(widespread air pollutants in the outdoor environment: particulate matter with a diameter <10 (im, sulfur dioxide, nitrogen dioxide, carbon monoxide, ozone, and lead.
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES / Public Health Service
�SENT BY:CDC ODD/HIV - O S C
3
jj
|
: 4- 6-93 :11:41AM :
XEROX 7033-
202 865 6800:* 3/
World Health Organization, 1983.
3. COC Famine-affected, refugee, and displaced populations: recommendationj.for public health
issu^fcxMMWR 1992;41(no. RR-13).
4. de Ville dtesGoyet C, Seaman J. Geijer U. The management of nutrjlitffrel emergencies in large
populations^SeQeva: World Health Organization, 1978.
5. Serdula MK, Aphtffft^M, Kunene PF, et al. Acute and^pfrfonic undernutrition in Swaziland.
J Trop Pediatr 1987;33!35»42.
6. Carlson BA, Wardlaw TM. A^gJ^bal. regional a«<J^country assessment of child malnutrition.
New York: United Nations Childreh^sFuniJ^rMO; UNICEF staff working paper no. 7.
7. Person-Karell B. The relationship bepi^ftQchild malnutrition and crude mortality among 42
refugee populations [Thesis]. Ajjartfa: Emoi^^Jniversity, 1989.
8. United Nations Children'sFiKftfTrhe state of the^scjd's children, 1991. New York: United Nations Children's Fund, id^2.
9. CDC. Population-^aSed mortality assessment—Baidoa and^Af^pi, Somalia, 1992. MMWR
1992:41:913-7.
Current
Trends
Impact of the E x p a n d e d A I D S Surveillance C a s e Definition
on A I D S C a s e Reporting — United S t a t e s , First Quarter, 1993
On January 1, 1993, the acquired immunodeficiency syndrome (AIDS) surveillance
case definition for adolescents and adults was expanded beyond the surveillance
definition published in 1987 [ 1 ) to include all human immunodeficiency virus
(HlV)-infected persons with severe immunosuppression (<200 CD4+ T-lymphocytes/nL
or a CD4+ T-lymphocyte percentage of total lymphocytes of <14), pulmonary tuberculosis (TB), recurrent pneumonia, or invasive cervical cancer. This report reviews AIDS
surveillance reports CDC received from local, state, and territorial health departments
during the first quarter of 1993 and summarizes the impact of the changes in the AIDS
surveillance case definition.
From January 1 through March 31, a total of 35,779 AIDS cases—13% of the cumulative total of 284,840 AIDS cases reported since 1981—were reported to CDC,
representing a 204% increase over the number reported for the same period in 1992
(11,770 cases) (Figure 1). Forty-eight states and Puerto Rico reported cases based on
the criteria added to the surveillance definition in 1993.
Of the 35,779 AIDS cases, 21,582 (60%) were reported based on the conditions
added in 1993, and 14,197 (40%) were reported based on pre-1993-defined conditions—a 2 1 % increase in reporting of pre-1993-defined cases over the number
reported for the same period in 1992. Of the cases reported with only 1993-added con-
�SENT BY:CDC ODD/HIV - OCS -
4- 6-93 :11:41AM
XEROX 7033-
309
MMWR
Vol. 42 / No. 16
202 865 6800:s 4/ 5
AIDS Surveillance Case Definition — Continued
ions, 19,284 persons (89%) had severe HIV-related immunosuppression only; 2021
had pulmonary TB; 245 (1%), recurrent pneumonia; and 41 (<!%), invasive cervical cancer.*
Implementation of the expanded surveillance definition has been associated with
an increase in the median interval between date of diagnosis and date of report of
AIDS cases; in preceding years, this interval had been 3 months. In contrast, persons
with conditions added in 1993 had a median interval between date of diagnosis and
date of report of 9 months, and persons with pre-1993 conditions reported in the first
quarter of 1993, 5 months.
Reported by: Local, state, and territorial health depts. Div of HIV/AIDS, National Center for
Infectious Diseases. CDC.
Editorial Note: The findings in this report indicate that an immediate impact of the
revised AIDS surveillance case definition on case reponing has been a substantial increase in the number of reported AIDS cases. The increase in the first quarter of 1993
is not expected to be sustained because the increase in cases reported during this
period reflects predominantly the reporting of the accumulated number of persons
•The total number of diagnosed 1993-added conditions (i.e., pulmonary TB, recurrent pneumonia, and invasive cervical cancer) is greater than the total number of persons with these
conditions because nine persons were reported with more than one condition.
FIGURE 1. Reported AIDS cases among adolescents and adults following expansion
If the AIDS surveillance case definition* — United States, first quarter, 1992 and
93
40,000
35,000-
Pre-1993 Definition
30,000 -
1993 Definition
(Severe Immunosuppression)
25,000
•
1993 Definition
.
(Added Opportunistic Illnesses )
7
g 20,000 •
O
15,000-
10,0005,000 0
First Quarter
1992
First Quarter
1993
All HIV-infected persons with severe immunosuppression (<200 CD4+ T-lymphocytes/jiL or a
CD4+ T-lymphocyte percentage of total lymphocytes of <14), pulmonary tuberculosis (TB),
recurrent pneumonia, or invasive cervical cancer in addition to the clinical conditions included
in the AIDS surveillance case definition published in 1987.
Pulmonary TB. recurrent pneumonia, and invasive cervical cancer.
�:
SENT BYCDC ODD/HIV - O S C
310
: 4- 6-93 :11:42AM :
XEROX 7033-
MMWR
202 865 6800:# 5/ 5
April 30.1993
\
AIDS Surveillance Case Definition — Continued
I
ith previously diagnosed conditions added to the surveillance definition in 1993 w h o
ould not be reported as AIDS cases until January 1, 1993. The interval between date
of diagnosis and date of report for persons with 1993-added conditions indicates that
AIDS cases in these persons were diagnosed earlier than in other persons reported
with AIDS.
CDC has estimated that the number of AIDS cases reported during 1993 will increase approximately 75% as a result of the expanded reporting criteria ( 2 ) . The
increase in the number of persons reported with prB-1993 conditions during this
quarter probably reflects changes in surveillance procedures associated with implementation of the 1993 surveillance definition. CDC will continue to report on the
ongoing evaluation of surveillance findings and the impact of the expanded AIDS surveillance case definition.
I
References
1. CDC. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR 1992;41(no. RR-17).
2. CDC. Projections of the number of persons diagnosed with AIDS and the number of immunosuppressed HIV-infected persons—United States, 1992-1994. MMWR 1992;41(no. RR-18).
international
Notes
Update: Dracunculiasis Eradication — Nigeria, 1992
The reported incidence of dracunculiasis (i.e., Guinea w o r m diseaa^nn Nigeria declined substanHqlly during 1992—the fourth consecutive y e a r ^ n which reports of
known cases declmad. This report summarizes dracunculiasis surveillance data for
Nigeria and d e s c r i b e s ^ o g r e s s toward eradication of this disease.
During the i g g i - i g g z ^ i d e m i o l o g i c year (i.e., July 1 ^ 1 - J u n e 1992), 201,453 cases
of dracunculiasis were repoHed in 4576 villages w h e r t t h e disease is endemic, a 25%
reduction from the number of c ^ e s reported duripo 1990-1991 (Figure 1). Since 19881989, the number of reported c a s ^ h a s d e c r e e d 68.5%, and the number of villages
where the disease is endemic has d ^ e a s ^ d 22.9% (from 640,008 cases in 5932 villages) (7).
Nigeria's Guinea Worm E r a d i c a t i o n A o g t e m (NIGER) intensified all major control
measures during 1992, including extending heaHh education and community mobilization to all villages with endepprfc dracuncuiiasisNpromoting education about the
disease in schools in areas witn endemic dracunculiafci^, distributing cloth filters to
more than 70% of the villao^s, and targeting at least 30%>af the affected villages for
provision of safe drinkino/Water. In addition, during 1992, Nigeha completed its transition f r o m annual retrprspective surveys to monthly reporting of cases by trained
village-based healtlvWorkers in each of the villages with endemic drabmiculiasis. During 1993, NIGER will introduce use of temephos (Abate®*) to treat u n S f ^ s o u r c e s of
drinking water in selected villages in which the disease is endemic.
•Use of trade names and commercial sources is for identification only and does not imoly
endorsement by the Public Health Service or the U.S. Department of Health and Human
Services.
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
�Tab 5.
This S e c t i o n Includes:
Journal of AIDS. June 1992/Vol. 5/No.
6
�O 5.
O 6.
9 3
1 O : O T
A M
* F D A /
C D E R ^ M L*
Changes In AIDS Incidence Trends In the U.S.
118
13
P
1 6
Tt138 • Pagel
Changes in AIDS Incidence Trends in the U.S.
This article appeared at pages 547-555 of the June 1992 issue (Vol. 5, No. 6) of the Journal of Acquired
Immune Deficiency Syndromes. // ir reprinted here with their permission. The article was written by Timothy
Green, John Karon, and Okey Nwanyanwu of the Division of HIV/AIDS at the Centers for Disease Control. For
further information, contact Dr. Green, Centers for Disease Control, 1600 Clifton Road, N.E. (M/S-E48), Atlanta,
GA 30333.
Changes in AIDS Incidence Trends in the United States
Timothy A. Green, John M. Karon, and Okey C. Nwanyanwu
Division of HIVIAIDS. National Center for Infectious Diseases, Ctniers for Disease Control. Public Health Service.
United States Department of Health and Human Services. Atlanta. Georgia, U.S.A.
Summary: Esiimaung the current prevalence of human immunodeficiency virus (HIV) and projecting the future incidence of AIDS require that trends in
incidence be analyzed and interpreted. We analyzed AIDS incidence trendi in
the United States by exposure category and selected demographic factors. In
1987, the trend in United Sutes AIDS incidence changed as growth in the
number of cases .diagnosed per quarter began to decline. The slowing in growth
is due in large JHUUO a plateau in quarterly inckkoce in men who have sex with
men in the New York City, San Francisco, and Los Angeles metropolitan
statistical arew (MSAs), andin injecting drug users in the New York City MSA
and New Jersey. Incidence has also reached a plateau in both adult/adolescent
and pediatric blood and blood product recipients. Quarterly U.S. AIDS incidence wasroughlyconstant during 1990, but appears to have increased to a
higher level during the first half of 1991. The variation in incidence trends
among subgroups suggests that several facton have affected the trend in total
incidence and that the burden of severe symptomatic HIV disease may be
shifting. Key Words: Epidemiology—United Sutes.
The incidence of AIDS in the United States continues to increase, with 33,230 cases reponed to the
Centers for Disease Control (CDC) in 1989 and
43,339 cases reported in 1990 (1). Estimating the
number of future AIDS diagnoses and the current
prevalence of human immunodeficiency virus
(HIV) infection is necessary for predicting the future of HIV disease and for planning health care,
social, and educational services. The most widely
used methods for making these estimates, extrapolation and back-calculation, require that historical
trends in AIDS case incidence be analyzed and in-
terpreted. Such analyses must take into account, at
a minimum, differences in trends among the various
modes of exposure lo HIV, and perhaps geographic
and racial/ethnic differences as well.
In this article we show that although quarterly
AIDS incidence (the number of cases diagnosed per
quarter) in the United States has not declined, the
quarterly growth in incidence (the increase from
one quarter to the next in the number of cases diagnosed) reached a maximum in early 1987 and has
declined since then. Analysis by exposure, geographic, and racial/ethnic categories shows the
slowing in growth to be a composite of trends rangAddrtu coimpondcnce and reprint requcsti to Dr. Timothy ing from a plateau or leveling in incidence in some
A. Crecn tt Centers for Otsease Control. 1600 Clifton Road. groups to no slowing in the growth in diagnosed
N.E (M/S-E4*). AUanu. GA JOJJJ. U.S.A.
cases in other groups.
Uie of inde n m s ii for identiTication only and do«t not
a e
jnpty endoncnent by tht Public Health Service or tht U.S.
Dp i t t or Health and H m n Strvtcei.
t Mmn
u a
METHODS
"oA of t e US Government. Not tubject to copyright in the
b
Ututtd Statu.
COC malniains national surveillance of AIDS through the re^inuicripireceivedAugust S. 1991; accepted December 19. ceipt of AIDS case repons submitted by state, territorial, and
local health depanmenu. All 5 sutes, the District of Cotumbia,
0
Copyright © 1992 by A ntlc Information Services, Inc.
1
September 1992
�O 5.
O e .
9 3
I O I O V
A M
*
F D A /
C D E R
United Sutet dcpendtnciet tnd pottettioni, and independent
Mtieni in free asaociation with the United StatctreportAIDS
caact to CDC unoj a uni/onn cue definition and cut repon
form (I). For surveillance purpoaet, each AIDS case ii counted
only once in a hierarchy of exposure cattforiei. Pereoai with
more than one reponed mode of exposure to HIV are dasiified
in the caiegory listedfirstin the hienrchy. except that men with
a history of both sexual contact with other men and iiuectinf
drug use make up a sepante category. Cases in persons with no
reported hiitory of exposure to HIV through any ef the routes
listed in tbe hiervchy are aitigned to a no-ideatified-risk'factor
(NUt) category and are Investigated by itatf and local health
departments. Persons who have an exposure mode identified as
aresultof thii bivestigaikm arc reclassified into tha appropriau
exposure category (2).
In this repon, we analyze the numbers of AIDS cases diagnoted during each calendar quarter, after s4juiimem for cstlmaicd delays inreporting.Because in our experience adjusted
incidence estimatei for the mott recent on* to two quarters arc
notreliable,we analyzad catet diagnotcdfromJanuary, 1964
through June. 1991. andreportedto CDC through December.
1991. Reporting delayt were cttlmaied uting a maximum likelihood statistical procedure (3), taking into account possible differences inreponingdelays among exposure, gaogrephic, racial/
ethnic, MSA-population-sizc, and age-at-diagnosis categorici
but utuming thatreportingdelayt within these group* have not
changed over time.
The adjusted quarterly meideitcc for each group analyred ineludat a percentage of the NIR cases thai bekmg to the same
geographic: and racial/ethnic categorici included in tha group in
quettion. Tbe redittributlon of NIR cases is baaed on the i*xand race-specific exposure category distributions of cases diagnosed between 1984 and 1988 that were initially assigned lo the
NIR category but have subsequently bean rtclasiified (Table I).
Each adiutted incidence scriet wu tmootbed using a weighted
moving avenge of the abutted incidence values. For each time
i, the weight applied lo the abutted incidence at time J is proportional io the value at r of a Gaussian probability density functkm with mean I and standard deviation 0.37 yean. Tbe proportionality constant it chosen so that, for each I. the weights sum
to one; the resulting weights are essentially xaro outside the
interval i l I year. Computationt were carried out using ihe
kernel-type scattcrplot smoother in S-PLUS (Statistical Sciences. Seattle WA).
TABLE 1. Expoiurt-caitfory diiiributloH by ttx and tact of
AIDS caiti diatnottd bttwtin 1964 and 19U. MitaUy
asslgntd io ike NIR category but mbitqiuHtty reclatiifled
Adult/adolucnt
Mm wbo ham tax with mta
litiectiai dnif us*
Mia who havi sex with am
aad iaieet dn«s
Hcnsphilia/ceafulalioa diiordir
Hcumaxual coatact
Bon ia Pattan-II country
Racaipt af Mood trauhiiioa.
Uadtteimiaad*
Pwllathc
Mother with/at risk for
HIV iafectloa
Other
W w
M
Othar
Whiu Other
0.7J
0.07
0J}
0J3
0.11
0.21
0.03
0.00
0.06
0.00
0.03
0.00
0.M
0.01
0.00
0.JT
0.00
0.00
0.66
0.00
0.04
0.01
0.02
0.10
an
0.17
0.06
0.06
O.C
0.11
0.11
0.12
0.10
0.10
1.00
0.00
" Caiesremataiaiia the NIR category aftereaaiph!U isnlUflliM
September 1992
P
L
1 T
Caseload Trends and Economic Pro]ectlons
11138 • Page 2
Eipmun ealccery
M
In thefigures,the plotted points indicate adjusted quarterly
incidence. Tht supehmposcd curves are obtained by connecting
ihe smoothed values with straight linet; the smoothed values
themselves are not plotted. However, since smoothed curves
cannotreflectnpid changes in incidence trends, both the adjusted incidence values and the smoothed curves must be used in
evaluating trends.
RESULTS
The growth in the number of AIDS cases diagnosed per quarter in the United Sutes increased
until early 1987, and then began to decline (Fig. 1).
The growth in incidence stabilized in 1988 and early
1989 before beginning to decline again in mid-1989.
Quarterly incidence was roughly constant during
1990, but appears to have increased to a higher level
during the first half of 1991. The initial slowing was
more pronounced and occurred earlier in men who
have sex with men but do not inject drugs (MSWM)
than in all other exposure categories combined. The
more recent slowing appears to have occurred at
roughly the same time in MSWM as in all other
exposure categories combined.
The trend in incidence in MSWM is a composite
of a number of trends that reflect geographic and
racial/ethnic differences. Incidence in the three
MSAs with the most cases—New York City, San
Francisco, and Los Angeles—shows a plateau beginning in mid-1987 in non-Hispanic whites (Fig. 2)
and a near-plateau beginning in mid-1988 in all other
racial/ethnic categories combined (Fig. 3). Outside
these three MSAs, the slowing in growth noted in
the United States as a whole is evident only in nonHispanic whites and is more apparent in MSAs with
population over 1,000,000 than in rural areas and
MSAs with population under 1,000,000 (Figs. 2 and
3). Quarterly incidence in non-Hispanic white
MSWM was roughly constant during 1990 in each
MSA category but. outside New York City, San
Francisco, and Los Angeles, appears to have increased to a higher level during the first half of 1991.
The growth in AIDS incidence in female and heterosexual male injecting drug users (IDUs) began to
decline in mid-1987 (Fig. 4). This pattern is a composite of a slowing in growth, and ultimately a plateau in incidence, in this exposure category in the
New York City MSA and New Jersey, and little
slowing in growth in the rest of the United States
(Fig. S). The growth in diagnosed cases in men with
a history of both sexual contact with other men and
injecting drug use has been declining since late
1986, and incidence appears to have leveled in early
1989 (Fig. 4).
AIDS Reference Guide
�O 5.
O 6.
1 O : O
AM
* FDA. /
C D E R /
p i e
M L
11138 « Pages
Changes In AIDS Incidence Trends In the U.S.
14000 .
12000 .
•
>
8IOOOO .
o
A - All exposure categories
M - Men who have aex with men but do not Inject drugs
O - All other exposure categories
A
*
I 19S4 I 1985 I 1986 I 1M7 I 19M I I S M I 1990 I 1991
Year and Quarter ol tNagnoelt
HQ. 1. AIDS Incidanca, adjusted for reporting delay*, tn tho United State*. January. 1984-June. 1991, baaed on cases reported
through Dwoamtxr, 1991.
The growth in incidence in persons reporting heterosexual contact with a person with, or at increased risk for. HIV infection shows no indication
of slowing (Fig. 6). In each of the last five years the
percent increase in annual AIDS incidence in this
group has been higher than that of any other expo-
sure category (Table 2). Diagnosed cases in this
group increased 40% from 1989 to 1990 while corresponding increases were no more than 12% in
each of the other exposure categories.
The remaining exposure categories contain relatively few cases. Incidence in persons presumed to
2500 .
2000 .
C - New York City, San Franeiaeo, Los.Angelm
L - Other MSAs with population > 1,000,000
S - Reat of United States
.
l
i-
H 1500 .
I
i
i
1000 .
500 .
0 .
I 19S4 I 1985 I 1986 I 1987 I 1988 I 1999 I 1990 I 1991 I
Year and Quarter of Diagnosis
FW. 2. AIDS incidence, adjusted lor reporting delays. In non-Hltpsnlc white men who hava tax with men but do not ln|ac
drug*. January, lOfrt-jurva, ia9l, baaed on caaea reported through December. 1991.
Copyright © 1992 by Atlantic Information Services, Inc.
September 1992
�O 5 .
O 6 .
9
1 O :O
AM
P 1 9
* F D A / C D E R, ML
Caseload Trends and Economic Projections
11138 • Page 4
1200
j
1000 .
n
u
o
L
1
600
|
C - New York City, San Francisco, Los Angeles
L • Other MSAs with population > 1,000,000
S - Rest of United States
-
600
3
Z
£ 400
c
0 .
I 1964 I 1985 I 1986 I 1987 I 1988 I 1989 I 1990 I 1991 1
•
Year and Quarter of Diagnosis
no. 3. AIDS incidence, adjusted for reporting delays. In men (excluding non-Hispanic whites) who have aex with men but do
not inject drugs. January, ise4-0una. 1991, basad on cases reported through December, 1991
have acquired HIV infection through heterosexual
contact because they were bom in a country where
heterosexual transmission has predominated for
several years—countries termed "Pattent 11" by
the World Health Organization (4>—has exhibited
little change since 1987, although incidence in each
of thefirsttwo quarters of 1991 is higher than in any
previous quarter (Fig. 6). Incidence in blood and
blood product recipients shows a leveling in both
adults and adolescents and in children under 13
years of age at diagnosis (Figs. 7 and 8). Incidence
in children bom to mothers with, or at increased
3500
3000 .
H - Women and heterosexual
M - Men who have sex with men
1964 I 1985 I 1986 > 1987 I 1986 I 1969 I 1990 I 1991 I
Year and Quarter of Oisgnosis
FW. 4. AIDS Incidence, adjusted for reporting delays. In Injecting drug users, January, 1964-Jun*, 1991, baaed on cases
reported through December. 1991.
O
September 1992
200 .
AIDS Reference Guide
�Q ^
0
6
.
g3
1 O : OT
AM
^
P 2 O
F D A / ' C D E R / M L
11138 • Pages
Changes In AIDS Incidence Trends In the U.S.
2000 J
N - New York City, New Jersey
O - Rest of United Sutes
Jj 1500
m
u
o
I 10
00
i
I
500 .
0 .
19M I 19S5 I 1986 I 1987 I 1968 I 1989 I 1990 I 1991 I
Year end Quarter of Oisgnosis
FIO. S. AIDS Incidence, adjusted tor reporting delays. In female and heterosexual male Injecting drug users. January. 1
June, 1991. baaed on cases reported through December. 1991.
risk for, HIV infection shows no indication of a
DISCUSSION
slowing in growth before 1990 (Fig. 8), but recent
fluctuations in quarterly incidence make it difficult AIDS jurveillance data indicate trends in the
to describe current incidence trends in this exponumSToTpersoris developing life-threatening dissure category.
ease duetoHIV infection. Identifying these trends
1000 .
H - Reported heterosexual contact
P • Bom in Pattem-ll country
*
900 J
u
°
z
600
400
f
£
200
J
-
r
t
'
i
*
.
'
'
*
*
!
'
*
^
0 .
I 1964 I 1965 I 1986 I 1987 I 1988 I 1989 I 1990 I 1991 '
Year and Quarter of Oisgnosis
lnCl
d
u s t w l
, o r
IH?" • • ~ M ? 5 . £ . ? f : • i
reporting delays, in persons reporting heterosexual contact with persona with, or at
mcreaaed risk for. HIV Infection, snd in persons bom in Panern-Jl countries. January, 1964-Oune. 1991, based on cases reported
through December, 1991.
Copyright © 1992 by Atlantic Information Services, Inc.
September 1992
�0 5.
o e.
1 O
9
:
*
O
P
F D A / C D E R / M L
2
1
Caseload Trends and Economic Projections
11138 • Page6
TABLE 2. Perctm change in annual AIDS incidence, adjusted for reporting delayt. compared with the previous
by exposure category
IW6
Expoturc caiefory
Adull/adolescenl
Men who have «cx with men
Injcctinf dnif ute
Men who have sex with men and inject drugs
Hemophilia/coagulation disorder
Heterosexual contact
Born in Paltem-U country
Receipt of blood mntfuaion,
blood components, or tissue
Pediatric
Mother with/at nsk for HIV infection
Other
1987
1988
1989
1990
60
69
67
49
102
40
46
69
M
59
98
34
18
39
IS
28
56
8
15
24
9
-10
43
IB
6
12
1
0
40
-4
79
63
8
-2
-6
62
-12
49
54
29
-3
23
-14
is important since they suggest characteristics of
persons who will require extensive medical care as
well as groups in which the number infected may be
increasing relativelyrapidly.However, since very
few AIDS cases develop within the first two years
rafter infection (5,6), trends in AIDS incidence do
i not describe recent trends in the incidence of HIV
^ V - ^infection itself.
^-The growth in AIDS incidence has declined over
the last 4-5 years. In some exposure, geographic,
and racial/ethnic categories, quarterly incidence
was roughly constant during 1990. In many of these
same categories, however, quarterly incidence appears to have increased to a higher level during the
first half of 1991. This phenomenon of roughly con250
year,
11
0
stant AIDS incidence for several quarters followed
by increases in incidence has been observed in a
number of previous instances [e.g., all cases in
1989; (Fig. 1); cases in non-Hispanic white MSWM
in large MSAs outside New York City, San Francisco, and Los Angeles in 1988 (Fig. 2); and cases
attributed to heterosexual transmission in 1988 and
1989 (Fig. 6)]. Trend data must be interpreted with
caution, therefore, and AIDS incidence trends must
continue to be monitored closely.
There are several possible explanations for the
observed changes in AIDS incidence trends. These
explanations include past declines in new HIV infections due to changes in behavior or saturation in
high-risk populations, a slowing in the progression
.
T - Transfusion
C - Cloning Factor
200 .
o
i
150 .
E
100
.
50
0
.
1984 I IMS I 1986 I 1987 I 1988 I 1989 I 1990 I 1991 I
Year and Quarter of Diagnosis
FW. 7. AIDS Incidence, adjusted for reporting delays, in adult and adolescent blood and blood product recipients January
ige4~June. 1991, based on eases reported through December, 1991.
September 1992
AIDS Reference Guide
�O 5.
0 6.
9
1 0 : 0 T
A M .
>- F D A / C D E R / M L
t
P 2 2
11138 • Page?
Changes In AIDS Incidence Trends In the U.S.
200 J
P - Perinatal transmission
B - Blood product recipients
i
i
S
o
150 .
100 .
5 0
0 .
1984 I 19*5 I 1986 I 1987 I 1988 I 1989 I 1990 I 1991 I
Year and Quarter of Diagnosis
m . 8. AIDS Incidence, adjusted tor reporting delays, In children under 13 years ol sge at diagnosis, January, 1964-Oune, 1991,
baaed on cases reported through December, 1991.
lo AIDS in HIV-infected persons due to antiretrovirel and otherwise improved medical therapy and
patient care, and changes in the timeliness and completeness of AIDS case reporting (7). Slowing in the
growth in incidence in particular geographic areas
may also reflect, in part, migration. Current incidence trends differ markedly among exposure, geographic, and racial/ethnic categories, suggesting
that the relative contributions of these and other
causes vary among subgroups.
The slowing of the growth in AIDS incidence in
MSWM reflects a plateau in incidence in the New
York City, San Francisco, and Los Angeies MSAs.
as well as a marked slowing of the growth in incidence in a number of other large MSAs (8). Quarterly incidence in MSWM in the United States was
roughly constant during 1990, and recent projections using back-calculation suggest that AIDS incidence in MSWM may remain relatively level during 1991-1995 (9).
Early slowing in the growth of AIDS incidence in
MSWM may be related to a reduction in risky sexual practices and a consequent decline in new HIV
infections during the early 1980s, the use of antiretroviral drugs and other prophylactic therapies such
as aerosolized pentamidine beginning in mid-1987
for the treatment of HIV-infected patients, and migration of HIV-infected but AIDS-free MSWM out
of large metropolitan areas. The later slowing out-
side large MSAs and the lack of slowing among
Mack and Hispanic MSWM in many large MSAs
may be due to less effective risk-reduction measures, particularly early in the epidemic; to poorer
access to, and use of, medical care and therapy for
HIV-infected patients; and to later introduction of
HIV into these populations. These issues have been
discussed in detail elsewhere (8).
In the New York City MSA and New Jersey,
where nearly 40% of recently diagnosed IDU cases
reside, AIDS incidence in IDUs has reached a plateau. Seroprevalence rates of 50-60% were observed in drug treatment centers in New York City
and pans of New Jersey in the mid-1980s (10).
These rates have subsequently stabilized (11). indicating a lower incidence of new infections since that
time. This combination of high seroprevalence early
in the epidemic and lower HIV incidence for the
past several years could account for the observed
plateau in AIDS incidence now.
The sustained increase in incidence in IDUs outside New York City and New Jersey may be due to
later introduction of HIV into this population, coupled with less effective prevention measures and
poorer access to, or use of, medical care and therapy for IDUs in general, compared with other populations at risk for HIV infection. In a study of
Maryland residents with AIDS, IDUs were less
likely than patients in other exposure categories to
Copyright © 1992 by Atlantic Information Services, Inc.
September 1992
�O 5.
o e.
3
1 o
A M
*
F D A
C D E R
L
P O 2
11138 • Page 9
Changes fn AIDS Incidence Trends in the U.S.
}. Breobncyer R. Gocdcn JJ. Ceniorini in an epidemic with
aa application to hemophilia-asiociated AIDS. Biomtirici
19»;45:32J-33.
6. Bacchetti P. Ettimatint the incubation period of AIDS by
comparinf population infection and diagnosis patterns. J Am
Slat Assoc 1990:83:1002-8. .
7. Cenun for Disease Control.' HIV prevalence estimates and
AIDS case projections lor the United States: Report based
upon a workshop. MMWR l990;39(No. RR.|6):1-3I.
t. Karon JM, Berkelman R L . The geographic and ethnic diversity of AIDS incidence trends in homosexual/bisexual men
in tbe United Sutes. J Acquir Immuiu Dtfic Syndr 1991:4:
1179-89
9. Braokmcycr R. Reconstruction and future trends of the
AIDS epidemic in the United States. Scitnct 1991:233:3742.
10. Cenun for Diseaac Control. Human immunodeficiency virus infection in the United States: A review of current
knowledge. MMWR l987J6(No. S-6):l-U.
11. Stonebumer R L . Chaisson MA. Weisfuse IB. Thomas PA.
Editorial review: The epidemic of AIDS and HIV-1 infection
among heterosexuals in New York City. AIDS 1990:4:99106.
12. Moore RD. Hidalgo J. Sugland BW. Chaisson R E . Zidovudine and the natural history of the acquired immunodeficiency syndrome. S Engl J Med 1991^24:1412-6.
13. Centers for Disease Control. Update: Heterosexual transmission of acquired immunodeficiency syndrome and human
immunodeficiency virus infection—United Sutes. MMWR
1989:3«:42J-<. 429-34.
M
14. Holmes K K . Karon JM. Kreiss J. The increasing frequency
of heterosexually acquired AIDS in the United Sutes. 19831988. Am J Publ Health 1990:80:838-62.
15. Centers for Disease Control. Continuing increue in infectious syphilis—United Sutes MMWR 1988:37:35-8.
16. C e n u n for Disease Control. Relationship of syphilis to drug
use and prostitution—Connecticut and Philadelphia. Pennsylvania. MMWR 1988;37:755-8, 764.
17 Ward JW, Holmberg SD. Allen JR. et al. Transmission of
human immunodeficiency virus (HIV) by blood screened as
negative for HIV antibody. S Engl J Med 1988318:473-478.
18 Centen for Disease Control. Safety of therapeutic products
used for hemophilia patients. MMWR 198gJ7:44l-4, 44950.
19. Gwinn M. Fleming P, Oxtoby M, Green T , Mofenson L ,
Hannon WH. George JR. HIV seroprevalence in childbearing women and predicted incidence of perinatally-acquired
AIDS, United Sutes. Presented at VUth International Conference on AIDS. Florence. Italy, June 16-21.1991 (abstract
W.C.34) 2:31.
20. Oxtoby M, Byen R, Simondt RJ. Rogen M, Berkelman R
Age at AIDS diagnosis for perinatally-infected children,
United Sutes. Presented at Vllth Inumational Conference
on AIDS. Florence. Italy. June 16-21, 1991 (abstract W.C.
36), 2:31.
21. C e n u n for Disease Control. Update: Acquired immunodeficiency syndrome—United Slates. 1981-1990. MMWR
l99l-.40:35g-63. 369.
Copyright < 1992 by Atlantic Information Services, Inc.
D
September 1992
�O 5 .
O e .
9 3
1 O : 3 3
A M
P O T
*= F D A / C D E R / M U
The Journal of
Health Care Organization,
Provision, and Financing
oci t m
Volume 29
Number 3
Fall 1992
x
Lasker/Shapiro/Tucker
Realizing the Potential of Practice Pattern Profiling
McNeil/PedersenlGatsonis
Current Issues in Profiling Quality of Care
GauthierlRogallBarrandlCohen
Administrative Costs in the U.S. Health Care System: The
Problem or the Solution?
Gillis/Lee/WiUke
Physician-Based Measures of Medicare Access
LiUie-Blanton/Felt/Redmon/Renn/MachImf}Vennar
Rural and Urban Hospital Closures, 1985-1988: Operating
and Environmental Characteristics that Affect Risk
Coslcr/Lambrinos
Zidovudine's Impact on Resource Use by Patients with
Symptomatic HIV Illness: A Large Sample Analysis
Research Notes and Data Trends
Hellinger
Forecasts of the Costs of Medical Care for-Persons with
HIV: 1992-1995
Young/Cohen
•,.^>-. •
The Process and Outcome of Hospital Care for Medicaid
Versus Privately Insured Hospital Patients
BamettlBlairlKaserman
Improving Organ Donation: Compensation Versus
Markets
Published by the Blue Cross and Blue Shield Association
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
b
�Tab
This Section Includes:
6.
H e l l i n g e r FJ. F o r e c a s t s o f t h e c o s t s o f
medical c a r e f o r persons w i t h HIV: 19921995.
I n q u i r y 1992: 29:356-65.
�P o s
Fred J. Hellinger
Forecasts of the Costs of
Medical Care for
Persons with HIV:
1992-1995
This study concludes that the cumulative (national) cost of treating all persons with
the human immunodeficiency virus (HIV) rose considerably over the past year and
will continue to rise over the next several years. It is forecast that the cumulative cost
of treating all persons with HIV will increase 48% from 1992 to 1995 (from $10.3
billion to $15 J billion). It is estimated that the average yearly cost of treating a
person with AIDS is $38,300 and of treating an infected person without AIDS is
$10,000. The lifetime cost of treating a PWA is calculated to be $102,000. This is the
first study to use, along with other data, data from the AIDS Cost and Service
Utilization Survey to estimate the cost of treating persons with the HIV. The study
also projects the number of AIDS cases to be 66,300 in 1992, 76,200 in 1993, 86.800 in
1994, and 97,800 in 1995.
This paper forecasts that the cumulative (national)
costs of treating aU people with the human immunodeficiency vims (HIV) will increase from $10.3
billion in 1992 to $15.2 billion in 199S. This forecast
is significantly higher than most, and reflects both
an increase in the average amount of services used
by an individual with HIV and the availability of
data on all services used by this population.
Only personal medical care costs (hospital, physician, drag, nursing home, and home health care)
are considered in this paper. Nonpersonal medical
care costs such as testing, education, and nonmedical support services (transportation, housing, employment assisiance, and benefits counseling), and
the indirect costs of HIV measured by lost productivity are excluded.
People with HIV may be divided into three
groups: those with AIDS; those who will be included if the proposed expanded definition is
adopted (i.e., people with T-cell counts below 200
wbo do not meet the current definition); and those
with T-cell counts equal to or greater than 200. The
individual and cumulative (national) medical care
costs of treating people in each of these three groups
are forecast for the years 1992 through 1995.
Available evidence suggests that, at a given point
in time, the number of persons with HIV without
AIDS receiving treatment for HIV is equal to twice
the number of persons with AIDS (PWAs), and is
evenly divided among those who have T-cell counts
below 200 and those with T-cell counts equal to or
greater than 200. In this study, projections of the
number of PWAs are used to estimate the number of
persons with HIV without AIDS who receive medical care for their illness.
The proposed expansion of the AIDS surveillance
definition includes all persons with HIV who have
laboratory evidence that their absolute T-cell count
is less than 200 or their T-cell percentage of total
lymphocyte is less than 14, if the absolute count is
unavailable (CDC 1991b). All individuals who are
classified as PWAs using the 1987 (current) definition are included under the proposed expansion.
Forecasts of the number of PWAs diagnosed from
1992 through H
extrapolation r
models. Data or
during the 102 1
1984 and Nove
two sutistical <
rived from the
to those derivt
models.
Data from th
Survey (ACSU
Care Policy ai
center AIDS C
National Instit
eases (NIAID)
Spectrum of Di
the Centen fo
Yoric State Dc
Department of
lature;
the
Hawaii are us
individuals wit
This is the
ACSUS. The c
awarded to Wc
ber 1989. ACS
hensive survej
consumed by
interviews wit
spondents (50
and 750 PWA;
male intraveno
ual individual!
1991 to Fall 19
sites (e-g-.hos
tient clinics, pi
10 cities (Balti
les, Miami, N
<
Francisco, ant
plemented wi
(e.g., hospital
nursing home
the first wave >
July 1991) are
estimates of tl
HIV.
The MACS
interviews an
twice a year
bisex
FndJ. HtOngtr, PkA., is director. Division of Cost and Financing, CGHSER. Agency for Health Care Folicy and and PitU
|h
Research. Address correspondence to the author at AHCPR, Suite 502,2101 E. Jefferson St., Rockville. MD 20852.
the MACS is
Ir
y 29: 336-365 (Fall 1992). O 1992 Blue Ci and Blue Shield Asjociatkm.
356
0O46-95*rW2/2<»:i.W5M I 25
�Celts qfMtMcttl Cvtfar Ptnons vitk HIV
DStS Of
'ith
nd
e cost
r the
tdy
medical care
e three groupi
gh 1995.
x a given point
< HIV without
equal to twice
PWAi), and is
eT<eU counts
nts equal to or
iections of the
>-U»i)Minharnf.
a receive med)S surveillance
SIV who have
te T
count
entage of total
solute count is
id
who are
rum ) definid expansion,
iiajujosed from
VUU.MD2Q&S2-
1992 through 1995 are derived using two sutistical Infected persons without AIDS with T-cell counts
extrapolation models and two back calculation less than 200, or with T-cell counts equal to or above
models. Data on the number of AIDS cases reported 200. The MACS is one of the few sources of data on
during the 102 four-week periods between January utilization of persons with HIV broken down by
1984 and November 1991 are used to estimate the T-cell count. This study uses informationfromthe
two statistical extrapolation models. Porecuts de- MACS about the use of inpatient, outpatient dinic
rivedfromthe back calculation models are similar and physician office visits, home health care, and
to those derivedfromthe statistical extrapolation long-term carefroma fix-month period in 1990.
models.
The next section of this paper fore cute the numDatafromthe AIDS Cost and Service Utilization ber of PWAs for the yean 1992 through 1995. Tbe
Survey (ACSUS),ftindedby the Agency for Health third section examines the cost of treating an indiCare Policy and Research (AHCPR); the Multi- vidual with HIV. Separate eatimatei of the cost of
center AIDS Cohort Study (MACS), funded by the treating a PWA, an infected person without AIDS
National Institute of Allergy and Infectious Dis- with a T-cell count below 200, and an tnCseted
eases (NIAID) of the NIH; the Adult/Adolescent person without AIDS with a T-cell count equal to or
Spectnim of Disease Study (AASD), sponsored by greater than 200 are constructed. The fourth taction
the Centen for Disease Control (CDC); the New estimates the cumulative costs of treating persom in
York State Department of Health; the Califoraia each of the three groups. Thefifthsection contains
Department of Health Services: the Florida Legis- somefinalremarks.
lature; and the Ooveraor's Committee en AIDS in
Projecting the Number o/Caeae
Hawaii are used to estimate the cost of treating
individuals with HIV.
In this section, the number of AIDS cases are
This is the first paper to use datafromthe projected for the years 1992 through 1995. The data
ACSUS. The contract to conduct the ACSUS wu set used to project the number of AIDS cases it
awarded to Westat, Inc., (Rockville, Md.) in Octo- comprised of the number of AIDS cases reported to
ber 1989. ACSUS is the largest and moat compre- the CDC during the 102 four-week periods from
hensive survey to collect data oe medical services January 1984 to November 1991 (CDC 1986-1991).
consumed by penoos with HIV. It involves six To prqject the number of AIDS caaes reported to
interviews with each of approximately 2,000 re- CDC, two polynomial equations are estimated. Us.
spondents (500 asymptomatic, 750 symptomatic, ing the estimated coefficients for these equations
and 750 PWAs; of whom 350 are women, 700 are and time period designations from 105-136, the
male intravenous drug users, and 950 are gay/bisex- number of eases reported to CDC during the yean
ual Individuals) over an 18-month period (Spring 1992 through 1995 are prq)ected.
1991 (o Pall 1992). Respondents were enrolled at 27 Prqjectioni of the number of reponed AIDS cases
sites (e.g., hospital Inpatient wards, hospital outpa- are translated into prqjectiooi of the number of
tient clinics, private clinics, and physician offices) inAIDS cases diagnosed during a year by deriving ah
1 cities (Baltimore, Chicago, Houston, Los Ange- estimate of theratioof the number of diagnosed to
0
les, Miami, Newark, New York, Philadelphia, San reponed eases. The ratio of the number of diagFrancisco, and Tampa). The interview data is sup- nosed AIDS cases to the number of reponed AIDS
plemented with Information from provider bills cases haa fallen since 1986 (CDC 1991a). This ratio
(e.g., hospital, clinic, pharmacy, home health, and was 1.40 (18.152/13,008) in 1986, 1.29 (27,052/
nursing home bills) and medical records. Data from
20,946) in 1987. 1.07 (32.883/30,847) in 1988, 1.08
thefirstwave of ACSUS interviews (March 1991 to (37,227/34,340) in 1989, and .90 (37,005/40,916) in
July 1991) are used in this paper to help construct 1990. It is hypothesized here that the ratio ef the
estimates of the cost of treating an individual with number of diagnosed to reported caaes will be 1.0
HIV.
during the yean 1992 through 1995.
The MACS includes information coUected from
The aforementioned ratios were derived using the
interviews and physical examinations conducted number of diagnosed caaesreportedto CDC as of
"vice a yearfromapproximately 5,000 gay and December 1. 1991. Each of these ratios will inbisexual men in Baltimore, Chicago. Los Angeles, crease, since some caaes diagnosed during the yean
•nd Pittsburgh (Phair et al. 1990). Information from1986 through 1990 have not yet been reported to
the MACS is used to estimate the cost of treating CDC,
357
in
h ti
�TtfMWVafem* 29, Fall 1992
T i b b l . ProJacttouorAXDS
1993
Midpoint (Modtb A and B)
Midpoint (Models C snd D)
O K Workshop
Number ef PWAs ilive during
any part of calendar year*
\m
66,667
S8,0OO-S3,aO0
76.117
78,370
61.000-46.000
176.811
203.191
1994
86,790
86,532
19K
«7,I0J
91,16)
231,469
260,146
AIDS it about 90.000 (160,(
taiio of the number of pen
below 200 without AIDS tc
1.29(90,000^0,000). Since s
counts below 200 do not re
is an overestimate of the t
persons with T-cell counts
medical care for HIV to thi
rtported, so the upward adjustmeat Is 17.6% (VM).
* The estimates of the n m e ef PWAs alive duriag any pan of the year were derived by multlplyinj tbe md on for Forecasting Treatment C
u br
i pi t
Models A aad B by 2.667. (See HeUinier, F. J. 1991. Foracaiting the Medical Care Costs of the HIV Epidemic:
1991-1994. Inquiry JBO); J1J-223.)
Estimates of the lifetime c
haverangedfromS27,751
Arno 1990; Scitovsky. air
number of penons with T-cell counts less than 200
Projection Midpoints
etal. 1986; Hay and Kixer!
without AIDS to the number of PWAs is one).
context meansfromthe tit
The midpoints of the projections of the number of
until death.) Recent estiir
AIDS cues thai will be diagnosed during the yean Evidence supporting this hypothesis comes from
range of estimates from
1992 through 1995 using models A aad B are pre- an array of sources including the Adult/Adolescent
most at the high
of th
sented in Table 1. ' The nridpoim of the projections Spectrum of HIV Disease study (AASD) conducted
medical care in i
ttui!
for models A and B are well within tbe range of theby the CDC. The AASD began active surveillance
care to the purchaser of h
e
latest CDC projections (CDC 1990). Thefiguresin at more than 50 sites in nine cities during 1990.
represent payments froir
Table 1 have been adjusted upwardtoreflect the Medical records obtained on 7,635 HIV-infected
provlden.
CDC estimate that «5% of all AIDS caaes arc re- individuals revealed that, "For every person with
AIDS at these sites, two additional persons with
^ported (Roaenbtum et al. 1991).
In this section, the cc
W
Tabl 1 also presents prctjectiotu for the number HIV were receiving medical services" (Farizo et al.
Table
estimated. Subsequently
of persons dlagnoaed with AIDS using the back 1992, 1798).
Infected person with H
calculation method.* This method uses data on the Dau presented in arecentarticle by Hellinpr
T-cell count below 200 i
number of AIDS cases and knowledge about the (1991) also showed that the number of people with
cost of treating infec
progression ratefrominfection to AIDS (i.e., the HIV without AIDS receiving medical care was
with a T-cell count equr
incubation period distribution) to "back calculate" equal to approximately twice the number of PWAs
estimated.
the number cf people who must have been Infected receiving medical services. Hellinger presented
with HIV In prior time periods to have resulted indaufrom10 sources including tbe state of Hawaii,
Persons with AIDS (PV
the observed number of AIDS cases (Brookmeyer the San Francisco City Clinic Hepatitis B Study,
The inpatient cost of tn
and numerous clinics and hospitals.
1991: Hay and Wolak 1990).
is derived by muldplyli
The midpoints of the estimates obtainedfromthe Dau from the MACS reveal that theratioof the
length of suy(ALOS),
•talistical extrapolation models and the back calcu- number of penons with T-cell counts less thflfl 200
and the number of
lation models are similar (see Table 1). The mid- without AIDS to PWAs is 1.28 (Schnger 1991). This
outpatient services is
points of tbe pngectiou derivedfromthe two su- figure is an overestimate of the ratio of the number
estimate of inpa
cc
tistical extrapolation models are used in subsequent of infected persons with T-cell counts less than 200
of treating a PWA dur
calculations of the cumulative (national) cost of the who receive medical care for their illness to the
HIV epUemic.
number of PWAs because tome infected penons
Inpatient costs - ALC
with T-cell counts less than 200 may not have bePrqjeed/ig AIDS Caw under the Proposed
perc
come aware of their infection had they not enrolled
Definition end People with HIV without AIDS in the MACS.
i
Receiving Medical Care
About 1 % of the people who have tested positive
6
Outpatient costs " C
It is hypothesised here that one-third of all people for HIV in the U.S. Army have T-cell counts below
c
with HTV receiving medical care are PWAs, one- 200 (CDC 1991b). Since CDC estimates that about 1
third have T-cell counts below 200 (and do not havemUlion people are infected with HIV in the United
AIDS), and the remaining one-third have T-cell Sutes, this suggests that about 160,000 persons
Calendar year c J •
counu equal to or greater than 200 (and do not havehave T-cell counts below 200. CDC dau reveal that
AIDS). This implies that the number of infected about 70,000 people were alive with AIDS as of
people without AIDS with T-cell counu below 200 December 1991 (CDC 1991a). Thus, the number of
Average length oj
Is equal to the number of PWAs (i.e., theratioof the persons with T-cell counts less than 200 without
suy (ALOS) for a
358
�1 O : 3 3
A M
Com ofMttictl Cvt for Ptnons with HIV
19M
)
I
91.803
91 Mi
9
260.846
3% of all CUM V l
ing the midpoint for
tw HIV Epidemic:
t
unts
than 200
PWAi is one),
thesis comes from
Adult/Adolescent
AASD) conducted
ictive sunrelllanee
:iues during 1990.
,635 HIV-infected
every person with
persons with
icei^Farlzo et tl,
ce^pa
rtlcie^y Hellinger
iber of people with
medical care was
e number of PWAs
[ellinger presented
che state of Hawaii.
Hepatitis B Study,
itals.
that the ratio of the
.-ounts less than 200
Schiager 1991). This
ratio of the number
counts less than VO
their illness to the
ne infected person*
0 may not have bead they not enrolled
.have tested posiii"*
. T-cell counts belo*
.dmatesthataboutll
h HIV in the United
ut 160.000 pet*"*
rnCLdata reveal tl*
3
jtl^^he numbc •
. than 200 withe*
S8
hospitalized in Florida was 1 days in 1990 (La5
crosse 1991). In New York sttte, the ALOS for a
PWA hospitalized during 1990 was 19.2 days (New
York Sute Department of Health 1991). Dau from
the California Department of Health Services ny
vealed that the ALOS for a Medi-Cal recipient
hospitalized with AIDS was 11.6 days in fiscal year
1989, and dau provided in a report Issued by the
Ooveraor's Committee on AIDS in Hawaii indicated the ALOS for a PWA was 17.4 days in eariy
Forecasting Treatment Costs
1990(Hiehle, Maxfield, and Kizer 1990; Ooveraor's
Estimates of the lifetime costs of treating a PWA Committee on AIDS 1990). The only recent dab
have ranged from $27,751 to $147,000 (Green and from a national survey isfromthe ACSUS tor the
Arno 1990: Scitovsky, Cline, and Lee 1986; Hardy periodfromMarch to July 1991. The ALOS for the
etal. 1986; Hay and Klzer 1992). ("Lifetime" in this 677 PWAs in the ACSUS was 16.3 days. In this
context meansfromthe time of an AIDS diagnosis study, it is assumed that the ALOS for PWAs is 16.3
until death.) Recent estimates have narrowed the days.
nnge of estimatesfrom$40,000 to $85,000 with
m s at the high end of this range. The "costs" of Average hospital charge per day: The avenge
ot
medical care in these studies represent the cost of hospital charse per day for a PWA in New York
can to the purchaser of health services, and usually sUU rosefrom$921 In 19891051,004 in 1990 (New
represent paymentsfrominsuren to health care York Sute Department of Health 1991). The average charge in 1989 for a hospital day of care la
praviden.
In this section, the cost of treating a PWA is California for a PWA was $1,130 (Hlehle, Maxfield,
attimatcd. Subsequently, the cost of treating an aad Klzer 1990). In this study, it It estimated that
Infected person with HIV without AIDS with a the avenge chargetora hoepital day ia $1,100.
Number of hotpltallzations per calendar year.
T-cell count below 200 is estimated, aad then the
The 1985 and 1987 NPHHI surveys found that the
cost of treating an infected person without AIDS
with a T-cell count equal to or greater than 200 is number of hospitalizations per PWA was 1.6 (Andnilis et al. 1987; Andnilts, Been, and Qase 1989).
estimated.
Tbe 1988 NPHHI survey buad It to be 1.7 (Ahdndis
and Kathbun 1991). DaUfromthe New Yoric Sute
Ptnoiu with AIDS (PWAs)
Department of Health indicau that the number of
The Inpatient cost of treating a PWA during a year botpiullzations per PWA was i.6 in both 1989 aad
it derived by multiplying estimates of the average 1990. It is hypothesized in this paper that tbe averkngth of suy (ALOS), the average charge per day, age number of hospitalizatioot per PWA is 1.6.
m the number of hospitalizations. An estimate of Inpatient costs. Using thefigurespresented in
d
outpatient services is derived and added to the the preceding paragraphs implies that the inpatient
utitnate of inpatient costs to estimate the total cost cost of treating a PWA alive duriag any part of a
cf mating a PWA during a calendar year.
year is $28,700 (16.3 x $1,100 x 1.6). ThisfigureIs
higher than recent estimates becauae of longer e»bpetieat costs - ALOS x Average hospital charge tlmaut for the avenge length of suy aad increased
per day x Number of hospitaliza- estimates of the average hospital charge per day.
tions
Outpatient costs. In New Yort state, outpatient
care constituted 1 % < f total costs in 1988, 22% in
2 r
friipaticDi costs - Outpatient visits + Long-term and 1 % in 1990. Although the proportion of
1989,
8
care + Home care + Drags
cost attributable to outpatient care fell between
1989 and 1990 in New York auu, the absolute
C^kodar year costs = Inpatient costs + Outpatient amountrotefrom$168 million to $174 million becosu
cause the amount spent oe medical care coets for
AIDS rose 28%fromS768 million in 1989 to $986
Avtragt length of stay. The average length of million in 1990 (New York State Depanmcnt of
(ALOS) for a Medicaid patient with AIDS Health 1991).
AIDS is about 90,000 (160,000 - 70.000), and the
ratio of the number of persons with T-cell counts
below 200 without AIDS to the number PWAs is
1.29 (90,000/70,000). Since some people with T-cell
counts below 200 do not receive medical can, 1.29
it an overestimate of the ratio of the number of
persons with T-cell counts below 200 who receive
medical care for HIV to the number of PWAs.
359
�O 5.
O B .
9
3
1 O
3 3
A M
P 1 2
* F D A ^ C D E R / M L.
li^mirjfVoUmt 29. Fall 1992
ratio of the mean to the median survival for PWAs
was estimated to be 1.09 (19.24/17.66) in San Francisco (San Francisco Department of Public Health
OtUanlm
NrwYorts
1988). Recent daufromthe New York City Depart(IHO)*
(HW)»
ment of Health indicate that the median survival for
lepeiiaat
72.8
13.3
someone diagnosed with AIDS Is 18 months. AsOutpatieat boipilal
S.7
S.3
suming that the ratio of the median to mean survival
Leeg-term cam
.7
1.3
Homs can
2.6
4.2
is 1.09 implies that the mean survival time is 19.6
AZT aad other drags
5.6
13.3
months (1.09 x 18).
• New Yoifc Stale Depanacat ef Health. AIDS In Ntw If the number of people that contract AIDS during
York Stutt 7Vp*f* I960, Albany, 1991.
Hkbk, GL. MaxfleU. W. T.. aed Kiaer. K. W.. Mtdi- the year occurs at a constant rate and the mean
Cal Studiti la AIDS Dtmogmpkici and Expenditurti far survival time is 20 months, then the average number
Persons with AIDS. 19/0-1989, CaUfomia Departmeat ofof months lived by e PWA alive during any part of
Hctltt Set nun: SiUHimiUu, March 1990.
a year is 7.5 months and the ratio of the number of
PWAs alive during any part of a year to the number
In Califomia, outpatient aervices comprised 27% of PWAs diagnosed that year is 2.667 (Hellinger
of all Medi-Cal expendituresforPWAs infiscalyear 1991). It is estimated here that the average medical
1989. 25% infiscalyear 1988, and 17.5% in fiscal care cost of treating a PWA alive during any part of
year 1987. Areportissued by the Florida Legisla- a year is $38,300. If the average number of months
ture in November 1991 calculated that outpatient lived by a PWA alive during any pan of the year is
can comprised 6 of medical care costs to treat 7 J , then the average cost per month of treating a
%
PWAs on Medicaid in 1988, 1 % in 1989. and 23% PWA is $5,100, and the avenge lifetime cost of
9
In 1990. Thisreportpredicted that thisfigurewould treatment Is $102,000 ($5,100 x 20).
rise to 24% in 1991 aad 26% ia 1992.
In this study, it is estimated that outpatient can Wetted Individuals Without AIDS with T-Cell
constitutes 25% of the total cost cf care (I.e., Counts Less Than 200
89,600). Using information provided In Table 2, the It is estimated in this section that the average cost
cost of the components of outpatient care an esti- of treating an infected peraon without AIDS with a
mated to be: S3,660 for outpatient visits (clinic or T-cell count less than 200 Is $13,525 per year (see
physician office), S420 for long-term care coats, Table 3). This Is slightly more than one-third of the
81,460 for home care costs, and $4,060 for out- cost of treating a PWA ($38300). The $13,525 estipatient drugs.
mate is higher than existing estimates of the avenge
Ufertme tnatmmt eosts per PWA. The preced- cost of treating an infected person without AIDS.
ing analysis indicates that the cost of treating a Estimatei of the cost of treating infected penons
PWA alive duriag any part of a year is $38,300 without AIDSrangefrom$1,850 to $9,700 per year
($28,700 * $9,600). It Is estimated here that the (Hellinger 1991). However, individuals with T-cell
mean survival of a PWAfromthe time of diagnosis counts below 200 comprise the most severely ill
until death ia 20 months, aad that the lifetime cost of
among infected penons without AIDS, and their
treating a PWA is $102,000. This is higher than cost of care exceeds the average cost of treating all
recent estimatea, and reflects enhanced longevity, Infected penons without AIDS.
higher hoepital charges per day, and slightly longer It is estimated that the hospitalization costs for a
lengths of suy.
person with a T-cell count less than 200 is $7,603 a
DtaUfromtbe San Francisco Department of Pub- year. Daufromthe MACS on utilization during a
lic Health Indicate that the mean survival time of the six-month period in 1990 indicated that the likeli4,994 individuals diagnosed with AIDS between hood of an infected penon without AIDS with a
1988 and 1990 was 19.24 months.' Most estimates of T-cell count less than 200 being hospitalized was .36
survival of PWAs arereportedas medians (the times that of a PWA (Schnger 1991). DaU from
length of time below or above which 5 % of PWAsACSUS reveals that the average length of suy for
0
will survive). Mean survival estimates (the average persons with HIV but without AIDS is 1 days, and
2
number of months lived by a PWA) are used to that the ALOS varies little between those with and
derive cost estimates because each month of sur- without symptoms. Assuming that each person was
vival entails the use of medical careresources.The hospitalized only once during this period, that the
TaUe 2. Medial care mpunrtlfre by aerviee
(pereentage ef toCal)
4
k
360
Table}. Erimatad
Inpatient, hospital
Outpatient
Visits (clinie or office
Long-tenn care
H m health can
o e
Drup
Outpatient subtoul
Total
average length of st
average charge per c
cost of hospitalizatit
$1,100 x 1.6 x .36).
Dau from MACS i
and 92.5% of
*
without AIDS receivi
6-month period in 1
indicates that the
over a 3-month x
a symptomatic vi
vey of clients of the
dation's AIDS 1
communities rc
8.06 outpatient phyi
3-month period prio
1992). It is assumed!
T-cell counts
tht
82% (6.2/7.6) of the
physician services
amounts to $3,001 p
DaUfromthe M
T-cell counts
tl
times as likely to
PWAs. Daufrom.
matic individuals arc
health services as f
home health care for
200 cost $423 (.29 x
cost for home tl
PWAs take an a <
v
tions during a 3-nu
from the ACSUS. S
average of 4.8 at
average of 2.7. It i
costs for those with
AIDS amount to $2
�COM qf Medical Care for renoiu with HIV
ir^por PWAt
.66) in San Franof Public Health
'ork City Departxiian aurvival for
11 months. Asto
survival
'ival
: is 19.6
4
tact AIDS during
x
the mean
:a
number
uring f part of
of the number of
ar to the number
2.667 (Hellinger
average medical
luring any part of
imber of months
(of the year It
nth of treating a
lifetime cost of
5 with T-Cell
i ^e g
h fn e
thWtnge cost
out AIDS with a
'15
year (see
i one-third of the
Ihe S13.525 estiof the average
i without AIDS,
infected persons
> $9,7 per yev
j »with T-cell
: severely ill
AIDS, and their
jst of treating all
Table 3. Estimated average cost of treating a penon with HIV (in 1991 dollars)
HIV
T-ed
H V portttv*
T
AIDS,
PWAI
T-cdleauihM
2,323
Inpaiieni, hospital
28,700
7,603
Outpatient
2,633
3.001
Visiti (clinic or office)
3,660
Long-term care
420
8
1
423
H m health care
o e
1.460
1.403
2.498
Drugs
4.060
,4,121.
Outpatient subtotal
AMOj
sm
Tool
38,300
AIDS,
13.323
average length of sUy is 12 days, and that the Infected IndMduaU Without AIDS with T-Celt
average charge per day is Si. 100 implies that the Counts Equal to or Greater Than 200
cost of hospitalization is 37,603 per year (12 x It is estimated here that the medical care costs of
$1,100 x 1.6 x .36).
treating an infected penon with a T-cell count equal
DaufromMACS revealed that 93.1% of PWAt to or greaur than 200 without AIDS is $6,444. DaU
and 92.5% of those with T-cell counts less than 200 from the MACS ladicaus that the chance of a
without AIDS received outpatient services during a penoa with a T-cell couat equal to or greater thaa
6-month period in 1990. Daufromthe ACSUS 200 without AIDS being hospiuHzed during a sixindicates that the mean number of ambulatory visits month period is 1 % of the chaaces of a PWA baiag
1
over a 3-month period was 7.6 for a PWA and 6.2 for hospitalized during the same period.
a symptomatic individual. DaUfroma recent surIt is estimated here that inpatient hospital charges
vey of clients of the Robert Wood Johnson Foun- for persons with fcatt counts equal to or greater
dation's AIDS Health Services Program in nine than 200 amount to $2,323 per calendar year!
communitiesrevealedthat the S17 PWAs avenged [(.ll)x(1.6)x(l2)x(Sl,100)].AaaotediAthepnvi.
8.06 outpatient physician or clinic visits during the out section, tlie ALOS tor asymptomatic and symp3-month period prior to the interview (Mor et al. tomatic penons in the ACSUS study wen of similar
1992). It is assumed here that infected penons withmapiitude and equal to 12 days. It ia assumed that
T-cell counts less than 200 without AIDS consume the avenge iapatieat chaise per day for penons
82% (6.2/7.6) of the outpatient hospital, clinic and with T-cell counts equal to or greater than 200
physician services consumed by PWAs. This without AIDS is equal to the avenge inpatient
amounts to $3,001 per year.
charge for a PWA.
Daufromthe MACS Indicau that those with
Dau from the MACS indicau that 74% of those
T-cell counts less than 200 without AIDS are .29 penons with T-cell couats equal to or greater than
times aa likely to use home health services as 200 without AIDS obtained outpatieat care during a
FO'lU. 1 a twu AGSUS indicate that sympto- 6-month period Wf99ariSiar'SSS t^ AC3VS
matic individuals are .43 times as likely to use home reveal that asymptomatic individuals had an averbealtb services as PWAs. It is assumed here that age of S.S outpatient visits dartag a Hnonth period
home health care for those with T-cell counts belowin early 1991 and PWAs had an avenge of 7.6 visits.
200 cost $423 (.29 x $1,460) or 2 % of the $1,460 It This implies that a penoa with a T-cell count equal
9
to or greaur than 200 without AIDS had .72timetaa
cost for home health services for a PWA.
PWAs take an average of 7.8 prescribed medlca- many outpatient vitiu as a PWA. It is hypothesized
tioas during a 3-month period according to daU in this paper that charges for outpatieat can tor
penons with T-cell counu equal to or greater thaa
from the ACSUS. Symptomatic individuals take an
avenge of 4.8 and asymptomatic individuals take an 200 amount to $2,635 (.72 x $3,660) per calendar
average of 2.7. It is hypothesized here that drug year.
costs for those with T-cell counts below 200 without In the MACS, penons with T<ell counU equal to
AIDS amount to $2,498 [(4.8/7.8) x $4,060].
or greaur than 200 wen 1/18 aa likely as PWAs to
u
:ation costs for a
n200isS7.603 a
lization during a
d
the likeliut AIDS with a
pitaUzedwas.36
991). Dau from
sngth of suy for
S is 1 days, and
2
n those with and
r^ferson was
. ^ P . that the
361
mil
�O 5.
O 6.
9 3
1 O : 3 3
A M
F
* F D A / C D E R / M l~
1 4
ImiulFyfVoUmt 29, Fall 1992
Table 5. Drugs
Table 4. Comuladve projected net of treating people with HIV in bilUoni of 1991 dollan
1992
1) Cost of Mttiag people with AIDS
2) Cost of treahag people with HIV
without AIDS sad T<eU count
below 200
3) Cost of mating people with HIV
without AIDS aad T<eU count
equal lo or greater than 200
Cost at tnatiaa all people with HIV.
(1) + (2) + 0)
Perceauwa inemse in eost at treatiag
im
199*
IMS
6.771
7.782
6.863
9.990
2.391
2.748
3.131
3.528
1.139
1.310
1.492
1.681
10.301
11.840
13.488
13.199
13.0
13.9
Cbcmkal/
12.7
P
O-CSF
N
OM-CSF
WtlewOBrtHprG-month pe- pected number of persons under treatment. The
cost
riod in 1990; About 1 of those with T-cell counts of treating this population is forecast to be
%
equal to or greater than 200 utilized home health $2,391 billion in 1992 and $3,527 billion in 1995.
services and about 1 % of PWAs utilized home
8
The cost of treating a person with HIV without
health services duriag this period. It is hypothesized AIDS and a T-cell count equal to or greater than 200
here that charges for home health care for personswas calculated to be $6,444 per calendar year. It is
with T-cell counts equal to or greater than 200 equalforecast that the number of people In this group that
S81 per calendar year [(1/18) x $1,4(3].
receive treatment is equal to the number of PWAs.
In the ACSUS, it was found that asymptomadc Thus, to obtain an estimate of the cumulative cost of
HIV positive Individuals took an average of 2.7 treating penons with HIV without AIDS with T-cell
prescription medications during a 3-month period counts equal to or greaur than 200, the cost of
while PWAs took aa avenge of 7-8 prescription treatment per penon ($6,444) is multiplied by the
medications. It is assumed here that drag cosu for expected number of penons under treatment. The
persons with T«cell couats equal to or greater than cumulative cost of medical care for this population
200 amount to $1,405 ((2.7/7.8) x $4,060] per cal- is estimated to be $1,139 billion in 1992 aad $1,681
endar year.
billion in 1995.
It is forecast here that the cumulative cost of
Projected Cumtdatfre Coats
treating PWAs using the proposed expansion of the
Table 4 presenu estimates of the cumulative costs definition of AIDS is 35% greater than the cost of
of treating people with HTVfrom1992 through 1995. treating PWAs under the current definition. For
The cost of treating a PWA alive during any part of example, the cost of treating PWAs in 1992 is forea calendar year is estimated to be $38,300. To obtaincast to be $6,771 billion under the current definition
an estimate of the cumulative cost of treating PWAs and $9,162 billion ($6,771 billion + $2,391 billion) if
la 1992, we multiply $38,300 by the number of all infected people with T-cell counts below 200 are
penons expected to be alive with AIDS at any time
defined as PWAs (see Table 4).
during 1992 (176.789). Thus, it Is estimated that it
The cost of treating all people with HIV is forewill oost $6,771 billion in 1992 to treat all PWAs. cast torise4 %from$10.3 billion in 1992 to $15.2
8
The estimates for the cumulative cost of treating billion in 1995. The rate of increase during these
PWAs ia the other yean are constructed in a similaryean is expected to decrease from 1 % in 1993, to
5
feahion.
13.9% in 1994, and to 12.7% in 1995.
The cost of treating a penon with HIV without
AIDS and with a T-cell count bdow 200 was esti- Final Remarks
mated to be $13,525 per calendar year. It is forecast Estimates of the cost of treating a PWA have
that the number of penons with HIV without AIDS climbed steadily over the past few years. Hellinger
with a T<ell couat below 200 receiving medical care (1991) estimated that the "lifetime" (i.e., from
equals the number of PWAs. To obtain an estimate AIDS diagnosis until death) cost of treating a PWA
of the cumulative cost of treating people with HIV was $57,000 in 1988, $75,000 in 1990, and $85,333 in
without AIDS aad T-cell counts below 200. we 1991. In this study the lifetime cost is estimated to
multiply the cost of treatment ($13,525) by the ex- be $102,000. The soaring cost of care is epitomized
Epoetin Alfa
L
Foscamet
f
utilize boarlN
362
ddl (didsnoiine) \
Clarithromycin 1
Azithromycin
by the en lero
Shield, the singl
with HIV. whoi
time of HIV in
$75,000 In the
1991 (Advltor's
This study c
persons with H
year, and that ii
several yean. E
that both the svt
average cost of
ofHIVlnfecttoi
than previously
One year a e
g
cost of treating
average cost o
without AIDS ^
�treatment. The
t foi
i to be
illion in 1995.
ith HIV without
greater than 200
endar year. It Is
n this group that
imber of PWAs.
imulative cost of
VIDS with T-cell
O y ^ c o s t of
m l ^ P by the
treatment. The
• this population
1992 and $1,681
lulative cost of
:xpansionofthe
the cost of
definition. For
> in 1992 is forejrrent definition
$2,391 billion) if
:s below 200 are
ith HIV is forcn 1992 to $15.2
ie during these
1 % in 1993, to
5
5.
a PWA have
i. Hellinger
e" (i.e., from
irea^aPWA
. a^n,333 in
islRKatedto
e is epitomized
by the experience of Empire Blue Cross and Blue paper, it is estimated that the yearly cost of treating
Shield, the single largest private iniurer of penons a PWA is $38,300, and that the average cost of
with HIV, whose cost of care per personfromthe treating an individual with HIV without AIDS is
time of HIV infection until death increased from about $10.000[($13,325 + $6,444V2].
$75,000 in the mid-1980s to more than $150,000 in The increase in the estimated cost of treating a
1 9 [Advisor's Insurance Digest 1992).
91
PWA is attributable torisesin the hospital charge
This study concludes that the cost of treating per day and the avenge length of stay. The avenge
persons with HIV rose considenbly over the past hospital charge per dayforPWAs has increased
year, and that it will continue toriseover the next from about SI .000 to $1,100, and the avenge length
several years. Evidence is presented here indicating of stay for PWAs in the ACSUS was 16.3 days for
that both the avenge cost of treating a PWA and the the period between March and July 1991. The estiaverage cost of treating an individual from the time mate of the average length of stay used in last year's
of HIV infection until an AIDS diagnosis are higher study was 15 days.
than previously calculated.
The estimated cost of mating a person with HIV
One year ago, it was estimated that the yearly without AIDSreflectsincreases In both Inpatient
cost of treating a PWA was $32,000, and that the and outpatient costs. Persons with HIV without
average cost of treating an individual with HIV AIDS in tbe ACSUS experienced average lengths of
without AIDS was $5,100 (Hellinger 1991). In this stay equal to 12 days. This is higher than the 8-day
363
�MX-
F 1 6
ImiuirylVelw* 39. Fall 1992
euimau used in list yev'i itudy. The ACSUS
itudy also revealed that people with HIV without
AIDS ute taite quantltiet of outpatient aervices
(e.g., asymptomatic individuals without AIDS had
an annualized average of 22 outpatient visits, and
symptomatic individuals without AIDS had an annualized avenge of 23 outpatient visits).
In addition, the widespread use cf new and expensive drugs has contributed to the high cosu of
treating penons with HIV. AZT was approved for
use in persona with T-cell counts below 500 in
March 1990, and ddl was approved in October 1 9
91
to treat HIV in penons who cannot use AZT.
Patients using AZT often experience anemia (low
red blood cell count) and neutropeniaflowwhite
blood cell count). Epoetin Alfa (EPO, brand name
Procrit) was approved in January 1 9 to treat low
91
red blood cell counts in patients with HIV (see
Table 5). This drug cosu about $200 a week for
three 7,000-unlt doses. O-CSF (granulocyte colony
stimulating factor) was approved In February 1 9
91
to treat cancer patienu with neutropenia, but also is
used by some physicians to treat neutropenia in
penons with HIV. This drug costs about $150 a day
(PI Penpective 1991). The use of expensive drugs to
prevent opportunistic infections (e.g., acyclovir to
prevent CMV,fluconazoleto preventftingalinfections, and pyrimethamine to prevent toxoplasmosis) (Smith 1991), also contributes to higher costs
(Bniens 1991).
Notes
Tht author U grottfid io Ira E. RtuUt. Agtieyfor HealthThe hack calculation (brecasu were derived using data
Cm Poticy and Rtinnh. USDHHS, for Us vafoeMe on the number of dlagnoted AIDS eases during tbe
cemmMM. Th* vfewi txpnsMti bt this enfeic vt solelyyean 1981 througb 1989, and pragnssion ntct prothose iff the author. No <&cial emdoneme* bf AHCPR is
tenicd by Bacchetti and Moss (1989) for the yean 1 8
91
Intended or thouU be btfemd.
thraugh 1987. In model C, it is aasuaed that the proinssioe me fell 2 % ie 1988,40K in 1989, aad 5M for
0
TOe eatimatod oocfldentsforModel A are:
the yean 1990 through (995 due to the use of AZT.
aerosol pentamMine, Bactrim/Septn, and a variety of
Model A: C T - 47.133 + 38 J47r - .Oni*
other drugs used to prevent opportuniitie infections
(131.503) (3J»t) (.055)
((Ml. Kosenbc*. aad Goedect 1990; VoIbeitUng «t al.
when Cr it Uu number of eases fepoftadrfiuiostha 7* 1990). In model D, it If assumed that the drop ia the
4-week period, r Is the number of the time period (time progression ntc occurred smoothly ever the 3 yean
period I UtheflNt4weeksef !9»«). and the values in
1988,1989, and 1990 (i.e., the progreision rata dropped
tht pweatbeses an the staadaid enrm of tbe cocfll17« In 1988,33% is 1989, aad SO* in 1990 and each year
deau above them. The square of the correlation coef- thenafter).
Adeot for model A is .864. The estimated eoaAcieau for
Panooal eommuoication with George Lemp (San
node) B an:
Pneclaco Dapannent of Public Health). December 16.
Modal B: € / * - 16.849 + .4647
1991.
(.939) (.016)
Personal commuaicatloa with Steve Blum (Office of
Tbe square ef tha eomlatloe coefflcientformodel B Is Epidemiological Research, New Yoric City Department
.891.
of Haafth). December 12, 1991.
Advisor's Insunnce Digest. 1992. "New York State Medence, sponteicd by AHCPR aad HRSA, U.S. Public
leal laaureaee Legislative Rafonn." Jaouaiy: 3-4.
Health Service, Miami, Fla., December 6.
AIDS TrntmuuNevt. 1990. "FDA Approves, NIH Rec- Bacchetti, P.. and A. Moss. 1989. Incubation Period of
oameodi AZTforAsympteoatics." No. 98, March
AIDS la San Fnacisco. Nature 238(March 16):2512.
253.
AadniUs. D. P., V. S. Bean, J. D. BenUcy, and L. S. Brookmeyer. R. 1991. Reconstracdon and Putun Trends
Oagt. 1987. The Provision and Plaaadng of Medical
of the AIDS Epidemic in UM United State*. Science
Can for AIDS in U.S. Public and Private Teaching
2S3(July 3):37-42.
Hospitals. Journal of the American Medical Asioci- Bniens. K. 1991. TPN: Total Paranterel Nutrition. PI
athsn 23«(IQ):1343-I346.
Perspective It (October): 14-15.
AndniUs. D. P.. V. S. Been, and L. S. Qage. 1989. Ihe Centen for Disease Control. 1991a. A1DSIHIV Surveil1987 U.S. Hospital AIDS Survey. Journal of the
lance Report, December 1.
American Medical Association 262(6):7t4-794.
- 1991b. Dnft of Revised CtaaaUlcatlon Systcoi for
rulis. D. P.. aad J. A. Ratbbua. 1991. 1988 U.S.
HIV Infection and Expanded AIDS Surveillance Case
•
Hospital AIDS Survey: Aa Analysis ef Hospitals in
DefinitionforAdolescents and Adults, October 22.
the Tide 1 Ryan White Care Aet Cities. Praiented at
1990. Estimates of HIV Prevalence end Prelected
the HIV-AIDS Health Servioei and Delivery ConferAIDS Cues: Summary of Workshop, October 31364
November 1,
Report 39(7): 110-1
1986-1991. M
o
port, issue number
through 40.
Farizo, K., J. Buehler
Spectrum of I >
nodeficiency Vint
Journal of the Amer
1798-1805.
Gall, M. H.. P. S. Ro*
Therepy May ExpL
dence. Journal of A
drones 3(4):296-30
Governor'i Committee
Esiitnaus for He*
tientt. Honolulu, H
Onsen, J.. and P. Arm
MDS . Journal of th
264(l0):1261-]266.
Hardy. A. M.. K Rauct
andJ. W. Cunao. 1
<
Fint 10,000 Casei
Syndrome ia the Un
ican Medical Assoc,
Hay, J. W.. and K. Klz.
for Persou with Al
st the Americas 1
Meeting, New Oita
Hay. J. W.,afldF. A. V
AIDS Prqjeclion M
c
ear Inequality-Conit
tutiaa, WorUag Pap
Hellinaer. F. J. 1991. Foi
ofthcHIVEpidemi
225.
Hlehle. Q.. W. T. MM
Medi-Cal Studies in
dltunt JOT Ptnons*
Department of 1
�1 1 : O 6
A M
Cosu of Medical Care for Persons with HIV
1 9 ^ 9 treat low
ts with HIV (see
$200 a week for
j
ocyie colony
l in February 1991
JI, but also is
. neutropenia in
ist
$150 a day
expensive drugs to
(e.g., acyclovir to
event fungal Infece
i toxoplasmotes to higher cosu
•n
red using dau
OS
durini (he
rograukm rate* pre«9) the yean 1 S
91
J
id thai tbe proOn]
andSCMfor
: to the use of AZT.
ftra,
a variety of
port^jgic infactioni
^hUTdrop in the
Jdy over the 3 yean
ijreM ion rate dropped
»in 1990 and each year
November 1, 1989. MorbUilry and Mortality Wtekty LaCroese. J. 1991. AIDS Caseloads and ggpandisurei In
Florida. Florida LeglaUtare. Jdat Lagislativs MaaReport 39(7): 110-119.
agement Committee, Tallahaasee.
1986-1991. Morbidity and Mortality Weekly Report, issue number 52 from each of the volumes 33 Mor, V., J. A. Fleishman, M. Dresser, and J. Ptotte. 1992.
Variadon in Health Service Use Among HTV-btfectad
through 40.
Penoos. Medical Care 30(1): 17-29.
Faruo, K., J. Buehler, M. Chamberiand. et al. 1992.
Spectrum of Disease In Penons with Human Immu- PI Perspective. 1991. "Guide to Opportunistie Infecnodeficiency Virus Infectloa in the United Stales.
tion i." No. II (October): 11-13.
Journal of the American Medical Aiioelatlon 267(13): New York Sute Department of Health. 1991. AIDS In
1798-1805.
New York State Through 1990. Albaay.
Call. M. H.. P. S. Rotenbcrs. and J. P. Ooedett. 1990. Phair. J., A. Muaoz, R. Detelt, et al. 1990. UM Risk ef
Therapy May Explain Recent Deflcits in AIDS InciPneumocystis Carinii Pneumonia Aaooi Men with
dence. Journal of Acquired Immune Deficiency SynHuman Immuoodeflciency Vlnis. Now England Jou*
dromes 3<4):296-306.
nal qf Medicine 3220):161-165.
Ooveraor's Committee en AIDS. 1990. Costs of Care RosenUum, L. S., J. W. Buehler, M. W. Moqae^ et al.
Estimates for Hawaii HIV-infected and AIDS Pa1991. Conplcicoess ef AIDS Repertlag ie the United
tUnu. Honolulu, Hawaii, June 5.
States, 1988. Presented at the Seventh Interaatlooal
Greea, J., and P. Arno. 1990. The Medkaidizatioo of
Cooteenee oa AIDS, noroaea. Italy. MX. 1314,
AIDS. Journal of the American Medical Association
Juoe 17.
264(I0):126M266.
San Prandseo Depanment ef Public Health. )9>8. AIDS
Hardy, A. M., K. Rauch, D. Bcbcnberg. W. M. Morgan,
in San Franeisco: Stutus Rtpon for Fiscal Year
and J. W. Curraa. 1986. The Eoonomie Impact of the
1987-88 and Pr&ctbns of Service Needs and Costs
Pint 10,000 Caaes et Acquired Immunadeflciaacy
for 1988-1993. Saa Fraadaco.
Syndrome in tbe United States. Journal cf the AmerSehngsr, L. 1991. Cost and Udliaadoa bsue^-the
ican Medical Association 255:209-211.
MACS. Praiented at the HIV-AIDS Health Servicet
Hay. J. W., and K. Klzer. 1992. Medi-Cal Expenditures
Research and Delivery Conference sponsond by AHfor Penons with AIDS: 1983-1988. Paper presented
CPR and HRSA, U.S. Public Heahh Seniee, Miami,
at the Amcricaa Economics Astodatioo Annual
Pla., December S.
Meeting, New Orieaas, La., January 5.
Hay. J. W., and F. A. Wolak. 1990. Bootstrapping HIV/ Scitovsky, A. A.. M. Oiae, aad P. R. Lee. 1986. Medical
Care Costs of AIDS Patieats ia San Frandteo. JourAIDS Pnqection Models: Back Calculation with Linnal of the American Medical Association 256(22):
ear laeauality-Coastnined Rcfreaiioa. Hoover Insti3103-3106.
tuttoa. WofUng Paper, No. B-90-3.
HeUinier, F. J. 1991. Forwutiag the Medical Care Cosu Smith. D. 1991. Tnatueat Stnte^es: iBtarview with
Paula Spani. AIDS Treatment News, no. 119 (Januof the HIV Epidemic: 1991-1994. Inquiry W):213ary II): 1-3.
225.
Hiehle, O.. W. T. MaxfleM. and K. W. Kizer. 1990. Volbsrdlng, P.. S. W. Lagaket, M. A. Koch, et tl. 1990.
Zidovudine in Asymptomadc Human ImmunodeAMedi-Cal Studies In AIDS Demoyaphlcs and Expenciency Virus latetion. New England Journal tfMtdditures for Persons with AIDS. 1980-1989. California
icint 322(14):Ml-9«9.
Depanment of Health Service*, Sacnmento.
<
: Lemp (San
Health). December 16,
>e e Blum (Office ef
tv
York City Department
nd HRSA. U.S. Public
December 6.
. IncubaUon Period of
rt 238(March I6);251Future Trends
United Sutei. Scitnct
nntenl Nuuttion. W
la. AIDS/HIV SurrilllaisiAcaUoo Sytiem for
UBS raillanceCM
i Adults, October » .
ev
wdPrq|ect«d
'c
October 31365
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
J_
�Tab 7.
This Section Includes:
AIDS Action Council- Health Care Reform
Statement of P r i n c i p l e s
�Health Care Reform Statement of Principles
The following are the criteria against which the HTV/AIDS
community will measure the various proposals for health care reform.
HTV/AIDS poses the same challenges to the health care system as
any chronic disabling condition. The reformed health care system
must be able to meet the needs of all Americans - and all Americans
are atriskof becoming disabled, whether the cause of their disability
is HIV, Alzheimer's, stroke, multiple sclerosis, or any other
condition.
To be acceptable, any proposed administrative structure must meet all
of the following criteria:
UNIVERSAL COVERAGE
Coverage should be universal and uniform. All U.S. residents should
be covered by the health plan and receive the same set of benefits.
Whether or not coverage and benefits are tied to employment; the
method for accessing health care and the scope of benefits should be
the same regardless of employment status.
COMPREHENSIVE COVERAGE
/S-5
Connecticut Are /VIP
Suite "00
Washington DC
20009
Fax 202 986 1345
Tel 202 986 1300
All medically necessary health care should be part of any benefits
package, including coverage for in-patient and out-patient primary
health care services, including full gynecological services, preventive
health services such as prenatal and well-baby care, prescription
drugs (including off-label use of approved drugs and ancillary and
actual costs of experimentaljreatments), hospice care, long- and
short-term home and community-based services, substance abuse
treatment, mental health services, outpatient rehabilitation services,
dental care, vision care, and the healing arts. Case management
should be provided to link people with appropriate health and social
services.
�AIDS Action Council health care reform statement of principles/Page 2
FREEDOM OF CHOICE
Individuals should befreeto choose their own health care provider. Individuals
should not be required to join a plan or delivery system if the individual does
not believe that program will meet his or her needs. In addition, individuals
should be allowed to change providers if they are not satisfied with the care
provided. This is particularly important for people with HTV infection, given
that a relatively small number of physicians are experienced in treating people
with HIV and there is no credentialing process that identifies those qualified or
experienced with HIV disease.
NO FINANCIAL BARRIERS TO CARE
Payment for coverage should be based on ability to pay, not based on a flat
premium schedule. Out-of-pocket payments should also be based on income. If
there is a system of copayments or deductibles, it should be designed so as to
prevent sick people from being disproportionately burdened. (For example, if
copayments are based on a per visit or a per prescription basis, sick people will
pay a larger portion of their income for health care coverage.)
GOVERNMENT RESPONSIBILITY
The federal government must exercise its authority to assure that health care is
delivered equitably and cost effectively through the following:
• Global budgeting, which sets targets for health care spending by sector
and is the most effective way of assuring appropriate distribution of
scarce health care resources while reining in inflation in the health care
field.
^
• Undertaking initiatives to redistribute the availability of health care
(especially health care practitioners) to underserved areas and populations,
including appropriate education andfinancialincentives to increase the
number of primary care providers.
• Guaranteeing the involvement of consumers in all aspects of
decisionmaking associated with any new health care system.
�AIDS Action Council health care reform statement of principles/Page 3
• Guaranteeing the confidentiality of all aspects of patient involvement
with the health care system.
• Achieving quality assurance through establishment of federal standards
of care and minimum standards for quality assurance methods.
• Requiring nondiscrimination in accessing the health care insurance
system, whether demographic or based on preexisting medical conditions.
• Assuring that appropriate care is given to all Americans, meaning that
the appropriate services are given for the appropriate level of need.
• The government must assure access to care under existing mechanisms
during a transition to the new health care system.
• The government must not "de-institutionalize" the traditional public
health programs even after a new system of health care delivery is
established.
• The government must assure full disclosure to and appropriate education
of all participants regarding access and options within the new system.
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
8
�Tab 8.
This Section Includes:
Lambda Legal Defense and Action Fund,
Inc. -Health Care Reform: Lessons From
the HIV Epidemic
�HEALTH
CARE
REFORM:
LESSONS
FROM T H
H IV
E P I D EM I
BY
M I C H A E L
T.
.../
Lambda
Legal
Defense
and Education
Fund Inc
I S B E l l
�— Chapter Eight
Lessons from the HTV Epidemic Principles forHealth Care Reform
j
1
Reform of the nation's health care system must be premised on the maxim that meaningful access to health
care is a basic human right. The experience of people living with HIV suggests six principles for ensuring each
person's enjoyment of this fundamental right.
(1)
The federal government should ensure
universal access in the U.S. and its territories
to portable, non-discriminatory medical
coverage in a single-tier health care system.
Universal coverage would improve health care access for the presently uninsured. It would facilitate cost
containment by simplifying health care delivery and by reducing opportunities for cost-shiftbg. Universal coverage
would also de-link the quality of health care from an individual's employment or family status. Portability — that is,
the freedom to take one's health benefits, unchanged, to the next job - would bring an end to the 'job lock' that
afflicts millions of Americans and would eliminate senseless, medically injurious waiting periods for coverage of
preexisting conditions.
For people with HIV, universal coverage in the U.S. and its territories would improve access to medically
necessary early intervenlion services and facilitate HIV prevention activities. Non-discriminatory, portable coverage
would bring an end to AIDS-specific benefit caps, discriminatory .redlining by commercial insurers and HMOs, HIV
antibody screening of insurance applicants, humiliating and injurious Medicaid 'spend-down^ provisions, arbitrary
waiting periods such as the 29-month lag in Medicare eligibility, state-by-state variations in the percentage of people
with AIDS who have medical coverage, bureaucratic delays in eligibility for medical services, and restrictions on
federal health care support to people living in Puerto Rico and other U.S. territories. By replacing the current
hodgepodge of fragmented delivery sources, universal coverage would permit patients with HIV infection to develop
the continuum of care recommended by medical authorities.
Universal coverage would ease Americans' anxiety that illness will lead tofinancialcatastrophe, a fear
which has proven chillingly accurate in the case of HIV disease. Under universal coverage, consumers with chronic
health needs would no longer need to resort to desperate measures — such as the purchase of "living benefits" or
complex Medicaidfinancialplanning - in order to stave off economic disaster.
Universal coverage would have economic benefits. By eliminating the link between the health profile of a
company's workers and the cost of the firm's premiums, universal coverage would discourage employment
discriminatioD against able-bodied persons who have or are at risk of developing a chronic illness. Portability of
health coverage would improve the mobility of U.S. workers and thereby enhance economic efficiency. By calming
Americans'financialfears regarding the future viability of their medical coverage, health care reform would shore
up consumer confidence and contribute to more robust economic growth.
Although even the most progressive health care systems have two or more tiers, the U.S. must attempt to
the greatest extent possible to care for all patients within a single system. A single-tier system of health care
delivery would improve medical care for the poor, who, as aresultof institutionalized medical apartheid, presently
depend on overcrowded public clinics or the emergency rooms of public hospitals. Reliance on a single system of
health care delivery would improve the quality of medical care provided to women, people of color, children, recent
immigrants, rural Americans and others who are disenfranchised by the multi-tier health care system. Creation of a
single-tier system would better protect each person's access to care by enhancing the country's collective political
investment in the system's integrity.
167
�People with chronic diseases, such as HIV infection, should have access to mainstream care. Every form
of health care delivery, even the most specialized, must be carefully placed within a uniury medical system to avoid
the stigmatizatioo and discrimination inherent in a segregated approach.
(2)
The federal government must guarantee a comprehensive
package of benefits that promotes good health,
emphasizes primary care, and leads to the most
efficient use of health care expenditures.
Meaningful health care reform would ensure each person's access to a minimum set of medical benefits.
This basic benefit package should be immunized as much as possible from political or fiscal pressures.
The basic benefits package should be rich rather than 'bare-bones. * Failure to ensure a full set of benefits
would inevitably lead to the development of a multi-tier system of health care delivery, with competing supplemental
insurance packages vying to appeal to the more affluent. A multi-tier system would encourage cost-shifting, risk
avoidance, and other subtler forms of discrimination in the availability of medical services.
In developing a basic benefits package, the federal government should attempt to place the health care
system on a sound footing. The package should emphasize primary care and preventive medicine. It should include
sufficientflexibilityto address the special needs of patients with chronic diseases such as HIV infection.
In particular, policy makers must transcend the present system's tendency to impose short-sighted,
counterproductive, arbitrary limitations on services in a misguided effort to save money. The U.S. can no longer
skimp on preventive care, long-term care, outpatient prescription drug coverage and other services that promote
good health and reduce medical expenditures in the long run. / • • ' y
Universal availability of a broad menu of services would permit the development of a rational contiuum of
care and would encourage coordinated case management, a strategy shown to save money and improve health care
delivery. A rich minimum benefits package would eliminate physicians' frequent need to tailor clinical judgments to
each patient's individual health coverage.
The record during the HIV epidemic demonstrates the need to offer a full set of minimum benefits.
Notwithstanding evidence that long-term care reduces AIDS treatment costs by as much as 12 percent, few patients
with HIV have access to such services. Many people with HIV could benefit from home health care, which is often
less costly than institutionalization, yet few third-party payers cover such services. Indigent patients progress to
AIDS much more quickly and often require costly hospitalization because they are deprived of early intervention
services or limited in their access to medically necessary prescription drugs. These and other irrational, injurious
and inefficient restrictions on health care access must be eliminsited.
In recent years, economists have advised that U.S. corporations must look beyond the next quarter's
earnings if they are to compete in a global economy. Similarly, the U.S. cannot expect to improve the public health
or to win the long-term battle against runaway health care costs unless it foregoes short-term false economies and
instead offers Americans the medical services they need.
Drawing on these principles, the basic benefits package must include but not be limited to:
-
in-patient hospital services;
—
out-patient hospital services;
physician, nursing and clinic services;
m
168
�m
*
-
UbonUory and x-ray services;
—
prescription drug coverage, Jsptfa in-patient and outpatient, with -flexible rules permitting coverage of
unapproved treatments in patients with lifethreatening diseases'*;
}
4
|
!
i
coverage for medical devices or substances required
in the administration of other medical services;
—
broad coverage for nutritional supports, including
oral supplements, tube-delivered supplements,
appetite enhancers, and total parenteral nutrition;
—
coverage on demand for a nnge of alcohol and
drug treatment options;
—
a variety of long-term care options, including
nursing homes, hospices, supportive housing or
other intermediate facilities addressing the specific
needs of patients with HIV or other chronic
conditions;
—
home and community based care, including
.coverage for personal care services; ..... v ' ,
—
private duty nursing services;
—
dental services;
—
accessible case management, if the patient desires
it;
—
physical therapy and related services;
—
services provided or prescribed by licensed
practitioners of the healing arts;
—
mental health services;
—
and all other services set forth in the Medicaid Act.
|
The federa] government must ensure meaningful access to each of these benefits, creating efficient referral
mechanisms, where necessary. To the extent a reformed U.S. health care system relies in part on managed care.
At the very lea«, rcimburwrnenlforunipprovedihcnpie* toed io tbe truunca of Ufc-UinMiciiijif eooditioairiwuldbe permi aed
when they utitfy eriierit icl forth in the Final Repon of Ihe Nitiona! CommiQcc to Review Currem Procedure* for Approval of New Drugs
for Cancer and AIDS. £ce Uuftu Comm'n, wpn n. 860. A preferable approach would prewme Ihe propriety of any therapy preicribed
by • physician for a pitients ufferinf from a life-threalening disease when Ute cflicacy of any such treatment has been documemed in *
peer-reviewed acbolaiiy jounul or when Ihe Ihenpy otherwise enjoys icatooable tcienciftc suppon.
169
�the govemmeot must ensure that each HMO, PPO,-or other managed care plan offers the full range of required
services.
(3)
Health care must be progressively financed.
Health care financing must be made fair and equitable. The federal government must guarantee to each
American,regardlessof his or her employment or socioeconomic sums, a standard, generous package of medical
benefits. Those most able to pay should bear the primary economic burden associated with universal access to care.
Out-of-pocket costs should be eliminated entirely; in the alternative, such costs should be carefully geared to each
patient's ability to pay orrequiredonly for procedures of diminished medical necessity.
Citing this country's disastrous experience with cost-unconscious demand for medical services, many health
care economists assert that costs can only be constrained if health care consumers are forced to bear out-of-pocket
costs in the form of deductibles and copayments. Yet the countries with the best trackrecordsin containing health
care expenditures, such as Germany or Canada, do notrelyheavily on out-of-pocket payments by consumers.
1461
Moreover, the long-term impact of reliance on out-of-pocket costs may be inconsistent with other health
care reform objectives. Studies suggest, for instance, that out-of-pocket costs frequently deter people from obtaining
vaccinations or other forms of preventive care that promote efficiency by preserving the patient's health. Similarly,
as the experience of people with HIV demonstrates, out-of-pocket costs frequently encourage patients with chronic
diseases to forego care required to forestall or prevent illness.
(4)
The federal government must retain
central authority to contain costs,
to correct imbalances in resource
allocations, to promote disease
prevention, and to ensure meaningful
access to medical care.
As the HIV epidemic makes plain, noreformin health care financing can by itself ensure meaningful access
to care. Numerous barriers, which frequently have little to do with health carefinancing,preclude access to proper
care for millions of Americans. The federal government must correct the deficiencies that often render access
illusory, even for patients with health coverage.
Tbe federal government must contain health care spending through enforcement of global budgets. Without
effective containment of health care spending, tbe U.S. cannot ensure meaningful access to care, strategically plan
future health care spending, or compete in the international economy.
As this chapter later explains, the health care system must be structured to promote efficiency. Little
evidence suggests, however, that free market economics alone can rein in runaway health care expenditures.
Moreover, unrestrained profitability would be inconsistent with the principle that health care is a fundamental human
right.
The federal government must therefore retain authority to control the growth in health care costs. The
government must limit physician fees, hospital fees, drug prices and other health care expenses to ensure that costs
do not grow faster than therateof inflation. The federa] government must establish and enforce global health care
budgets.
1
gee N.Y. Tunei, Jan. 23, 1993.
170
�Tbe government must ensure that fees remain sufficiently high to encourage both innovation and the
delivery of high-quality medical services. With respect to physician fees, Canada appears to have struck a
responsible balance; even with regulated fees, interest in entering tbe medical profession appears higher among
young Canadians than among their counterparts in the U.S. Likewise, the government must ensure that drug prices,
which are rising faster than other medical costs, increase at more reasonable rates, while preserving sufficient
profitability to encourage innovation in the pharmaceutical industry.
Although price controls are certain to generate intense opposition within tbe private sector, health care
reform presents important political opportunities, as well. Notwithstanding the medical profession's historic
opposition to regulation of physician fees, the federal government ought to be able to strike a bargain with U.S.
physicians by guaranteeing a reduction both in administrative paperwork and in outside interference in clinical
decision-making. Similarly, although the pharmaceutical industry has long been hostile to tbe mere suggestion of
price controls on prescription drugs, universal coverage of prescription drugs would drastically expand the industry's
market in the U.S.
The federal government must ensure a proper allocation of health care resources. While universal coverage
within a unified system should reduce the economic incentives that cause the San Francisco Bay area to enjoy 70
MRI scanners while other communities have none, the present imbalance in resource allocation results only partially
from tbe sources offinancing.Moreover, even iffinancingreform could by itself lead to effective allocation of
medical resources, it would take decades to correct the existing maldistribution. Government must take additional
steps to ensure a proper allocation of resources.
With respect to the medical workforce, the federal government must reinvest in scholarship and loan
forgiveness programs to encourage medical graduates t^/tocate in underserved areas. Through financial support and
promulgation of strict accreditation criteria, the federal government must also ensure that medical schools provide
sufficient training in primary care.
Expensive medical technology, hospital beds, and other medical services invite consumption by their very
existence. Resource gluts encourage the delivery of unnecessary care. Through data-driven technology assessment
and strategic planning, the federal government must ensure that medical services are rationally dispersed throughout
the U.S.' *
4
Bv collecting accurate date, the federal government must ensure that the health care delivered in the U.S. is
of the highest quality. The U.S. must invest in comprehensive medical outcomes research. Some frequently
oversell the accuracy and utility of existing outcomes data, as evinced by the Sute of Oregon's absurd claim to
possess the expertise required lo priontize every existing medical procedure. The U.S. must devote substantial sums
to improve on such efforts.
1463
While efforts must be taken to avoid unnecessary paperwork, health care providers, particularly HMOs and
other managed care plans for which meaningful data are largely unavailable, should be required to report medical
outcomes. Outcomes reponing should include demographic data to assist government in identifying patterns of
discrimination in the health care system or other barriers to health care access. For example, data collection
procedures must be calculated to identify undesirable patterns of care, such as the limited access of HIV-infected
women, IDUs and people of color to AZT or other standard treatments routinely delivered to more affluent, male,
white, non-IDU patients with HIV disease. Likewise, government must be capable of detecting excessive delays in
the delivery of medical care and differentia] rates in AIDS-related mortality among health care facilities. Once such
M
" §ee Ue & Lamm. Europe'! Medical Model. N.Y. Timei, Mar. 1, 1993.
See Teitimony of Dr. Phillip R. Ue, Chainnan. Pbyiician Piytnenl Review Comm'a, before the Senile Comm. oo Labor and
Human Reaourcct, Dec. 17,1992 CLeO (calling for develop mem of •andaidized national dau lyttcmi).
171
�deficiencies are ideotified, the federal govemment must exercise its authority over resource allocation and medical
education to ensure the elimination of such barriers to care.
Armed with accurate data, the federal government should devise meaningful systems for quality assurance.
Every consumer must have timely access to effective grievance procedures. Federal outcomes oversight efforts for
the protection of patients must substantially improve on those currently undertaken for Medicare and Medicaid by
HCFA, which aggressively monitors unnecessary expenditures but frequently glosses over quality-of-care
concerns.
1444
The U.S. must invest in targeted medical education. Too few physicians have sufficient training to provide
HIV-related care or to deliver the specialized services required by patieats with other chronic diseases. Similarly,
too few physicians understand the special medical needs of underserved communities such as the homeless, recent
immigrants, injection drug users, women, children, and adolescents. Working with the states, the federal
govemment must ensure that medical schools train students regarding these and other matters and that targeted
continuing education is available to correct deficiencies in clinical expertise.
The federal govemment must ensure that health care providers have access to information they need to
deliver high-quality medical care. Few mechanisms exist for the swift, effective distribution to health care providers
of important medical information. Due to the federal government's failure to publicize HIV-related risks faced by
many women, for example, too many medical practitioners have failed to recognize signs of HIV infection in their
female patients, leading to delays in diagnosis and premature morbidity and mortality.
f
Drawing on the results of outcomes research, the federal government should invest substantial sums in the
publication of clinical guidelines for a wide range of medical conditions. Guidelines must be updated regularly,
particularly for life-threatening conditions for which therapies and>(reattnent modalities change frequently, and
should be disseminated widely. Clinical guidelines should not serve as "cookbooks" to constrain clinical decisionmaking, particularly for life-threatening conditions which frequently rely on unapproved or innovative treatments,
but should target providers who may lack basic information. These guidelines should help ease the malpractice
crisis by setting forth a baseline standard of care.
/
Although the federal government has issued excellent infection control guidelines for the prevention of HIV
transmission in health care settings, it is evident that too few workers understand these rules orregularlyadhere to
them. Inadequate infection control training encourages workers to avoid patients with HIV and potentially places
workers at risk of contracting the vims. Government shouldrequirethat workers be properly trained in all aspects
of infection control and should suppon efforts to monitor workers' adherence torecommendedguidelines.
The federal government mustrebuildthe nation's public health infrastructure. The U.S. mustreinvestin
disease prevention, health education, disease surveillance, and public health leadership. Public health departments
should be depoliticized and have as their sole objective the protection of the public health.
jj
Every person in the U.S. must have meaningful access to HTV prevention information. Prevention efforts
must befrank,honest, culturally appropriate, and targeted to the communities most at risk. At-risk communities
should be integrally involved in the formulation, planning and implementation of prevention programs. Govemment
financial support for HTV prevention must increase significantly.
w
*' Federal ovenight of medical outcome! and iwift murvcttion lo correct pinenu of improper canritouldnot be -"fifirtrrt with
utilization review of elinical dccuioni by individual phyaiciam, whichritouldbe drutically reduced in areformedhcahh cere ty«em. Ai
deicribcd later in thii chapter, t reformed lyctm ihould maximize the freedom of doc ion and patieitfs, in coaailiation with one another, lo
decide on their own the moct appropriau count of Ihenpy.
NooelheleM, coniumen ahould be protected from malpractice, much of which the pietem syuem taib lo delect. Convfchcnsive
outcomei reporting would aaaut in Ihe identification of tacoopetcnt providcra who pose a danger io patieas.
172
�••
'
vV |
Every aspect of the public health response to the HTV epidemic - including educational efforts and
surveillance strategies - should be designed to encourage people at risk to seek appropriate medical care. Through
education and energetic public health leadership; the federal govemment should encourage Americans to exhibit
care, compassion and commitment toward people with HTV. The U.S. must also become the leader in the global
fight against AIDS.
Tuberculosis requires a similar public health response. Government must ensure meaningful access to a
range of TB treatment options. Government must aggressively monitor the epidemic and intervene quickly to close
gaps in access to needed services. The U.S. must invest in TB prevention strategies, ensuring proper infection
control in health care settings, correctional facilities, and other congregate living situations. The U.S. must address
urban decay, homelessness, drug addiction and other social problems which have facilitated the resurgence of TB.
The federal govemment must sumo out discrimination in the health care setting. Discriminatory denials of
carer, groundless referrals, and breaches of confidentiality impede effective integration of underserved patients into
the health care system.
In most circumstances, federal law presently prohibits discrimination in the delivery of health care services
based on a patient'sraceor ethnicity, gender, disability or perceived disability. Yet widespread discrimination in
the health care setting continues to afflict countless patients due to inadequate enforcement of existing sututory
protections.
The federal government should energetically investigate allegations of HTV-related discrimination in the
provision of health care services. Where appropriate, tbe federa] government should terminate funding to facilities
that engage in illegal discrimination. Widely publicized federal enforcement actions have previously improved
medical practitioners' compliance with existing anti-discrijmhation laws.
" 1445
No federal law protects gay men and lesbians from discrimination in the provision of medical services. The
President should work energetically to ensure passage of national gay civil rights legislation.
The U.S. should take steps to ensure that the confidentiality of medical information is broadly protected.
At present, no federal sutute protects the confidentiality of medical records. The President should support
enactment of such a law, which should, inter alia, prohibit the unauthorized disclosure of HIV-related information,
provide for a private right of action for violations, and permit the recovery of compensatory and punitive damages.
In addition, the U.S. must take the following steps to ensure the delivery of appropriate medical care:
—
Tbe federa] government must invest in strategies
to combat fraud in the health care system,
with particular attention to abuses in die
home health care industry.
—
The U.S. should prohibit physician self-referral.
—
Although efforts must be made to provide mainstream
services to all consumers, tbe U.S. should, at
least in the short nm, continue to suppon
M U
Sec Burdin v. U.S. Dee'i of Hetlih snd Humin Servi.. 934 F.2d 1362 (5th Cir. 1991) (relying on •mi-ducriniiaition provuions in
Medicaid Ac I to penalize emergency room phyiician who Umnsfcrred patiem, wbo had an emergency medical condiiion, lo another hospital
for a non-medical reason); In rt Wetichcster Medical Center, supra n. 1265 (ordering the termination of federal funding to medical center
that refused to hire KIV-infccied pharmacist, in vioUtion of Section 504 of the federal Rehabilitauoa Act).
173
�facilities caring for underserved populations.
Such facilities include community and migrant
clinics, IHS facilities, school-based clinics,
public HIV testing sites, and clinics serving
homeless persons. In addition, Congress should
fully fund the Ryan White CARE Act to support
the numerous community-based endeavors required by
the HTV epidemic.
The U.S. must guarantee the availability
of a wide range of family planning and
reproductive health services.
The U.S. must ensure the delivery of standard
medical care to every prison inmate, through
increased funding, improved staff training,
ind meaningful quality assurance.
The U.S. must significantly increase funding
for drug and alcohol addiction treatment
and for research into improved approaches
to addiction treatment.
The federal government must substantially
./
increase funding for biomedical research,
/
in general, and regarding women's health needs,' in particular.
•
The federal govemment must support efforts to
make health care delivery culturally appropriate
for underserved communities, such as recent
immigrants. The U.S. should offer financial
support for clinics serving such groups. In
particular, the U.S. must support efforts lo
improve the linguistic competence of health care
providers and case managers.
The U.S. must develop national housing, child care,
and transportation policies that facilitate access
to health care.
The U.S. must work with states and localities
to improve the foster care, family law and
educational systems to address the needs of
children orphaned by the AIDS epidemic.
The U.S. must invest in the development of
a comprehensive health care infrastructure in
Puerto Rico and other U.S. territories.
174
�-
Toremoveinhumane deterrents to care, the U.S.
must cod the ban on immigration to the U.S. by
people infected with HIV.- -
(5)
The health care system should be efficient.
Although global budgets constitute an important strategy for keeping tbe lid on health care spending, health
care delivery must itself be efficient. Otherwise, administrative waste will consume limited health care resources,
causing the quality of medical care provided within globalrestrictionsto decline over time.
Health care financing must be drastically simplified. Tbe number of third-party payers should decline
substantially. Claims forms should be standardized, and billing should occur electronically. Risk-based
underwriting must be prohibited; providers found to have discriminated against perceived poor risks should be
subject to severe penalties.
Physicians must be freed to do what they were meant to do, i.e., practice medicine. Accordingly,
administrative paperwork associated with medical practice should be limited to the greatest extent possible.
Utilization review of clinical decisions should be reduced or eliminated altogether. Enforcement of
reasonable global budgets would, when combined with aggressive outcomesreportingand swift dissemination of
practice guidelines, betterreduceunnecessary care than the wasteful, time-consumingreviewby non-medical
personnel of physicians' decisions.
(6)
The health care system should be accountable
to consumers and should maximize
the freedom of consumers and providers.
The health care system should/preserve to the greatest extent possible the consumer's freedom to choose his
or her provider. This freedom is especially critical for patients with life-threatening diseases such as HIV infection.
Within enforceable global health care budgets, physicians should have maximum freedom to practice
medicine. Under health care reform, collaborative decisions of individual patients and their physicians should
become the touchstone for "medical necessity."
Health care providers must be accountable to consumers. Having collected comprehensive outcomes dau,
the government must implement effective quality assurance protocols. Patients, particularly those subject to
managed care, should havereadyaccess to swift, fair due process reviews of medical decisions.
175
�I
Chapter Nine
Managed Competition - Does It Meet the Test?
— - ••
i
i
Health care economists, public policy analysts and govemmeot officials have in recent years hotly debated
the merits of various health care reform models. In the last year, "managed competition" has emerged as a likely
model for health care reform.
Using the six reform principles derived from the HIV epidemic as a backdrop, this chapter analyzes the
concept of managed competition. This chapter describes managed competition, studies tbe ability of managed
competition to satisfy the six key health care reform principles, describes additional impediments to implementation
of managed competition, and sets forth strategies for improving managed competition.
A.
Managed Competition: A Primer
One of the reasons why managed competition currently enjoys such broad political support is that
proponents of the approach define the term according to their individual tastes. Nonetheless, certain
generalizations are possible regarding managed competition.
1467
'Managed competition plans . . . . share a few basic assumptions. One is that, left to themselves, health
insurance and service markets are inefficient. Insurers and providers do not naturally compete on tbe basis of
quality, economy, and consumer satisfaction. To create a competitive market requires an activist regulatoryrolein
which the objective of regulation is not protecting tbe financial solvency of insurers but promoting competition
among health plans." ?
/
4
?
Under managed competition, health insurance purchasing cooperatives ("HIPCs") would offer consumers a
choice of coverage from a limited menu of giant regional health plans. While some proposals would require
universal participation in these HIPCs, many would permit larger companies to opt out by purchasing coverage for
their workers. Although tbe menu of health plans could theoretically include a variety of delivery options, including
traditional fee-for-service care' , virtually all observers believe that consumers would principally choose from
among various managed care plans, most of them owned by large insurance companies.
I4W
410
1471
**" "A year ago, managed compctiiion did DM even figure in mon public diacuaaioni of health care reform. As Ute news media
preaeated it, the menu of reform had three major allematives: a Canidiaa^tyle, eingle-paycr sysum of national health insurance; Bush's
plan to reform the currenl insurance market and give tax credits for limited coverage lo the poor, and 'play-or-pay,' Ihe proposal embraced
by the Senate Democratic leadenhip torequireemployen lo insure their worker* or pay into a public insurance program." SUIT. Healthy
Compromise: Univeraal Covtrsee and Manseed Competition Under a Can. The American Prospect 44 (Winter 1993) CSlarr").
Jee Salisbury. Beware of 'Managed Competition" Convent. L.A. Timea, Jan. 13, 1993.
Cabel & Rice, ££22 n. 167. $ct Enthoven & Kronick 1. supn n. 41 at 2534-35.
Enthoven & Kronick I, supra n. 41 at 2534.
•«* Starr, EJEB ». 1466.
gee Reiman. gigg n. 331 at 134. See also Enlhoven A Kronick L lupra n. 41 al 2534-35.
177
�A national health board would stipulate the minimum benefits package which each health plan would be
required to offer. Managed competition would incospprate the standard elements of small group market reform,
including portability of coverage and a ban on medical underwriting.
In any given region, health plans would presumably compete in aregulatedmarket by attempting to offer
superior services at lower prices." By pooling consumers' purchasing power, HIPCs would theoretically
retain
leverage to force health plans to function in consumers' interests. To enable consumers to distinguish between
plans, the federal govemment would invest substantia] sums in medical outcomesresearchand make outcomes dau
readily available.
75
1471
1474
For purposes of cost containment, consumers would probably berequiredto make limited copayments.
According to standard managed competition proposals, tax subsidies for health benefits would be limited to the value
of the standard benefits package; employers or consumers desiring more extensive benefits would be required to
purchase them with taxable dollars.
1475
Although managed competition proponents agree that the poor and the uninsured should be integrated into
the same health care system that serves large employer groups and HIPCs, various managed competition plans
envision different means and timeubles for achieving this result. Under the managed competition proposal advanced
by conservative congressional Democrats , health coverage for the indigent would be de-linked from welfare
programs; persons living under the official poverty line would receive health care free of charge, while those with
incomes between 100-200 percent of poverty would obuin subsidized care.
1474
Recently, policy makers have expressed interest in wedding the market-driven approach of managed
competition with the seemingly contradictory notion of goveramenHinposed global health care budgets. "
According to such plans, global budgets would be esublished by a national health board, presumably the same one
which identified the minimum benefit package. Enforcement mechanisms for these global budgetsremainunclear.
14
B.
Managed Competition and the Six Criteria
for Meaningful Health Care Reform
In light of the six criteria for health carereformset forth in chapter eight, does managed competition make
the grade? An analysis of managed competition highlights its selling poults, as well as potential trouble spots.
"n Enthoven & Kronick I, wipri n. 41 at 2233.
""Id. at 2534-45.
Id. at 2535.
gee Enthoven & Kronick I, mpn n. 41 tt 2534. Prei* re pom ugfcH Hut the Clinton admininration viewa this etc menl of
managed competition aa politicallyriakyand potentially diq>cnaablc. Managed competition Ihcoriau contend, however, thai reviaiont to Ihe
ux code ate central to their proposed approach.
"* §ee The Managed Competition Act of 1992. H.R. 5936 (102d Cong.).
W l 1 1 S l n e l
Journal, Feb. 16, 1993. See Teaumony of Dr. Paul Starr before the Senate Comm. oo Labor and Human Reaourcet.
Dec. 17. 1992.
178
.
I
j
\
�(1)
Would managed competition facilitate
uni?ersal, portable, non-discriminatory
coTerage in a single-tier biealtb care system?
Managed competition wouldreducethe number of uninsured by expanding coverage to all workers
employed by small businesses. Although the exact means for covering the poor differs from one managed
competition proposal to the next, managed competition would likely improve the quality and availability of health
care for the indigent. Managed competition's inclusion of standard small group marketreformswould presumably
ensure portability of coverage andreduceopportunities for discrimination against the chronically ill.
Although proponents assert that managed competition would eventually lead to equitable coverage within a
single health care system, questions remain regarding the timetable for integration of the poor into managed
competition. During the presidential campaign, President Clinton suggested that access to care should be expanded
for the uninsured as managed competition produced savings. Not only would this apparent savings-before-access
approach necessarily delay therealizationof universal coverage, but, as described below, many economists express
skepticismregardingmanaged competition's ability to slow health spending.
Without swift phase-in, the poor would likelyremainfor some time in an expanded Medicaid system, with
attendant barriers to primary care, utilizationrestrictionson medically necessary services, and inadequate provider
participation. Ultimately, if managed competition were to fail to produce savings, this expanded Medicaid system
could become permanent, preserving the traditional second-class treatment of the poor.
The precise rules governing HIPCs would further determine whether the health care system would
unitary or have multiple tiers. If the employee cut-off for business participation in HIPCs were low, for exam
HIPCs would primarily attract small groups, the poor and the uninsured, limiting HIPCs' leverage to comp
the best coverage.
>tn
The comprehensiveness of the minimum benefits package could also affect the equity of health care
delivery. If the benefits package were limited in scope, health plans would compete for the wealthiest customers by
offering missing services or other amenities. Unless steps were taken to ensure equitable access to all forms of
health care delivery, multiple tiers of health delivery would quickly develop, with tbe poor locked into poor-only
HMOs, the middle classrelyingon the standard package provided by managed care plans, and more affluent patients
enjoying gold-plate services offered in exclusive care settings.
Due in large measure to itsrelianceon managed care, managed competition might not ensure nondiscriminatory treatment of people with disabilities or other chronic medical needs. According to federal authorities,
private managed care systems presently engage in more extensive medical underwriting than commercial insurers.
Although HMOs would presumably be constrained by various ground rules under any system of managed
competition, these entities would likely be owned and operated by large commercial insurance companies, which
have accumulated a deplorablerecordin serving the chronically ill.
Under managed competition, as with the status quo, profit would remain the name of the game for health
care providers. Patients with chronic diseasesrepresenta clear threat to the profitability of HMOs. Tbe same
incentives to undertreat which presently inhere in HMOs would continue to exist under managed competition.
In light of managed competition's reliance on managed care, mere promulgation of a standard minimum
benefits package could not ensure equal access to health care for the chronically ill. HMOs that desired to curtail
service to patients with chronic health requirements would need merely employ a limited range of specialists, ay
"» Inadcquaic buiiaen panicipaiion ia ihc HIPC* would alio impede job mobility, a* a job cha^e would cocoinuc to mean dump tion
in • worker'* hcahh coverage.
179
�b arrangingreferrals,relyon gatekeepers with limited medical expertise, or adhere to other, subtler practices
calculated to discourage persons with chronic diseases from enrolling in the plan.
Quality assurance would be difficult, if not impossible, in these plans. As Relman observes, "[N]o
practical system of monitoring by public agencies could be expected to ferret out all the subtle ways in which
managed-care organizations, controlled by third parties determined to show a profit, might stint on services. * *
147
HMOs offering high-quality service to patients with chronic diseases would likely suffer financially by
attracting poorer risks. Addressing this concern, various managed care proponents have proposed to risk-rate
compensation to HMOs to ease the financial strain on plans with disproportionate numbers of chronically ill
enrollees would not be economically disadvantaged.
While risk-based compensation to the plans would
certainly be preferable to leaving unchecked the financial incentives at the heart of managed care, introduction of a
risk-based yardstick could encourage segregation of people with chronic health conditions in a limited number of
plans, with the stigmatization and discrimmation inherent in any such "separate but equal* approach. Moreover,
there is no assurance that such risk projections would be accurate; indeed, cost projections during the HTV epidemic
have provenremarkablyinexact. Withoutreasonablyaccurate risk projections, HMOs might seize on these
compensation formulae as merely another opportunity to profiteer.
1-0
Presently, HMOs oftenredlineentire neighborhood andregionsto avoid covering perceived high risks.
While proponents of managed competition have emphasized that competing health plans would be required to offer
the same set of minimum benefits, it is equally imperative that competing plans serve identical coverage areas.
Otherwise, HMOs might, through careful selection of coverage districts, engage in subtle, yet highly discriminatory,
risk avoidance.
(2)
Would managed competition produce a
comprehensive minimum benefits package?
•/
It is impossible to predict which medical services a national health board would include in a minimum
benefits package. Skeptics fear, however, that fiscal limitations could prompt the board torestrictthe scope of the
package, which would not only lead to the development of a multi-tier system of health care delivery but could also
perpetuate the current de-emphasis of numerous cost-effective, medically necessary services. Critics further
question the advisability of limiting tax subsidies for health benefits, fearing that such a move would create longterm fiscal incentives to steer consumers toward bare-bones coverage.
If subject to a minimum benefits package, Medicaid beneficiaries might actually lose access to services on
which they currently depend. Although existing Medicaid eligibility rules are overly restrictive, and although
patients who depend on the program are often forced to obtain care in emergency rooms and overcrowded clinics,
Medicaid's benefit package typically includes a variety of services - such as dental care, outpatient prescription
drugs, case management, and mental health services - which are often missing from private plans. If the minimum
benefits package wererestrictiveratherthanrich,many Medicaid recipients could actually lose many of the benefits
they currently enjoy.
/•
Patients with HTV or other chronic diseases require arangeof services, including outpatient drug coverage,
preventive primary care, numerous medical specialties, a variety of long-term care options, and coverage for home
health care. Some of these might well be among thefirstservices deleted in a limited benefits package.
"* Relrnin, sm* n. 331 at 134.
Ma
° SSS. Temmooy of Dr. Sbodunni Sober before tbe Senile Comm. oo Labor and Human Reaourcca, Dec. 17.1992 fSotaer').
180
�Although chronically ill patients require numerous standard medical services, they frequently require
flexibility in their use of such services. Take outpatient prescription drugs, for example. Whereas inclusion of
prescription drugs in the benefit package' would expand the access of chronically ill Americans to medically
necessary services, the failure to ensureflexiblecoverage for unapproved drugs or for off-label applications would
withhold meaningful care from many patients with life-threateoing diseases.
Even if the minimum benefits package included therangeof services required by people with HIV or other
chronic diseases, managed care plans could nonetheless fail to ensure the availability within the plan of sufficient
HTV-related expertise or the full nnge ofrequisitemedical specialties. Even when specialty services were readily
available, gatekeepers who were incompetent or inadequately versed in the special medical needs of the chronically
ill could fail to recognize the need forreferral.Utilization review, compensation-based incentives to underserve,
inadequate grievance procedures, and other impediments inherent in existing systems of managed care would
presumably continue to afflict chronically ill patients under managed competition.
(3)
Would managed competition
be progressively financed?
Failure to promulgate a rich set of minimum benefits would lead to greater out-of-pocket costs, which
would inevitably burden the poor and middle class more than the rich. Under a limited benefits package, workers
employed by large U.S. corporations — who typically enjoyrathercomprehensive (if rapidly deteriorating) coverage
— might lose benefits they currently enjoy. In such circumstances, workers would be forced to buy such services
directly or purchase supplemental coverage.
• • . .- •
.
•/
Little evidence exists that out-of-pocket costs produce wiser, more efficient health care choices by
consumers. On the contrary, data persuasively suggest that deductibles and copayments deter consumers from
obtaining preventive services and other forms of primary care which promote good health andreducemedical
spending in the long nm.
/
(4)
Under managed competition, would govemment
retain central authority to contain costs
and make additional systemic corrections?
According to Bruce Vladeck, the President's nominee to head HCFA, "Thereasonso many providers
suppon managed competition is because they don't believe its cost containment mechanisms will really work, and
they're afraid the mechanisms in other proposals for health-carereformwould." " Skeptical of managed
competition's potential to slow the long-term growth in health care spending, many proponents of the approach
propose linking it to global health care budgets. How such budgets would be enforced, however, remains aitogethei
unclear.
Questions regarding the ability to control costs are particularly apt with respect to proposals to permit large
companies to opt out ofregionalHIPCs. Unless each HIPC dominated a given region, it is difficult to envision ho\
HIPCs could force providers to live within global health care budgets.
Managed competition could not by itself eradicate race-, gender-, or disability-based discrimination in the
provision of health services, nor would it assist inrebuildingthe nation's public health infrastructure. Whereas
universal coverage would ease many of the financial incentives which have facilitated the improper distribution of
health care resources, managed competition alone oould not ensure an adequate supply of primary care providers,
4
'N.Y. Tunes, Mir. 7, 1993.
181
�eliminate deficits in expertise regarding the medical needs of neglected communities, or entice young doctors to
locate in rural communities or in underserved inner-city areas,
i
i
;
-- .
..
Market economics cannot ensure that health care providers have swift access to the information they need to
provide state-of-the-art care for complex conditions such as HIV disease. Nor could managed competition ensure
the removal of cultural barriers to meaningful health care access for recent immigrants, or erase impediments
created by the lack of national housing, transportation and child care policies.
While experts widely agree on the need to invest in additional outcomes research, such research must elicit
the proper information in order to be meaningful. Consequently, substantial attention should be devoted to creative
means for collection of data regarding health care quality and .
(5)
Would managed competition be efficient?
Managed competition would marginally improve the efficiency of health care financing. The number of
insurers would likely decline under managed competition, and claims processing would presumably be simplified.
Yet it is doubtful that the efficiency associated with managed competition would produce massive savings.
Although managed competition would render obsolete the present chaotic system of health care delivery, it would
replace it with an experimental, complex administrative apparatus. According to experts, market reforms are not
likely to save substantial sums in the long run.'*
2
Managed care, a principal element of managed competition's cost control strategy, has, at best, a spotty
record in restraining health care spending. From an efficiency standpoint, many profitable HMOs currently
devote a greater share of each premium dollar to administrative overhead than competing indemnity plans. *
Mn
14
1
Because managed competition has never really been tried on a grand scale, proponents of the approach
cannot convincingly specify the number of health plans which each region would require to form a competitive
market. Without such data, proponents could not 'ensure genuine competition instead of covert collusion to
maintain prices." * Although global budgeting and price restraints might limit the inflationary effects of
collusion among health care providers, such savings would result from govemment interventionratherthan from the
intrinsic efficiency of managed competition.
4
(6)
3
Would health care delivery be accountability
to consumers under managed competition?
Under managed competition, more Americans would obtain care in HMOs or from other forms of managed
care, yet quality assurance is frequently lacking in such plans. In addition, many managed care plans often lack
adequate grievance procedures and frequently fail to ensure consumers' access to tbe full range of necessary
services.
gee Coaon, Clinton T.nken With Health Svretn Sutui Quo: Criiic. Seek lo Pick Aoin Miiured Compethion. 269 J. A.M.A. 1229
(Mar. 10, 1993) (cilinf Congrcnionil Budget Office sudy).
»•» gee Relman. wpn a. 331 al 134.
Phyiician't Weekly. Jan. 18, 1993.
M
J
* Relman. KEn n. 331 al 134.
182
�Numerous surveys confirm that Americans desire the freedom to choose their health care providers.
Managed competition, however, would steer most consumers toward delivery systems predicated on severe
restrictions on consumer choice. Loss of freedom of choice would be particularly problematic for patients with HIV
disease or other chronic conditions, for whom a personal relationship with a sensitive, qualified medical specialist is
essential.
Far from maTimiring physicians' freedom to practice medicine, managed competition would subject the vast
majority of U.S. doctors to the "intrusive surveillance and control methods [of managed care]* which have proven
immensely unpopular with the medical profession. " While some managed care plans have experimented with
strategies to expand the clinical freedom of participating physicians, trend-setting HMOs have adopted more, not
less, restrictive utilization review in recent years.
14
Physicians, like most other commercial actors in the health care system, bear their share of blame for the
present financing crisis. In the end, however, the medical profession must play a central role in health care reform.
The U.S. would appear to have little to gain by alienating the bulk of practicing physicians. * "[G]ood costeffective care . . . . is most likely to be achieved when properly motivated physicians working with patients are
given tbe authority to manage medical care without the intrusion of third parties but within dearly defined cost
limits. "
14
7
m
Not only must doctors be freed to deliver high-quality care to patients, but, under managed competition,
HIPCs should be structured to maximize their potential to protect consumers. Dr. Phillip Lee has observed, 'HIPCs
can play an active role in a community's health care delivery, targeting areas where inappropriate care is common,
initiating projects to improve the delivery of care, spurring other entities to initiate their own projects, and
monitoring the involvement of both fee-for-service andrnknagedrcareplans and their providers in these
projects.'
HIPCs would be unlikely to realize this potential unless steps were taken to ensure the
centrality of consumers to the daily operation and governance of these entities.
1-9
Presumably, ERISA's applicability to health plans would end under managed competition. Otherwise, selffunded employers that opted out of the HIPCs would retain potentially dangerous discretion to operate their selffunded health plans in ways contrary tti the interest of their workers.
Managed competition would depend on engaged, educated consumers who, when provided with necessary
incentives, would make rational health care choices. Standardizatiion of benefits offered by competing health plans
would be critical to empowerment of consumers. If health plans were permitted to offer a competing array of
amenities or additional services, bewildered consumers would have little means of comparing one plan with another.
I
j
I
j
Indeed, given managed competition's likely reliance on managed care, it is difficult to envision how
managed competition could possibly enable consumers — particularly those with chronic, complex health care needs
— to make rational, educated choices among competing plans. As Relman argues, the subtle avenues by which
HMOs might deprive consumers of necessary care would seem to be limitless. Even with the very best outcomes
data, devising a system to enable consumers, particularly those with complex health needs, to navigate sucb a
complicated terrain would be challenging, to say the least.
— Jd. at 134.
1
V'k* * R* *'. U f i Mike i Deal - Nerwining i S«nltment Between PhnicUn. and Society. 327 New En*. J. Med. 1312.1313
(Oct. 29, 1992) Cthe importance of a eompasaioaaic. committed, and qualified workforceof provide™ nnpUci a auoof aocial iotenat in Ihe
job aatiifactioo of phyatciani").
Id. at 134.
"o Lee, ESQ a. 1463.
.183
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
^
�Tab 9.
This Section Includes:
Smith, Mark. "Primary Care and HIV
Disease"
�COMMENTARY
Primary Care and HIV Disease
MARK D. SMITH. MD. MBA'
Nearly one million Americans are infected uitb tbe human
immunodeficiency virus (HIV). With tbe advent of increasingly effective therapy, including intervention early i n tbe
course o f infection, there will be a growing needfor physicians technically and attitudinally prepared to provide
p r i m a r y care f o r HIV-infected individuals. For any disease, however, determination of which physicians provide
tbe bulk o f care depends on severalfactors, including tbe
prevalence and cbronicity o f tbe disease, tbe complexity,
rate o f change, and toxicity of therapy, and tbe socioeconomic characteristics o f patients with tbe disease. General
internists will clearly constitute a large p a n of tbe pool o f
practitioners caring f o r HIV-infected patients, especially
i n tbe earUer stages o f infection. I t seems reasonable to
expect every general internist to be competent in f o u r
aspects o f HIV care: counseling about transmission and
prevention; tbe proper administration and interpretation
o f diagnostic tests; monitoring and care o f patients i n
early stages o f infection; and recognition of complications
o f advanced infectionf o r proper management or referral.
Academically based generalists will have a major role i n
research, teaching, and patient care i n tbe AIDS epidemic.
Tbe organization and delivery o f primary and specialty
care f o r HIV-infected people in f u t u r e years will continue
to evolve with changes in therapy a n d i n tbe demography o f
tbe epidemic. Key words: AIDS; HIV; academic general
internists; p r i m a r y care. J G E N INTER* M E D 1991;
6(suppl):S56-S62.
THIS ISSUE o( the Journal of General Internal Medicine
represents a significant step in the effort to meet one of
the most serious challenges facing the American medical care system: the provision of primary care for individuals infected with HIV.
When AIDS was first recognized in the United
States some 11 years ago, its manifestations were a few
unusual infections and cancers, usually cared for by a
handful of specialists. From this early stage, the epidemic has grown to encompass a wide array of conditions in a large pool of infected persons, and what was
once regarded as a terminal disease with a short life
expectancy is now clearly a chronic condition, albeit
one that is ultimately fatal. The capacity of AIDS or
infectious disease specialists to meet the needs of the
epidemic has been surpassed. Increasingly, therefore,
the question must be raised: who will provide primary
care for the nearly one million individuals in America
infected with HIV?
1
2
WHAT DETERMINES A
"PRIMARY CARE" DISEASE?
It has been suggested that care for all stages of HIV
disease should be within the scope of competence of
every generalist . As desirable as "primary care" might
be for all patients and all diseases, a great many diseases
are usually cared for principally by subspecialists or in
episodic fashion. In many cases the diagnosis, even if
made by a general internist or family physician, dictates
that the patient be transferred to a subspecialist; most
malignancies and advanced congestive heart failure are
examples of such diagnoses. At least six factors appear
to determine the attitudes of the medical profession,
the health care system, and the public about which
diseases are cared for by primary care physicians. They
are: the prevalence and the cbronicity of the disease;
the complexity of its diagnosis and management; the
rate at which the therapy changes, the toxicity of therapy: and the socioeconomic characteristics of the people who have the disease.
3
Prevalence
A high prevalence alone is not sufficient to lead to
primary care physicians' treating patients for a disease.
The United States had experienced, as of July 1990,
139,000 cases of AIDS (end-stage HIV disease)
throughout the ten years of the epidemic.'' In comparison, for example. 155.000 cases of colorectal cancer
are predicted for the United States in 1990 alone.' Yet,
while most physicians regard the prevention, detection, and preliminary diagnosis of colorectal cancer as
being within the purview of generalists, its management is usually left to subspecialists.
Chronicity
Chroniciry also plays a role in the determination of
provider. Long-term diseases such as hypertension and
anhritis are more apt to require long-term relationships
between provider and patient Acute, episodic disorders can be cared for by a patient's primary care provider if he or she has one. and are also more likely to be
treated by emergency rooms, "urgent care'' centers,
and subspecialists
Complexity
"The AIDS Service. The Johns Hopkins University. 1830 East
Monument Street, Suite 7400, Baltimore. Man land 21205
Address correspondence and reprint requesu- to Dr Smith
S56
Complexity also plays a role We expect the generalist to be able to manage hypertension and diabetes,
�JOURNAL OF GENERAL IKTEBNAL MEDICINE. Volume 6 (January/February
not only because these diseases are relatively common
in the United States, but also because their management
is relatively simple, requiring only a few therapeutic
agents that are usually well known to generalists. On
the other hand, diseases that require invasive procedures for diagnosis or management or that have complex therapeutic regimens (such as combination
chemotherapy or dialysis) are often left to specialists.
Rate of Therapy Changes
Another feature of "primary care" diseases is that
their therapy changes relatively slowly. For example,
although calcium channel blockers and ACE-inhibitor
therapy have become increasingly important in the
management of hypertension, their acceptance has
taken years. While medicine certainly continues to
make progress in the treatment of diabetes, congestive
heart failure, arthritis, and other "primary care" diseases, that progress is measured in years and decades,
not weeks and months.
Toxicity of Therapy
The toxicity of therapy is also an issue, particularly
when the therapy is one not familiar to many physicians. Thus, chemotherapy is generally left to oncologists, in pan because of its complexity and narrow therapeutic window.
Supplement). 1991
S57
TABLE 1
Clinical Subspecialties Involved in the Care of HIV-infected Adults
Frequent
Dermatology
Gastroenterology
Infectious diseases
Neurology
Oncology
Ophthalmology
Pulmonary medicine
Psychiatry
Less frequent
Cardiology
Endocrinology
Hematology
Nephrology
Obstetrics/gynecology
Rehabilitation medicine
Surgery
Clinical pharmacology
cialty). It is, rather, a combination of these qualities
with the nature of the system within which he or she
works. AIDS patients can be taken care of within any
number of settings; the arrangement of that practice
and the support available to practitioners will be key.
Nevertheless, a grasp of the general features that put
some diseases within the purview of generalists and
consign others to specialists is constructive when considering the future of primary care for HIV disease.
7
Demographics of People with the Disease
Last, patients with some conditions, such as tuberculosis, gonorrhea, and pregnancy with the "co-morbidity" of poverty, are cared for by certain doctors (or
doctors in certain types of health care facilities) not so
much because of the intellectual or procedural difficulty of their care, or because of the speed with which
therapy changes or that therapy's toxicity, but rather
primarily because of the demographics of the patients
who have that disease. The private medical system in
the United States continues to demonstrate certain
"market failures" — diseases whose care simply is not
sufficiently attractive to providers for economic and/or
social reasons; such care is largely left to sexually transmitted disease (STD) clinics, prenatal clinics, community health centers, and other public or quasi-public
providers. Some of these facilities render comprehensive primary care for many conditions; others, including TB and STD clinics, treat only a single disease.
All of these distinctions are, of course, relative;
often a combination of factors, rather than any single
one, is decisive. Whether specialists or generalists care
for a given disease also depends on whether the region
is urban or rural, in the Northeast or the deep South, a
staff-model HMO or solo private practice, and other
factors. It should be emphasized that this is not primarily a question of the intelligence, commitment, or
skill of the individual physician (of whatever spe6
WHAT IS PRIMARY CARE.
AND HOW DOES IT APPLY
TO HIV DISEASE?
Most definitions of "primary care" emphasize four
qualities that distinguish it from other types of medical
services: comprehensiveness, continuity, coordination, and attention to psychosocial and environmental
issues of care. Each of these has particular importance
in assessing the needs of HIV-infected patients.
8
Comprehensiveness
Comprehensiveness is a cardinal requirement of
care for HIV disease because the disease affects every
organ system in the body. Physicians experienced in
HIV disease care will feel comfortable with many subspecialty-related problems, but diagnostic and therapeutic procedures normally performed by subspecialists will continue to be necessary (see Table 1). Given
the extraordinary diversity of the clinical manifestations of HIV disease, therefore, it is important for the
physician to pay attention to multiple potential diseases and effects.
Continuity
Continuity is important because of the life-long
nature of HIV infection. Most individuals enjoy a long
�S58
Smith. PRIMARY CARE AND HIV DISEASE
primary care physician in coordinating such services is
doubly important for patients with AIDS.
TABLE 2
Necessary Attributes of Primary Care for AIDS
Psychosocial and Environmental Issues
I. Competent
A. Knowledgeable about presentation of disease, medicines, and
procedures
B. Specialists available, when necessary
C. Appropriate technology available
II. Compassionate
A. Free of stigma attached to homosexuality or drug use
B. Tolerant of a variety of lifestyles and values
C. Sensitive to issues of death and dying
Considerate
A. Economically accessible
B. Geographically accessible
C. Responsive to diverse medical, emotional, and social needs
Psychosocial and environmental issues also pose a
challenge in AIDS care. Patients often require assistance
with issues such as entitlements, legal needs, and advocacy with insurance companies. Patients justifiably fear
discrimination in employment, housing, and even
health care because of their disease, and they are frequently confronted with the possible revelation of
some aspect of their behavior (homosexuality or drug
use) that may have been kept a secret from their closest
friends and family. Patients thus face not only a fatal
disease, but one that stigmatizes them as well —a predicament that calls for caregivers to look after the
whole life of the patient, and not merely the dysfunction of one or more organs .
Table 2 lists necessary attributes of primary care
for HIV-infected patients. While this is a set of ideal
standards that may not be completely achieved in practice, they are within the grasp of most general internists, particularly for patients in the early stages of HIV
disease. It therefore seems reasonable to expect every
generalist to be able to do four things within this context; counsel patients and their families about the
transmission of HIV and how to prevent infection;
diagnose HIV infection through appropriate counseling, testing, and interpretation of test results, manage
early infection, including making appropriate decisions about CD4+ lymphocyte monitoring and the initiation of antiviral therapy and Pneumocystis prophylaxis; and recognize potential manifestations of
advanced HIV disease such as cancers and infections, so
that the patient may be appropriately managed or referred to an AIDS specialist. A growing number of
12
15
IV. Cost-effective
A. Maximizes prevention of further transmission
B. Maximizes health maintenance for those already infected
C. Minimizes unnecessary utilization
D. Minimizes unnecessary use of expensive facilities and drugs
period of asymptomatic infection. Without major
breakthroughs in antiviral therapy, however, we may
expect that progressive immunosuppression by HIV
will eventually occur and prove fatal. Patients will progress from being relatively well to i l l ; from ill to terminally ill. Because of this relentless progression, patients
who have HIV infection need physicians who will continue to follow them as they decline. Issues such as
institutional placement and advance directives for terminal care are best handled by a caregiver who has
come to know the patient over time.
Another feature of HIV disease that calls for continuity of care is the additive nature of antimicrobial
therapy. Patients with AIDS may have multiple opportunistic infections; with each new infection, the period
of acute treatment must be followed by life-long suppressive therapy. Therefore, patients who have AIDS
may eventually receive Pneumocystis prophylaxis,
chronic antifungal and antitoxoplasmosis therapy,
treatment for peripheral neuropathy, and suppressive
therapy for herpes simplex. As the patient's pharmaceutical regimen becomes more and more complex,
the physician's familiarity with the case is increasingly
important.
9
TABLE 3
Current Requirements for Persons Needing Medical Services Because
of HIV Infection
All stages
Concern about prognosis
Mental health
Prevention of transmission
Preventive care (e.g.. immunizations)
Primary care for non-HIV-related medical problems
CD4-I- cell count > 500
Laboratory monitoring
Coordination
Coordination is important in the setting of HFV
disease, not only to ensure communication between
different medical specialties, but also to foster collaboration between physicians and non-physician caregivers. Patients who have HIV disease will often require
complex home care regimens or institutional placement prior to death.
The traditional role of the
1011
CD4+cell count 200-500
Antiviral therapy
Laboratory monitoring
CD4+ cell count < 200
Antiviral therapy (perhaps combination)
Prophylaxis against opportunistic infections
Treatment of acute intercurrent infections and tumors
Long-term suppression of opportunistic pathogens
Decisions about terminal care
�JOURNAL OF GENERAL INTERNAL MEDICINE. Volume 6 (January/February Supplement). 1991
S59
and patient care. In each of these areas the general
internist will play an important role in the provision of
HIV-related primary care and the training of future
caregivers.
TABLE 4
Obstacles to Increased Primary Care for HIV
. Fear of contagion
d. Prejudice toward affetted groups
Teaching
3. Inadequate reimbursement
4. Inadequate medical knowledge
5. Fear of effects on practice
6. Institutional opposition (e.g.. from bospital administrators)
7. Discomfort with death and dying
8. Lack of professional support
generalists — particularly those in high-incidence HIV
areas — will be comfortable managing late-stage disease as well.
OBSTACLES TO PRIMARY CARE OF HIV
Most primary care for HIV-infected patients over
the past decade has been delivered by a relatively small
number of providers. A substantially expanded capacity for priman- care will be necessary in the future, as
medical care moves beyond its present focus on the tip
of the HIV "iceberg" — those patients who have fullblown AIDS With mounting evidence of the effectiveness of early intervention, all of the approximately one
nillion infected Americans will now need medical
.valuation and appropriate care - (Table 3)
Physical examinations, histories, the monitoring
of CD4+ lymphocyte counts, administration of antivirals, decisions about immunomodulator therapies,
and Pneumocystis prophylaxis will all require additional physicians to take care of HIV-infected patients. The shonage of primary care physicians ready,
willing, and able to care for HIV-infected patients in the
United States has a number of contributing causes.
Table 4 presents the main obstacles to increased care of
HIV among primary care physicians.
Public and private teaching hospitals and their associated outpatient facilities currently provide the bulk
of care in the United States
Clearly, the problem is
multifactorial, and no one solution will be successful in
increasing the pool of physicians available and
equipped to take care of people with AIDS. Physiciansin-training are becoming more and more familiar with
HIV care in the course of their training; few teaching
centers these days are without substantial AIDS experience. Academic general internists, therefore, will play
several key roles in the provision of care today and the
preparation of tomorrow's caregivers.
14
15
16
17
18
2 0
1 9
2 1
ROLES FOR ACADEMIC GENERAL INTERNISTS
Academic general internal medicine, like the rest
of academic medicine, continues to perch upon the
traditional "three-legged stool"; teaching, research.
In many academic medical centers, general internists arc at the core of the clinical teaching program
Particularly in the outpatient setting, where much of
HIV care is now concentrated, depanments of medicine usually look to divisions of general medicine for
didactic teaching and teaching by example for medical
students and housestaff. In addition to general outpatient medicine, general medicine has also pioneered
the examination of the medical interview as a subject
for study and exposition. The HIV epidemic reminds us
that areas of interviewing historically given short shrift
in medical school training, such as patients' sexual behavior and drug use, are vitally important for the development of an adequate database and good patient care
General internists have also taken the lead in teaching
other imponant areas in the "art" of medicine that are
vitally important for caregivers in the midst of an epidemic, such as ethics, death and dying, and cultural
sensitivity. They will have an imponant role in teaching
the applications of these areas to the current generation
of students and residents.
Research
In addition to AIDS-related clinical trials, in which
many general internists participate, the AIDS epidemic
desperately needs research in other areas that general
internists have pioneered and led: cost-effectiveness
and cost - benefit analyses of alternative diagnostic and
therapeutic strategies ; development, validation,
and application of instruments to measure quality of
life for patients with HIV infection ; investigation of
the qualities of care rendered by different physicians,
within different institutions, and by different models of
care delivery ; and alternative financing mechanisms. As effective intervention lengthens the survival of people with HIV infection, traditional endpoints for clinical trials — death or serious
morbidity — will, thankfully, become rarer. There is.
therefore, a great need for internists with health services research skills to assist other AIDS researchers in
the development of other measures as endpoints for
new AIDS treatments.
22 2 4
25 2 6
27 2 8
29
30
PATIENT CARE
The level of sophistication of this issue of the Journal of General Internal Medicine demonstrates that
general internists are already rendering high-quality
care to HIV-infected patients. In many areas, infectious
disease specialists who once were primary providers
for AIDS patients now provide only consultative rather
�S60
Smith. PRIMARY CARE AND HIV DISEASE
than primary care, which is increasingly in the hands of
generalists. Generalists' contributions to primary
care, however, go beyond its current provision. They
also serve as role models to students and house officers;
in medical practice generally, particularly outpatient
practice, it is important for young physicians-in-training to see generalists caring for patients with AIDS. In
addition, part of the process of diffusing AIDS care
among more community practitioners will require the
development of referral networks and backup with tertiary centers so that such practitioners may feel comfortable in taking on care, particularly when they may
not have much experience with HIV-infected patients.
Academic generalists, therefore, can serve as a link for
both patients and providers between community settings and the sophisticated referral centers that will be
necessary for some aspects of HIV care.
31
32
them is the increasing complexity of therapy. It now
seems likely that the next four to five years will see the
emergence of new antiviral agents such as 2',3'-dideoxyinosine (ddl). the growth of combination antiviral therapy, ' the possibility of effective immune
modulator therapy . prophylaxis for additional opportunistic infections, and improved treatment (including
lifelong suppression) for others. It is hard to imagine
physicians undertaking to learn and manage such therapy without significant numbers of HIV-infected patients in their practices; it is equally hard to imagine
every generalist having significant numbers of HIV-infected patients in his or her practice, given the relative
concentration of the epidemic.
Another factor that may encourage continued concentration of care is the pursuit of quality. Some early
evidence suggests that patients who have Pneumocystis carinii pneumonia (PCP) have better outcomes
when cared for by experienced providers. Since PCP
is by far the most common of the opportunistic infections associated with AIDS, it may be that even larger
disparities exist in the quality of care for toxoplasmosis
or infections with the Mycobacterium avium complex. Therefore, health care systems, regulators, and
patients may push for the restriction of care to those
physicians who have substantial experience.
A third factor continuing to concentrate care is the
financial issue. As more and more patients with AIDS are
concentrated among minorities and among F drug
V
users and their first- or second-generation sexual
partners, lack of medical insurance will naturally force
such patients into settings where ali uninsured patients
go: public hospitals, public clinics, and other facilities
where physicians care for the poor. Despite the medical mainstreaming of AIDS care, patients may be concentrated among a few providers because they lack private medical insurance, unless there are major changes
in thefinancingof American health care in general or of
AIDS care in particular.
In the final analysis, however, the configuration of
primary care for HIV disease 15 years from now will
probably depend mainly on the way in which therapy
evolves. If, by then, there is a relatively simple, stable
regimen for HIV — perhaps one or two medicines that
prevent progression and decrease or prevent transmission, then it is likely that most generalists will feel fairly
comfortable in providing such care. If, in other words,
there is an "insulin " or a "thiazide" for HIV infeaion,
its care will probably diffuse much like that for diabetes
and hypertension. If. on the other hand, the state-ofthe-art 15 years from now consists of complex regi mens
being actively researched — that is, if care resembles
the current state of oncology — then HIV care will
probably continue to be as concentrated as is care for
malignancies today.
This is not to say that such care requires formal
44
What does the future hold for HIV disease primary
care? Will every internist, family practitioner, or general practitioner in the year 2005 regard AIDS as we
currently do rheumatoid arthritis or angina pectoris?
Will all neighborhood health centers and community
clinics routinely provide care to patients with AIDS? I
don't know; I am not sure that it is possible to know.
Some forces certainly are acting to encourage more dispersion, but they are counteracted by other forces that
.ontinue to create concentration of AIDS care.
The main force acting for more diffusion of AIDS
care is the sheer number of patients who need such
care. Many neighborhood health centers, private physicians, and others who three years ago regarded AIDS as a
rarity have now seen several cases and have set about
educating themselves because of their realization that
they undoubtedly have HIV-positive patients in their
practices. Additionally, the development of more and
more effective therapies for HIV infection and its various sequelae means that the epidemic fits more into
the curative model with which most physicians are familiar and comfortable. There is, in other words, more
for them to "do."
The problem of inadequate knowledge is easing
somewhat, as information is increasingly available not
only from "mainstream" medical journals but from a
growing number of AIDS-specific peer-reviewed journals, other periodicals, textbooks, " computerized databases, and continuing medical education opportunities. Last, as the disease is "medicalized" and
more accepted by society, it loses some of its stigma
among physicians as well; it thus becomes less personally or professionally threatening to physicians to care
' i r patients.
On the other hand, some factors seem to portend
continued concentration of HIV care. Chief among
33
34
43
35
42
t
;i
4
i
48
27
CONCLUSION
i
47
i:
«
e
tl
n
ii
A
P
P
Tl
C
r<
�JOURNAL OF GENERAL INTERNAL MEDICINE. Volume
subspccialirv training in infectious disease, oncology,
or even internal medicine; on the contrary, it is evident
at many good general internists, family practitioners,
.d general practitioners can. with enough interest and
experience, become expert AIDS care providers. But it
may be that care for patients with advanced disease is
so complicated that it can be competently rendered
only by physicians who have substantial training and
experience. Regardless of the specialty affiliations of
those providing care, the humanistic qualities of primary care will continue to remain an imponant requirement for patients suffering with this most intriguing, most demanding, and most tragic of diseases.
Academic generalists must play an important role in
passing on these qualities to new generations of
physicians.
The author gratefully acknowledges the helpful comments of Molly
Coye. Richard Chaisson. and Lois Eldred, and the assistance of Clair
resc Ward in the preparation of the manuscript.
REFERENCES
1
2.
3.
4
5.
6.
7.
8.
9
10
1 1.
12.
13.
14.
15.
16.
7
Pneumocvstis pneumonia — Los Angeles
MNfWR. 1981;
3<>:250-2'
Chaisson RE. Living w i t h AIDS JAMA 1990;263:43-1
Northfelt DW. H a r v a r d RA, Shapiro MF. The acquired immunodeficiency syndrome is a priman' care disease. Ann Intern Med.
19HH;l09:?'73-5.
AIDS/Hiv Suneillance; CDC July 1990.
Cancer Statistics 1990. CA. 1996;40:17-20
Uaumgardner J. Physicians' senices in the division of labor
across markets J Polit Economy. 1988;96:948
MakadonJH, Delbanco SF, DclbancoTL Caring for people w i t h
AIDS and HIV infection in hospital-based p r i m a n care practice. J
Gen Intern Med.1990:5:446-50.
Stoeckle.lD Tasks of p r i m a n care. In: Goroll AH. May LA. Mulley
AG eds Priman care medicine: office evaluation and management of the adult patient Philadelphia: J B Lippincott.
1987:1-4
Smith MD An integrated p r i m a n care system for HIV disease.
Maryland Med J. 1990:39: J 73-6.'
Benjamin AE. Long-term care and AIDS: perspective from experience w i t h the elderly. Milbank Mem Fund Q 1988:66 4 1 5-43
Drostc T Going home to die developing home health care ser
vice for AIDS patients Hospitals 1987;61 ( 1 6 ) 54 8
Blendon RJ. Donclan K. Discrimination against people w i t h
AIDS: the public's perspective N Engl J Med 1988:319:1022-6.
Volberding PA The clinical spectrum o f the acquired immunodeficiency syndrome: implications for comprehensive patient
care. Ann Intern Med. 1985:103:729 33
Andrulis DP. Weslowski VB. Hospital trends and the financing of
medical care for AIDS Patients In: LcVcc VCN. ed New perspectives of HIV-related illnesses: progress in health senices research. Rockville. MD: National Center for Health Semces Research. 1989
Volberding PA, Lagakos SVI. Koch MA, et al Zidovudine in
asymptomatic human immunodeficiency virus infection. A con
trolled trial in persons w i t h fewer than 500 CD4-posiiive cells
per cubic millimeter. N Engl J Med. 1990;322:94 1-9.
Fischl MA, Richman DD, Hansen N , et al. The safety and cfticacv
of zidovudine (AZT) in the treatment of subjects w i t h mildlv
symptomatic human immunodeficiency virus type 1 ( H I V ) in
fection. A double-blind, placebo-controlled trial. Ann Intern
Med. 1990;112:727-37.
State-of-the-art conference on AZT therapy for early HIV infec
t i o n . Recommendations for zidovudine: early infection. JAMV
1990;264:1606-7
6 (January/February
Supplement).
1991
S61
18 Cooke M, Koenig B, Beery N, Folkman S. W h i c h phvsicians w i l l
provide AIDS care? VI International Conference on AIDS. San
Francisco. 1990; abstract S D.50.
19 Kelly J, St Lawrence JS, Smith SJr. et al. Stigmatization of AIDS by
physicians. Am J Public Health. 1987;77:789.
20 Andrulis DP, Beers VS. BcntleyJD. Gage LS. The provision and
financing of medical care for AIDS patients in U.S. public and
private teaching hospitals. JAMA. 1987,258:1343-6
21 Andrulis DP, Weslowski VB, Gage LS. The 1987 US hospital AIDS
suney.JAMA. 1989;262:784-94.
22 Smith MD, Murray RB. McAvinue S, Suldan M. Chaisson RE. Cost
effectiveness of aerosolized pentamidine (AP) prophylaxis
against Pneumocystis c i r i n i i pneumonia V i t h International
AIDS Conference. San Francisco, June 22, 1990. abstract F.D.
8.30
23 Freedberg KA. Tosteson ANA, Cohen CJ. Cotton DJ Arc AZT and
prophylaxis for Pneumocystis carinii pneumonia (PCP) cost-effective? V i t h International AIDS Conference, San Francisco. June
22. 1990, abstract F.D 832.
24 Clean PD, Barn MJ, Mayor KH. Brandt AM, Gostin L. Fineberg
HV. C o m p u l s o n premarital screening for HIV JAMA.
1987;258:1757-62
25. Ellcrman D, Smith MD. Shrinking options: q u a l i n of life for
persons w i t h AIDS in Philadelphia . Vth International Conference
on AIDS. Montreal. Canada. June 1989.
26 Wu AW, Mathews WC. Brysk LT, et al Quality of life in a placebocontrolled trial of zidovudine in patients w i t h AIDS and AIDS-related complex in AIDS clinical trials. J AIDS. 1990 683-90.
27 Bennett CL, Garfinkle JB. Greenfield S, et al The relation between hospital experience and in hospital mortalitv tor patients
w i t h AIDS-related PCP. JAMA. 1989:261:2975-9.
28. Ball JK. Farley DE. Turner BJ. Does familiarity breed effectiveness
in the treatment of AIDS patients? Evidence f r o m a national sample of U.S. hospitals. Vith International Conference on AIDS. San
Francisco. June 2 1 . 1990. abstract S D. 794.
29. Makadon HJ, Seage GR. Thorpe KE. Fineberg HY Paying the
medical costs of the HIV epidemic: a review of policv options. J
AIDS 1990;3:123-33.
30 Smith MD. Community perceptions and reality-based AIDS research. In: LcVce WN. ed. New perspectives on HIV-related i l l
nesses: progress in health services research. Rockville. MD: National Center for Health Semces Research. September
1989;I35-42.
3 1 RizzoJ. Marder W. Willke R. Physician contact w i t h and attitudes
toward HIV-seropositivc patients: results f r o m a national survey
Med Care. 1990;28:251-60.
32. Cooke M. Sande MA. The HIV epidemic and training in internal
medicine. Challenges and recommendations N Engl J Med
1989;321:1334-8.
33 AIDS. The Journal o f Acquired Immune De/ic/eno Syndromes.
and AIDS Research and Human
Retroviruses
34. AIDS Clinical Care. AIDS Alen. AIDS Medical Repon AIDS Tar
geicd Information Newsletter, AIDS Patient Care, and CDC AIDS
Vieekly
35 Sande MA, Volberding PA. The medical management o f AIDS.
2nd ed Philadelphia: W.B Saunders. 1990.
36 Cohen PT. Sande MA. Volberding PA. The AIDS knowledge base.
Waltham. MA: The Medical Publishing Group. 1990
37. VolberdingPJacobsonMA.eds.AlDSclinicalreview New York:
Dekker, 198938 DeVita VT, Hellman S. Rosenberg SA, eds. AIDS, etiology, diag
nosis. treatment, and prevention. 2nd ed. Philadelphia: J.B. Lippincott. 1988
39 Wormser GP. Stahl R. Bottonc EJ eds AIDS, acquired i m m u n e
deficiency syndrome, and other manifestations o l HFV i n f e c t i o n
epidemiology, etiology, immunology, clinical n u n i festal ions,
pathology, control, treatment and prevention. ParV Ridge. NJ
Moves Publications. 1987.
40 Holmes KK. ed. AIDS, a guide for the primary physician. Seattle.
WA: University of Washington Press. 1988.
4 I Pinching AJ. Weiss RA. Miller D. eds. AIDS and HIV infection: the
wider perspective. Edinburgh, Scotland: C h u r c h i l l Livingstone.
1988.
42. Y o u l c M . AIDS: therapeutics in HIV disease NewYorV. C h u r c h i l l
�S62
Smith. PRIMARY CARE AND HIV
Livingstone, 1988
4 3 The AIDS Compact Library.
44 Lambart JS. Scidlin M. Reichman RC, et al. 2',3'-dideoxylinosine
( d d l ) in patients w i t h the acquired immunodeficiency syndrome or AIDS-related complex. A phase I trial. N Engl J Med.
1990;322.1333-40.
45 Meng TC. Fischl MA. Richman DD AIDS Clinical Trials Group:
phase l / l l study of combination 2',3'-dideoxycytidine and zidovudine in patients w i t h acquired immunodeficiency syndrome
DISEASE
(AIDS) and advanced AIDS-related complex. Am J Med.
1990(May 21);88(5B):27S-30S.
46. Pederson C, Sandstrom E, Petersen CS, et al. The efficacy of
inosine pranobex in preventing the acquired immunodeficiency
syndrome in patients w i t h human immunodeficiency virus infection. N Engl J Med 1990:322.1757-63
47. Blendon RJ, Aiken LH, Freeman HE. Access to medical care for
black and white Americans: a matter of continuing concern.
JAMA 1989;261:278-80.
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
]O
�Tab 10.
This Section Includes:
Smith, Mark."An Integrated Primary Care
System f o r HIV
�An Integrated Primary Care System
for HIV Disease
MarkD. Smith MD, MBA
The autlwr is Associate Director, AIDS Senice. The Johns Hopkins Hospital.
rom the early days of the AIDS epidemic it has been apparent
that successful care of HIV disease would force a re-examination
of existing health care systems. The initial manifestations ot
acquired immune deficiency werepneunwc\stis carinii pneumonia
and Kaposi's sarcoma. Later, other opportunistic infections such
as cryptococcal meningitis, toxoplasmosis encephalitis, and typical and atypical mycobacterial infection were recognized. More
recently, HIV-related dementia and wasting syndrome have become part of the official AIDS definition. In addition, it is clear
that a number of non-Kaposi's neoplasms (particularly lymphomas) and increased susceptibility to nonopportunistic
pathogens are part of the spectrum of HIV disease.
The good news, outlined more specifically elsewhere in this
special issue of the Maryland Medical Journal, is that the number of
therapeutic options for HIV disease continues to increase. Not
only is there a wide range of treatments available for opportunisticinfections, but also the first licensed anti-HIV drug. Zidovudine,
now clearly has a role in both early and late infection. Furthermore, new antiviral drugs and immunomodulators should further
broaden physicians' therapeutic options. Health care providers
and systems, however, now face a dilemma: we know more than
we can deliver. Thus the growth in the available therapeutic options and the increasing life span of people with HIV disease (even
in its later stages, AIDS) present clinicians with a problem: how to
construct and maintain a primary care system capable of handling
patients with the disease.
1
:
2
3
Although a "cure" for HFV infection
has not yet been found, there are
increasing numbers of options for
treatment. Like other more familiar
diseases, HIVIAIDS present lifelong
management responsibilities for
both patient and clinician.
4
Why is AIDS a Primary Care Disease?
The primary care movement has grown in ihe last 20 years as a
reaction to the increased specialization and compartmentalization
of American medicine. One definition of primary care deals with
the patient care functions of primary care doctors: to provide
access, continuity, and integration. Another text has a somewhat
broader view of the primary care function:
5
1) Medical diagnosis and treatment; 2) Psychological diagnosis and treatment; 3) Personal support of patients of all backgrounds, and all states of
MMJ February 1990
173
�illness; 4) Communication of information about diagnosis,
treatment, prevention, and prognosis; 5) Maintenance of
patients with chronic illnesses; 6) Prevention of disability
and disease through detection, education, persuasion, and
preventive treatment.
Care of persons with H I V disease clearly requires all
six of these tasks. AIDS also has a number of distinctive features that make it particularly suitable for a
primary care approach:
HfVInfection is a Chronic Disease Although a "cure"
has not yet been found, there arc increasing options for
treatment: thus HIV, like the more familiar diabetes
and hypertension, presents lifelong management
responsibilities to patient and clinician alike.
HIV Infection Lt a Multisystem Disease H I V and opportunistic pathogens can affect every organ system of
the body. Unless the patient has a physician possessing wide knowledge in a number of fields and capable
of judicious use of specialists were necessary, the
patient is at risk of being parcelled out among opt h a m o l o g i s t s , d e r m a t o l o g i s t s , neurologists,
gastroenterologists, etc.
HIV Infection is a Highly Labile Disease Not only can
patients go ten years or more after infection without
showing symptoms, but also acute exacerbations of
disease such as opportunistic infections may be
punctuated by relatively long stretches of less
symptomatic disease. Physicians experienced in H I V
care know that patients may look well one day and be
deathly ill the next; conversely, some who require intense inpatient hospital treatment may, after antibiotic, antineoplastic, or other therapy, revert to their
baseline state. The implication for the health care
system is that HIV-infected patients, particularly those
who are symptomatic, must have access to providers
who can respond rapidly to changes in their clinical
condition.
What Are the Requirements of a Primary Care
System for HIV?
Outpatient Care The linchpin of primary care for
HIV, as for other chronic diseases, is outpatient care
(Figure). The past few years have shown that such
Care
^
^
Care
care, to be effective, must have three central characteristics: aggressiveness, multidisplinary character,
and linkages with other parts of the system.
Because of pressures to reduce inpatient hospitalization, H I V outpatient care has become substantially
more aggressive as physicians have become more and
more comfortable with new techniques. Many
providers caring for HIV-infected patients now
routinely perform such diagnostic and therapeuticprocedures as lumbar punctures, transfusions, intravenous hydration, administration of chemotherapv.
and induction of sputum for diagnosis of pulmonary
infections; each of these was either developed in
response to the AIDS epidemic or would have been
unthinkable except on an inpatient basis as recently as
five or six years ago.
A multidisplinaty approach is key. Because patients
present with a variety of medical, emotional, and social
problems, it is important for outpatient care to be able
to call on the expertise of medical subspecialists as well
as psychologists, psychiatrists, and social workers.
Smooth working relationships and, if possible, convenient arrangements for rapid referral to such colleagues as neurologists and opthamologists have been
an important part of delivering primary care at The
Johns Hopkins Hospital and other major H I V
centers.
Outpatient medicine can sometimes be hectic: often
there is not time to pursue extended workups or discuss
patients' social and emotional needs. It is therefore
particularly important that outpatient care have strong
linkages with the other forms of care HIV-infected
patients will need: communication with inpatient
units, chronic care facilities, and home health
providers will substantially ease the tasks of caring for
patients in a busy outpatient setting.
Inpatient Care Despite the intensity of outpatient
H I V care, most patients eventually will require
hospitalization for one or more periods. It is important that providers caring for HIV-infected patients
have access to appropriate inpatient settings when
they become necessary. Virtually all hospitals in
Maryland have cared for at least one AIDS patient in
the course of the epidemic. Some, like Johns Hopkins,
have dedicated A I D S units; most do not.
Whatever the arrangement, it is important that
hospitals be able not only to provide competent
medical care to AIDS patients but also to provide a supportive and compassionate environment free of stigma or hysteria.
Home Care The H I V epidemic has stimulated substantial innovation in home care.
Many companies, both not-for-profit and
profit-making, have become expert in tlie
delivery of intensive home-based antibiotics
and other therapies. Many patients can be
maintained at home with treatments that a few
years ago would have required hospitalization.
Such solutions are cost-effective and usually
preferred by the patient.
7
W
Home
f
8
Care
Figure. Spectrum of Facilities Needed for HIV Care
174
MMJ
Vol 39 No 2
�Chronic Care Some patients will, toward the end of
their disease, become so debilitated that they need
more intensive care than a home environment can
provide. At the same time, their conditions (HIV
dementia, wasting syndrome, and others) will not
respond to acute intervention, and therefore do not
require inpatient hospital care. For those patients,
chronic care facilities are a natural option particularly
if they do not have home/family situations that permit
constant attention at home. Currently, few nursing
homes accept AIDS patients; the provision of more
chronic care beds for HIV-infected patients is one of
the most pressing priorities for the health care system.
Terminal Care Although life expectancy continues
to increase and new therapies are promising, we may
expect that most of those currently ill with AIDS will
eventually die of it. It is important to discuss with
patients and their loved ones their wishes regarding
terminal care. Although many AIDS patients were
initially hostile to the concept of hospice because of its
connotation of "giving up," hospice services now find
increasing acceptance in the AIDS community. Around the country hospices (both institutional and
at-home services) have sprung up to care for people in
the terminal phase of HIV infection. While hospice
care is not an alternative that all patients will choose,
it often will provide appropriate palliative services to
keep patients comfortable in their final days.
Psychosocial Suppon Many AIDS patients will have
tremendous need for psychological, emotional, and
social support. Some of this need stems from the fact
that AIDS, like many forms of cancer and other
chronic fatal illnesses, saps both the patients' physical
strength and financial resources. In addition, however, society continues to harbor substantial prejudicial attitudes toward people with AIDS. Patients
therefore face a variety of threats to their social supports (job, insurance, and even family support) that are
uncommon in other diseases. Many states, including
Maryland, have developed case management systems in
an attempt to respond to these needs. Case management seeks to provide patients with an individual who
can help coordinate services such as transportation,
application for entitlement, and access to other needed
medical and social services. AIDS patients will vary
widely in their need for such services. Some, particularly those in advanced stages of disease or with
substance abuse problems, will need substantial assistance in managing their affairs. Others will be quite
self-sufficient and able to coordinate services and make
their own decisions. Whatever the patient's needs,
physicians should familiarize themselves with the
community based support available in their area and
refer patients to organizations and agencies that may
assist them with this range of needs.
9
1Q
n
12
Obstacles to the Development of a
Primary Care System
While substantial steps have been made toward the
MMJ February 1990
creation of an integrated system to care for HIV-infected patients, much still has to be done in Maryland
and elsewhere. The principal obstacles are: inadequate trained personnel, lack of proper financing of
care, and institutional wariness about AIDS.
The problem of the provision of enough trained
medical personnel to care for AIDS patients
throughout the stages of their illness is a difficult one.
While most community settings will rely on physicians
to provide this care, The Johns Hopkins Hospital, San
Francisco General Hospital, and a number of other
major HIV centers have relied heavily on so-called
"mid-levels" (nurse practictioners and physician assistants). We clearly will need more physicians to be
involved in this care. The shortage of physicians to
care for people with AIDS is a multifactoral problem
(Table). For any individual physician the main
obstacle may be one or more of the specific areas cited;
nevertheless, it will be the task of the medical community, academic centers, and local, state, and federal
health officials to overcome each of these obstacles
confronting physicians.
Table. Obstacles to Physicians' Caring for
HIV - Infected Patients
Information overload
Inadequate reimbursement
Fear of infeaion
Fear of loss of other patients
Institutional (e.g. hospital) pressure Burnout
Dislike of homosexuals
Dislike of drug users
Adequate financing of HIV care is a serious problem both to private physicians and to institutions.
Many patients with AIDS must rely on Medicaid from
the start, or they eventually will become impoverished
and turn to Medicaid. Inadequate Medicaid reimbursement for outpatient care, home care, and other parts
of the care spectrum is a serious impediment to providing quality care to AIDS patients.
The last major obstacle to creation of an integrated
primary care system for patients with HIV infection is
the wariness with which many institutions approach
care for these patients. Hospitals, nursing homes,
home health care agencies, and others face some of the
same financial problems faced by physicians. But their
inability (or perhaps unwillingness) to solve these
problems - felt most strongly in chronic care facilities
- is a formidable obstacle to the creation of a rational,
compassionate, and cost-effective system.
References
1. Volberding PA. The clinical spectrum of the acquired immunodeficiency syndrome: Implications for comprehensive patient
care. Ann Intern Med 1985;103:729-33.
2. Centers for Disease Control. Kaposi's sarcoma and Pneumocystis pneumonia among homosexual men - New York City and California. MMWR 1982;30:305-308.
3. Chaisson RE, Volberding PA. Clinical manifestations of HIV
infeaion. In Mandell GL, Douglas RG, Bennett JE, eds. Principles
and Praaice of Infcaious Disease. New York: Churchill Livingstone
Inc., 1990: 1059-91.
4. Monfardini S. Italian Cooperative Group for AIDS-Relatcd
175
�tumors. Malignant lymphomas in patients with or at risk for AIDS
in Italy. JNCI 1988;80-855-60.
5. Alpcrt JJ, CharneyE: The Education of Physicians for Primary
Care. DHEW Publication No 74-31B. US Government Printing
Office, 1975.
6. Stocckle JD. Tasks of primary care. In Goroll A H , May LA,
Mulley A G . Primary Care Medicine: Office Evaluation and
Management of the Adult Patient. Philadelphia:Lippincott and Co.,
1987: 1-4.
7. Grinkcr WJ. Out-of-hospital care for persons with AIDS. In
Rogers DE, Ginzberg, eds. The AIDS Patient: An Action Agenda.
Boulder: Westview Press, 1988: 42-49.
176
8. Bennett J. Helping people with AIDS live well at home. Nurs
Clin North Am 1988;23(4):731-48.
9. Benjamin AE. Long-term care and AIDS: Perspectives from
experience with the elderly. Milbank Mem Fund 1988;66(3):41.5-43.
10. Drostc T. Going home to die: Developing home health care
services for AIDS patients. Hospitals 1987;61(16):54-58.
11. Clark C,CurleyA,HughesA,JamcsR. Hospice care: A model
for caring for the person with AIDS. Nurs Clin North Am
1988;23(4):851-62.
12. Blendon RJ, Donelank. Discrimination against people
with AIDS: The public's perspective. N Engl J Med 19S8;319
1022-26.
•
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
�Tab 11.
This S e c t i o n Includes:
Bennett, Charles L., G a r f i n k l e , J e f f r e y
e t a l . "The R e l a t i o n Between H o s p i t a l
Experience and I n - H o s p i t a l M o r t a l i t y f o r
P a t i e n t s with AIDS Related PCP" JAMA
May 28, 1989- Volume 261, No.20
�Original Contributions BB
The Relation Between Hospital
Experience and In-Hospital Mortality
for Patients With AIDS-Related PCP
Charles L. Bennett, MD; Jeflrey B. Garfinkle. MS; Sheldon Greenfield, MD; David Draper, PhD;
William Rogers, PhD; W. Christopher Mathews, MD, MSPH; David E. Kanouse. PhD
There Is marked debate by physicians and policymakers regarding the creation at hospitals with low familiarity? Wc
of regionalized acquired immunodeficiency syndrome (AIDS) centers. A central have selected PCP because it is a seriissue is whether outcomes of care, particularly mortality, differ as a function of ous complication of AIDS that often
leads to death and, unlike many other
hospital experience with patients with AIDS. We evaluated the experience of
complications of AIDS, because it is
257 patients with AIDS and Pneumocystis carinii pneumonia treated at 15
amenable to therapies that are available
California hospitals between October 1986 and October 1987. An overall 15.2% currently. Patients with AIDS and PCP
In-hospital mortality rate was observed. However, a markedly lower In-hospital infection pose difficult diagnostic and
mortality rate was observed in the group of patients treated at hospitals that had therapeutic dilemmas for health care
a high level of experience with patients with AIDS (>30 human Immunode- workers, such as diagnosing additional
ficiency vims-related discharges per 10 000 hospital discharges) relative to the life-threatening opportunistic infecgroup treated at hospitals with less experience (<30 human Immunodeficiency tions besides PCP. Therapy also is diffivims-related discharges per 10 000 hospital discharges): 12% vs 33%. Other cult because of the high incidence of
factors significantly associated with in-hospital mortality Included intensive care severe toxic reactions to anti-PCP therunit use, admission from an emergency department or through an interhospital apies in patients with severely comprotransfer, and a history of hospitalizations. A logistic regression model indicated mised hematologic and immunologic
systems.
that, after controlling for severity indicators, AIDS experience remained significantly related to mortality. Our findings suggest that policymakers should
for editorial comment see p 3016.
consider three options: creating regional AIDS centers, implementing policies
that promote a rapid but carefully monitored Increase In experience of lowvolume hospitals with human immunodeficiency virus-infected Individuals, or
There is some evidence that experiproviding highly focused educational efforts at low-AIDS-experlence facilities. ence in treating a disease might be un
Without such policy initiatives, differences in mortality rates like those we have important factor in minimizing adverse
found might persist as cases of AIDS begin to occur in every area of the country.patient outcomes. Previous studies
1
{JAMA. l!IH9;2r>l:297G-2979)
CARE FOR patients with acquired immunodeficiency syndrome (AIDS), including those with rneumocystis carinii pneumonia (PCP), is now concentrated in a few cities in the United
States and within certain hospiUla in
these cities. As cases of AIDS begin to
occur across the country, there is debate whether increased regional support by federal, state, and local govemment* for B)>cciric AIDS centers iu an
alternative to the natural diffusion of
AIDS care into local hospitals.
This study focuses on one aspect of
the debate over care of patients with
AIDS by addressing the followinf; isnue:
Are patients with AIDS who have a
diagnosis of PCP infection morelikely
to be discharged alive when they are
treated at hospitals that have a
high familiarity with patients with
AIDS compared with poticnts treotcd
From Iho RAND/UCLA Centef loi Ho»ith Policy Study
ancl Ilia Deparlmnnl ol Medicine. Umveony ol CaMoima al Lot Anoolea Sclmol ol Mmjinna ID' Hommltl: llm
owa) Dr Bunnuti 1 a Naiional Cenior tor HoaUh Sor3
vir.m nosoorch poutdoctoroi lolkw at Iho RAND/UCLA
D e p a r l m n n l o l M e d i c i n e . Tulta UnivefBily e n d t h e f4ew
4
fcnoiand ModicalCenior Ooiton M m i ( O f Qmenlmidl
Iho Dapeftrnantt ol Iconomict and Siainnca (On
Draper and HoQeri) and Behavioral Uciencea (Dr Konontei Hie HAND Corp. Santa Monica. CaM (Mr OarlmMol nnrt the Dopartment ol Medicmo. Umvornly ol
Caliiorrua al San Diego Sctiool ol Medicine (Dr Mulh-
JAMA, May ? 0 , W O - V o l ? 6 I . No 20
m
C e n i o r lor I *i»alth f Vihcy Suirly
The opmionfc a n d c o n c l u s i o n o o ' p r r i f s o d h o r o m aro
iOloly thoae n l l l m nulhnrfl a n d s h o u l d not h e c o n s l r u o d
us r e p r e s u n t i n o i i ' r j o p i n i o n s or p o l i c i e s o l Tho R A N D
C o r p or the N a l i o n o l Center lor H o o l l h b u r v i c e s
Fiofcearch
R e p r i n l renuoMS l o I h o n A t . ' D C o r p o r n l i f i n . 1700
Mam si. Santa Monica CA oatoi, (Di Uunnuit)
have shown that hospitals that provide
care for large numbers of patients receiving specific surgical procedures
(such as open heart surgery, vascular
surgery, transurethral resection of the
prostate, and coronary artery bypnsH)
experience lower in-hospital mortality
rotes relative to hospitals that perform
these procedures in fewer patientx."
Few studies have addressed the reliitionship of hospital volume to in-hospital mortality for medical conditions.
These studies have evaluated patients
with diseases such ns acute myocnrdinl
infarction, peptic ulcer disease, subarachnoid hemorrhage, and bums." One
study focused on the relationship between hospital volume and mortality in
patients with short-term heart disease;
no relationship was found.'
Other researcliers have attributed
differences in in-hospitul mortality
rates to one of two causes: (1) hospitals
with low mortalitv rales are further
Mortnlily u n d AIDS nelntotl PCP-- tlnnrmtl nl ul
mmmmmmm
2075
-
�f
Table 1.-Chara«erislics ol the 15 Hospitals Studied
Hoapttal
No.
Teaching
Sutua*
Occupancy
Rataf
No. Ol
Mat
Ho. ol
Dlachargae
(In 1000»/y)t
WDSFamlllarlty
Indei*
22
11
80
80
1
Ligw
Hlgh-Famlllarlty Hoaptula
70
500
2
Ught
60
3
None
00
200
e
110
4
70
>5O0
45
60
5
Ught
Lighl
60
200
18
30
6
Heavy
BO
400
19
40
7
None
60
100
6
80
8
Nona
Low-Famlllarity Hoapttal•
200
70
10
5
9
None
70
400
14
12
10
None
70
400
15
7
11
Nona
70
400
19
5
12
Heavy
90
300
18
17
13
None
80
100
7
13
14
Light
60
200
15
7
15
None
70
400
20
4
500
'Basad on me ratio ol the numoer ol lull-time •qunaJeni rasldenta to the number ol beds: 0 - n o n e : between 0
and 0.25 - light, and >0.25 - heavy teaching
tTo avoid indenblicanon ol tne hoapnals. we report rounded values lor hospital-level characteristics.
•AIDS (acquired immunodeliclency syndromeHamillamy Index was defined as ine total number ol discharges ol
Individuals with human Immunodelidency virus Inlectlon per 10000 total hoapiUI discharges.
alonp on the "leaminp: curve" in caring
for the relevant patients or (2) a referral
system for patients might have been
esublished so that hospitals that
achieve better patient outcomes admit
relatively more patients with a particular disease.' With regard to AIDS, if
differences in in-hospital mortality
rates that reflect a relative advantage
for high-volume hospitals exist after adjusting for severity of illness, then (1)
regional AIDS centers might be necessary to minimiie the risk of death in
patients with AIDS, (2) hospitals with
low numbers of patients with AIDS
might need to increase rapidly their volume of patients with AIDS and PCP
infection, or (3) highly focused, intense
educational efforts might need to be directed toward hospitals with low volumes of patients with AIDS. With a
rapid increase in experience or education, these hospitals can move up the
learning curve related to the diagnosis
and treatment of PCP infection.
PATIENTS AND METHODS
Patient Population
We obtained a modified version of the
Uniform Hospital Discharge Data Set
from 15 hospitals located in metropolitan areas of California. Hospital and
patient confidentiality are maintained
in these analyses. Approval of the human subjects protection committee of
The RANI) Corp, Santa Monica, Calif,
was obtained prior to beginning the
study. The selected hospitals vary substuntially over a number of iitlributes,
sudi us teixliing status, ntunliernf beds,
occupancy rules, and frciiuency of
2976
JAMA. Miiy..'!).
I'jH'.)-
•/oll.'lil.
No. 'M
AIDS-related conditions. Hospital-level data included factors that reflect the
origin of patients at an individual hospital (the proportion of patients with human immunodeficiency virus [HIV]
-related disease that were admitted
from the emergency department or
transferred from another hospital) and
the proportion of patients with HIV infection who are insured by the statesubsidized Medicaid program (MediCal)
relative to all patienU with HIV
infection.
We selected patients with PCP infection diagnosed between October 198G
and October 1987. All patients were
documented to have HIV-related disease as defined by the International
Classification of Diseases modification
for HIV instituted in October 198G. Patient-specific data included age, sex,
race, diagnoses, type of admission
(routine vs urgent), patient origin
(emergency department, interhospital
transfer, physician office, or home), disposition on discharge, specific resource
use (eg, intensive care units or bronchoscopy), length of stay, insurance status, outcome of hospitalimion (survival
vs death), and disease severity. Severity of illness was based on one or more of
the following criteria: admission from
the emergency department or through
an interhospital transfer, presence of
concomitant infection or tumor, or prior
hospitalizations. The marker admission
was defined us the first episode of I'CP
infection fpr the individual patient nt a
given hospital. The data tracked individual patients both retros|K'ctively and
prospectively from the marker admission, with evaluation of all premurker
and postmarker admissions for e: oh patient to the individual hospital fur 2
years.
For each hospital, an AIDS-familiarity index was defined as the total number of discharges of HIV-infected individuals per 10000 total hospital
discharges. The hospitals were classified into two groups based on AIDS familiarity. High AIDS familiarity was
defined as greater than or equal to 30
HIV-related discharges per 10 000 total
hospital discharges; low AIDS familiarity was defined as less than 30 HIVrelated discharges per 10 000 total hospital discharges. The results of all
comparisons were not significantly different when different cutoff points for
AIDS familiarity were used. Additionally, for comparison we alternatively
defined AIDS familiarity as the total
number of discharges of HIV-infected
individuals at each hospital without ad
justing for discharge volume, and the
natural high-familiarity vs low-familiarity cutoff point was unchanged.
}
ft:
Data Analysis
The individual patient was the unit of
analysis. A logistic regression model
was adapted from the mortality model
of DuBois and colleagues.' The outcome
was in-hospital mortality; the predictors were insunmce status, intensive
care unit and bronchoscopy use, patient
origin, and hospital-specific features
such as AIDS familiarity and number of
beds. We retained independent variables when t values were adequately
large (P^.10). In cases that involved
simple comparisons between two
groups, unpaired t tests (two tailed)
were used on continuous outcome variables, and Fishers Exact Ttst (two
tailed) was used for dichotomous outcome variables. When comparing mortality rates at groups of hospitals, individual hospital mortality rates were
weighted by the hospitals' total number
of discharges during the observation period. Relative risk estimates and %7r
confidence intervals at low- and highAIDS-familiurity hospitals were derived from the final logistic regression
model.
RESULTS
The If) hospitals varied in capacity,
location, and amount of teaching (Table
1). All hospitals were located in melropolitan areas in California. Occupancy
rates varied from W/r tn !KI',;. The sample included both small hospitals (-. Uin
beds) and large hospitals (>r>(Ki beds).
Tital numbers of yearly discharges
ranged from a low of CilXH) in (lie small
hospitals to a high of 'l.> (MK) in the largr
MntUilily m i d A I D o n<iliitLMl I'CP - l l o n n . i t l H j l
>
I
�Table 2.-Charac1erlslicj ol Survivors vs Nonsurvivors*
Dmervnca
(SurvlvoraNonaurvlvora)
± 1 SE
Nonaurvlvora
(N = 3«)
Survlvore
(N = 218)
38
37.9 ± 9 1
ST
-
9.5
P
0 9 i 1.6
NS
NS
156
-21.7t:85
•01
56 4
16.5i8 5
04
- 2 3 8 - 83
963
92.3
40r45
88.0
67.0
21.0-7.8
01
183
Concomiianl lumor/Hilaetior. %
0.216.3
58.9
41.0
•AIDS-ax parlance nospilal. %
15 4
37.2
72.9
AOminw) via amargancy depanmantmanslei, %
33.3
-15.0; 8 0
•: 01
26.2
33.3
-7.1 r B 1
NS
Oi
NS
'NS Indlcalss not significant; and AIDS, acquired immunodeliclency syndrome
Table 3.-Characteristics of Patients in H i g h - and Low-AIDS-Expenence Hospitals'
Charactsrlatlc
Htgh-AIDS-Famlllarlty
Hoapltal
Low-AIDS-Famlllarlty
Hoapltal
Dltterence
(High-Low)
H SE
P
7
8
Total No ol HIV-related dlscnarges
816
99
Total No. ol patterns with PCP
221
38
2.8
28
-0.2i0.4
NS
Mortality rale. X (patients with AIDS-related PCP only)
12.2
33.3
-21.1 : B . 2
<.01
Intonsiv* care unit use rite, %
21.3
19.4
1.9±7.1
NS
Emergency Oepartment/traneler admission, %
394
47.2
-7.8:8.9
NS
Any premarker admissions. %
Concomitant mieenonrtumor, %
21.3
16.7
4.6:6.8
NS
24.9
41.7
-18.8:87
Average length ol stay lor eurvrvore r i SD, d
13.6:9.5
14.9:11.3
-1.3:2.0
NS
Average length ol s u y lor all patients r 1 SD, d
No. ol hoapnals
Mortality rate. % (all medical patients)
04
13.5196
154^106
- 1 9:1.9
NS
Broncftosropy rate, %
71.5
639
7.8:86
NS
MediCal redplert. %
13.8
27.8
-14.2:7.8
38 3 ?9.2
36 3 ? 7.9
2.0:1.5
97.3
88.1
11.2:5.9
Mean ; SD age, y
Male, S
.04
NS
.01
•AIDS Indicates acquired Immunodeliclency syndrome; HIV, human Immunodeliclency virus; PCP, Pneomocyslis cart™ pneumonia; and NS, not slgnileant.
4
est hospital, (lb avoid identification of
hospital with those of patients who did
the hospitals, all hospital-level characnot. These analyses were done both
teristics are reported as rounded figwith the smaller group of patients who
ures.) AIDS-familiarity indexes ranged
had PCP diagnosed by bronchoscopy
from 4 to 110 HIV-related discharges
and the larger group, which also inper 10000 hospital discharges. Eight
cluded patients with PCP who were dihospitals had AIDS-familiarity indexes
agnosed by other methods (sputum cyless than 30. The average overall 1986
tology or open lung biopsy or based on
mortality rate for all diseases at the hosclinical impression). Because the results
pitals was 2.7%, with no significant difof these analyses were not substantially
ference in average overall in-hospital
different when either the smaller or
mortality rates between high- and
larger group was used, the lorger group
low-AIDS-familiarity hospitals. The
was analyzed for this study.
level of teaching in the high--AIDS-faThe major patient-level variables asmiliarity hospitals was somewhat highsociated with in-hospital mortality were
er than that in the low-familiarity faciliintensive care unit use, admission of the
ties, but not significantly so. Only one
patient from an emergency department
hospital in each group was a major
or through on interhospital transfer,
teaching affiliate of a medical school.
history of premarker hospitalizations,
From these 16 hospitals, 200 000 paand omission of bronchoscopy from the
tient discharges were screened and 257
diagnostic evaluation (Table 2). Admisinitial episodes of PCP at individual hossion of the patient from un emergency
pitnls were identified. Each episode was
department or through an interhospital
Hssociated with u '12-)-l httcntalioiial
transfer and a history of premarker adCliisitifirntuni uf hinmnr.H — Ninth Re- missions reflected severity of illness.
nminii code (HIV disease). We comPatients admitted through an emergenpared values of independent variables
cy department or through an interhoin the group of patients who died in the
spital transfer are likely to be more seJAMA. May :>fi. 1 9 U ' J - V o l 2111. No 20
verely ill. Consequently, it is not
surprising that a higher proportion of
these individuals died in the hospital.
Similarly, individuals with a history of
premarker admissions are likely to have
major impairment in their immune systems and/or serious lifelong infections
such as cytomegalovirus infection and,
therefore, are less likely to survive the
hospitalization. Nonsignificant patientlevel variables included age, sex, insurance status, and presence of a concomitant infection or tumor. Most of the
tumors were Kaposi's sarcoma; the concomitant infections were mainly due to
herpes or Candida. These complications reflected dysfunction of the immune system but were unlikely to affect
mortality for the marker admission.
The only significant hospital-level variable was AIDS familiarity.
We compared the group of patients
treated at high-AlDS-familiarity hospitals with the group treated al
low-AIDS-familiarity hospitals (Table
3). Patients treated in hospitals with
low AIDS familiarity were substantially more likely to die in the hospital tv
Moflnlily (incl AIDS n n l a l o d P C P - Uunnoll ol al
2077
�i;
tions, usually for treat mem of ther-v infectious processes. Similarly, uidividCoetllclent
SE
P
Pr»dlr.lo'
uals who were admitted from MI
0.47
- 1.67
••- 01
High acQ:iiri>d irrniunodrlicwniY syndrome lomilianly
emergency department or through an
02
039
0.90
Emergoncv dopartmenilranslei admission
interhospital transfer were likely i.i
• .01
1 43
043
Ever in intensive care unit
have had a more guarded immediate
0.44
• .01
1.28
Ever with a premarver admission
prognosis compared with individuals
NSt
Concomrtam mlectionrtumor
0.00
Ot?
with electively scheduled admissio:
NS
108
0.S9
Mea.Cal
This severity factor was associal ed with
-139
0.46
in-hospital mortality in the model ny
Imercepi
DuRois et al* and also was significant in
•Vhe oonaspooding ordinary laasl-squares regression wrtn the same dichotomous outcome (in-hospilal death vs
our model.
suiMval) had an R' o." I S ' * . « ' values with dichotomous outcomes tend to be low even when ditleronces ol large
clinical signilicance are present because dichotomous outcome scales are relatively inlormation poor in relation to
Although the presence of concomiconlmuous oulcomes »
tant infections or tumors was included
tNS indicates not significant al the 5% level
as a severity of illness marker, it was
not significantly associated with in-hosnot been explored in previous studies.
lative to those treated at high-AIDSpital mortality after other severity facWe used longitudinal computerized disfamiliarity hospitals (33% vs 12%,
charge data from 15 hospitals from 1980 tors were accounted for. The majority of
P= .04). Despite the markedly different
individuals with these severity markers
and 1987. These hospitals were located
mortalityrotesin the two groups, markhad oral candidiasis or hen>es infections
in metropolitan areas in Califomia and
ers of severity of illness were not signifi(in contrast to the more severe infecvary in characteristics related to teachcantly different except for concomitant
tions associated with premarker admising affiliation, size, occupancy rate, and
infection or tumor. Use of the hospital
sions). Although herpes or candidiasis
number of yearly discharges. Overall
intensive care unit, patient length of
mortality rates (AIDS plus non-AIDS) likely reflected the presence of an alstay, and age were not significantly diftered immune system, these factors
were not significantly different on averferent in the two groups of hospitals.
were unlikely to affect mortality for the
age in the high- and low-AIDS-familDiagnostic evaluations seemed to be
marker admission.
iarity hospitals. Our data are consistent
similar in the two groups, as measured
Omission of bronchoscopy also was
with previous data in the overall obby proportion of patients whose diagnoassociated with in-hospital mortality in
served mortality rate in individuals
sis was made by bronchoscopy. Patients
the univariate analyses and might be a
with AIDS-related PCP infection
cared for at low-AIDS-experience hosmarker for individuals who have had
(16%).*" High mortality rates were sigpitals were more likely to be covered by
additional pulmonary infections that
nificantly related to low AIDS experithe state Medicaid program relative to
were not diagnosed or treated. Infecence of a hospital, use of the intensive
other insurance carriers (28% vs 14%,
tions such as cytomegalovirus pneumocare unit, a history of hospitalizations,
P= .0-1), to have a concomitant infection
nia are not uncommon in individuals
and admission of the patient from an
or tumor (42% vs 25%, P= .04), and to
with HIV infection and might go undiemergency department or as an interbe female (14% vs 3%, P= .01).
agnosed without bronchoscopy.'
hospital transfer. The markedly higher
When logistic regre: sion was used to
The limitations of our data set must
in-hospital mortality rates noted for papredict in-hospital mortality (Table 4),
be considered. The data allow us to evaltienU treated at low-AIDS-experience
several significant factors were identiuate only those diagnostic strategies
hospitals relative to high-AIDS-experified. These included one hospital-level
that included a procedure such as bn t\ence hospitals persisted, even after advariable (AIDS-familiarity index) and
choscopy. It is likely that mi.st of the
justment for severity of illness.
three patient-level variables (use of the
patients who did not undergo ImmchosSeverity of illness in patients with
intensive care unit, origin of patient
AIDS and PCP is likely to reflect both enpy were presumptively treated on the
from the emergency department or
the severity of the underlying immuno- basis of clinical and laboratnry findings.
through an interhospital transfer, and
Detailed chart review would lie necescompromised state as well as severity
history of premarker admissions). As in
sary to confirm the diagnosis nf I ' l T in
directly related to the ongoing PCP inthe univariate analysis, these patientthese patients.
fection. Our severity markers included
and hospital-level variables were signifu history of hospitalizations, admission
An additional limitation of claims data
icantly associated with a poor outcome
from an emergency department or as an
stems from the omission of risk group
durini; the marker admission. The presinterhospital transfer, and the presence
for HIV infection. This limitation is likeence of a concomitant infection or tumor
of n concomitant infection or tumor. Prily to be more important on the Mast
was not predictive of d'-ath after the
or hospitnliiations were associated with
Coast, where intravenous drug abusers
other severity factors had been taken
make up a large proportion of the AIDS
into account. The logisticregressionin- severe life-threatening infections such
as cytomegalovirus infection, choriorepopulation; in California, more than
dicates that, after adjusting for severtiniti", or intracerebral toxoplasmosis.
W'/, of the AIDS population are homnity, the chances of dying in the hospital
These infections are difficult to eradisextml or bisexual men." Further, the
al a low-AIDS-familiarity facility are
cate and would be expected tu occur
15 hospitals were private, with an overabout 3.1) times higher than at a highmore frequently in individuals with a
all MediCul rate of 15'A . It is even less
AIDS-fumiliarity hospital (O.W. confiseverely compromised immune system.
likely that large numbers of intravenous
dence limits, l.iHoC.H).
Kales et ul" noted that the presence of drug abuser.; were included in these
COMMENT
these infeclinns wits significantly assohospitals. The relationship tinted in our
ciated with m-!ios()ital mortality in pastudy should be evaluated in ntlier risk
We evaluated the associalimi nf IIUMtients with AIDS who were hospitalized
groups mid in other geographic areas,
pital AIDS familiarity with iii-hi>s|iit.il
with PCP infection. This is consistent
particularly with res|>ect to intnivenuiis
mortality rates of patients with 11IVwith our finding of markedly higher inrelaterl I ' d ' infection while controlling
drug abusers.
fur differcnrcK in patient and hospital
hospitul mortality ralei: assnciated with
Our model explains only a part ol t he
cliuracteriHtics. This lelalionship has
pa' ients who havi- had prior hospitali/ji- total variability in in-lmspilal timi talily.
TaWe 4 -t.cxy.s'.ic f<c<)ies^'on ol In-Hospiial MoilaMy on Palienl- and Hospilai Lovt! Variables'
207B
JAMA. M.iy I'O. lUM'J- Vol ;>0I.N(i
L'O
Mrill.ility . m i l All 1'; h e l . i l i n l I'i : l ' Mm
j
4
i-
I
•Cs.
I
I
iv--'
ir/-'
�ilnexpluin'-'d variability might reflect
Vu'.neasured differCfiices in severity of
il-i-s? '.'i ir-accuraey of the Uniform
hiv>l.;».!-i 'J:sr .fa cre Data Set. In pail.-.'ts •••••::! AIM-.', additional severity
:'iiciors :ie .-iiiularly recorded clinical
data (slveee.--to-arterial gradient for
oxygen, fivi-rt'ssed carbon monoxide
diffusioii cap.' i:,-. , lactic dehydrogenase
levels, and hemoglobin levels at the
time of admission). " ' These variables, in addition to more general approaches to case-mix adjustments,"""
are not included in this study; we plan to
report on current work including them
at a later date. More detailed severity
adjustment would be expected to account for some but not all of the large (21
percentage points) discrepancy in mortality rates between the high- and
low-AIDS-experience facilities.
Previous studies have addressed issues regarding the experiencj level of
the hospital related to procedures such
as coronary artery bypass graft surgery, vascular surgery, and transurethral resection of the prostate. Patients
who undergo these procedures are more
likely to survive when treated at highvolume hospitals. Possible explanatory
factors include skills acquired by operative and postoperative staff through experience over time, use of complex medical equipment during surgery, or other
characteristics. However, few studies
have noted an effect of hospital experience in nonsurgical diseases. Kelly and
Hellinger' note that patients who suffer
an acute myocardial infarction are more
likely to survive when their attending
physicians treat high volumes of patients with heart attacks; however, /io.«pifaf volume was not a significant factor
in their analysis. During the present
era, in which diagnostic and therapeutic
,
>
r
-
11
11
1
strategies for PCP pneumonia are becoming standardized, familiarity with
the disease might be required to facilitate timely medical decisions. In contrast to the findings of Kelly and Hellinger, we find that hospital familiarity
might be an important factor in determining in-hospital mortality for patients with AIDS.
Ar. alternative explanation for the
finding of a hospital volume-outcome
relationship is that informal referral
patterns have been established. These
selectively channel patients into hospitals that are associated with better
outcomes.' However, with respect to
AIDS, this seems unlikely. There was
no significant difference in interhospital
transfers into or out of high-AIDSfamiliarity hospitals relative to lowAIDS-familiarity hospitals. Additionally, severity of illness markers were not
significantly different between highand low-volume AIDS hospitals. If selective channeling were occurring, differences in the proportions of patients
transferred into or out of high-volume
hospitals as well as differences in severity of illness markers would be
expected.
AIDS cases currently are highly clustered in coastal cities and within certain
hospitals in these areas. However, in
the next 5 years, many cases will occur
in cities that currently have few patients with AIDS. If the results reported in this study are generalizable to all
parts of the country, then policies of
federal, state, and local governments
and decisions by hospital administrators will have a large impact on in-hospital mortality for patients with AIDS,
many of whom will be cared for in hospitals that currently are unfamiliar with
AIDS.
5
Caution should be used in generalizing to the nation from our findings,
which are based on a relatively small
number of hospitals and patients in only
one area of the country. Tb set responsible national policy, our work should be
validated using larger, more geographically diverse samples of hospitals and
patients with AIDS. But the striking
difference we found in mortality between high- and low-AIDS-experience
facilities suggests that |>olicymakers
concerned with hospital AIDS treatment should consider steps to alter the
natural diffusion of AIDS care at hospitals with little AIDS experience. For
patients with AIDS, the choice of hospital might be important. It is likely that
technology diffusion related to AIDS is
not uniform and that hosj itals with low
AIDS familiarity are at a marked disadvantage in the treatment of patients
with AIDS. Consequently, policymakers need to consider (1) forming regionalized AIDS centers, (2) providing intense educational efforts directed at
low-familiarity hospitals, or (3) rapidly
increasing AIDS experience for low-familiarity hospitals accompanied by close
monitoring of quality of care. These hospitals might benefit from "learning by
doing." By promoting the rapid diffusion of AIDS care to many hospitals,
improved survival rates as well as improved access to care might occur. Our
results might help govemment and hospital officials plan for the future, when
AIDS will be an important health problem in every major city in the country.
The authoisi thank Jack Delloviti.. MI). Allien P.
William.-.. Jr. MI), liohcrt H. Hronk. MI). Sell.
Steven A. Milen, MD, David Pranove. PhD. IlonaldT. MitKuyasu. MD. and AndrewSaxun. Ml), fur
their editorial commenlft.
ReUrencot
1. Clnft AK, Chirtr*in K. UmlOHtTun SH. T-vatment uf inrcctiohri MH.tociuU'<l with human immunudefirirncy vinia. A' tStigl J Med. l!)m:3t8:U:i»M4H.
2. Luft IIS. Hunker JI', Enthoven AS. Should operntinn* ht reKinnali"^' the empiricul relation bvIween aurtorst! volume and mortality. S Engl J
Med. 197a;a01:1364-13C9.
3. Kelly JV, Hellinger FJ. Heart dinea** and honpital deathti: an empirical - ^udy. Health .Sen' fUn.
issT^ies-ags.
4. Luft US, Hunt SS. Maerki SC. The volumeoutcome relationship: prarticc-make.t-iK-rfefl or
aeleclive referral patlrrnn. Health Sen' Het
19M7;2i!:157-IK2
5. Maerki SC. Lull IIS, Hunt SS. Seli'dinu caleifohen of |iulienLri fur reffinnuliutinn: imjiliciitionj; .tf
the rt-lhtmnnhiptitlwtM'n volume anil nuUtime. Alril
Car, |11H.;:24:14M.|W(.
fi. Shortcll S, Lolierfu J. HiiH|iiliil and Htaff tiryunizatinn anil ((uulily of rare: n-MtiitA for riiyontnliul
inrarctinn and hpiK'Udi'domv. Stett t'urr I'.iMl-.lii:
1041-lllM
7. Wolfe It, Hoi 1.1), Horn Jl). -t ul Mortuhly
dlfferenren anil wuuni) clonure uinmi^ i.|M'riu)ireil
burn fki-ililien. JAMA. lfiH:ii. .ii:7ii,l-1MI.
r
JAMA. May 20. 19(i'J - Vol L'G I. No I'O
K. Dulloin KW. Brook KH, I!o er» WH. Adjusted
himpital death ruten: a iKitential acreen for qualilv
of medical rare. An: J /lifclic Health. l(tS7;77: lll'.:'I1C7.
9. Armritronjf D. Gold JWM. Dryannki T, et al.
Treatment of infeclinnN in patients with the acquimt immunodeliciencv ayndrome. Aim /nlrrx
Med. 1885:103:738-743. '
10. Leone GS. Milla J. Hopewell PC, et al Papaone-tnmethoprim for Pneiunnryatin rnnnit pneumonia in the acquired immur.odencienry rmidrume.
Ann Inteni Med. ISHfi; 10^:45-18
11. Wharton JM. Coleman Ul., Wufay CM. el al.
Trimethoprim-BUlfomelhoxaiole or |>entuniidme
fur /'NriiiNM-yafia rarinii pneumonia in the acquireil immunoc. ( 'Tncy nymlrome: a pnmpeclivi'/
K
r,
riili'lMfhirt-d triiil.
"tern .Mnl. lllSl'i;|ll. ,::l7-44.
12. KulviC.I'. M'
T'.rre* It A. et al. Karly
pn-diclor*, ' in *'
rlallty fur I'liruninryn.
fin r n n n i i tme
Hiin-d iinniun'Mleficieri'-v MVildronie.
led 1<IK7;M7:I-Ii:i.
1117.'
13. Kiner K. H.HlriKue/J. M.-ll.ilhm.l GK. el al A
Qtnttititnlirc A " ' l f f / . f i x " f A l l t S tn Cnlitnrnut. Syrninii'iilo: r.nlifoniiii Depnrtrnt'til of l l e n l l l i S I T \ii-cn, 1:IH7.
r
I I . Grant IH. ArmHtronK* D. Management nf infertioui* complicalinn* in the ai'(|uin-il immuniHliTiciencv KViidrume Am J Med. l'.IHI'i:Sl-.Vj 7^.
15. Siover DK, While DA, Itomanu PA. et ul
S|H'Clrum of pulmonary diHeaaen iwHociutci) with
the acquired immune ileficiencv. Am J Mnl.
19Mr.-.7S:4L".»-437.
16. llrewhler AC, Karlm IlC. llvde I.A. .lae..!..CM. MKDISGKPS: a clinically ha.«ed upiimaeh to
claaiiifvinK hospiial patienu at admiHsion. /m/iiiry
19S. :a:377-3K7.
17. HomSD. Hum II A. Sharkey I'D Tin' -ev.-nty
nf illnean imlcx at, a neventy adiuntmrnl to diuv'""'"
lie-related Kruup - Health Care t'inanr lirr.
l.M.t:anniial.'<u|i|>l::i.'l-45.
IS. Knau* WA, Draper KA. Wat'in'r DP A ».-vi-r
ity of dinea-o' clii!"*ifirtttiiin nvi*li*iii. f 'nf ('nr.- Mnl
r
)
M
,
I'.isr.^-.i.Hls.SL-.i.
I!) Grei'iifielil S. Aronnw H . Kluhhiifr K M . Wi.ti.n
ahe D. I'lawr. in niortuhty IIHIII: th'- )iii/;ir.l-. of
iirnnrinK roniorlml iliKeiifi'^. J A M A . Wif.*;'- '" JJ-'1
:'ll. (irinlonT. Klinnel W. I l u l | i e n i i M I Y' dirliil.ilitv of roroniirv hearl diheuwe. J t ' h r n m r / ) M
i!l7ll.:!: :l:7--l-!li.
,
,
Moil.lhly and ;MU ..-f(iil,lleil I ' C I '
- Hi.'nii, |l ,:[ , , l
:
2379
iii
m
�Clinton Presidential Records
Digital Records Marker
This is not a presidential record. This is used as an administrative
marker by the William J. Clinton Presidential Library Staff.
This marker identifies the place of a tabbed divider. Given our
digitization capabilities, we are sometimes unable to adequately
scan such dividers. The title from the original document is
indicated below.
Divider Title:
)X
�Tab
This S e c t i o n
Includes:
12.
Bennett, Charles L., Adams, John e t a l .
" R e l a t i o n Between H o s p i t a l Experience
and I n - H o s p i t a l M o r t a l i t y f o r P a t i e n t s
w i t h AIDS R e l a t e d Pneumocystis c a r i n i i
Pneumonia: Experience from 3,126 Cases
i n New York C i t y i n 1987." JAIDS Volume
5, No.9
1992
�05/06/93
05/06/83
09:28
19:11
©202 233 3594
g410 635 7905
VACOCRS
MPBAUHDC LI BR
LIBTECHWLD142D
*•»- VACOCRS
©002/010
01002/010
J^uni^An^f^ /MMM/ Dtflttttef Syndromtt
iRivta h«l. Ltd.. Ww Yort
«
Relation Between Hospital Experience and In-Hospital
Mortality for Patients with AIDS-Related Pneumocystis
carinii Pneumonia: Experience from 3,126 Cases in
New York City in 1987
'ttCharles L. Bennett, •John Adams, *Paul Gertler, •RoIIa Edward Park,
••Stuart Oilman, 'Lance George, •fiMartin Shapiro, and •S'ftobert H. Brook
*Tht RAND Corporation, Santo Monica. California; ^Dfpartmtnt afhitdkin*, Dukt Unhtnlty, and tCtnttr
Htolth Policy Rett arch end Education, Duke Unlvtrtlry, Durham, North Carolina; fDtpartment (4 Medtcln
School of Mtdtctw, ^tpartmm of Health Strvlcti. UCLA School of Publtc Htalth, 'Vtttrant Affairs Mtdtcal
Centers at Wist Lot Angtlts, Lot Angeles, California; and '*tht Wtsttrn Region Special Studies Research C
of the Department of Veterans Affairs, Long Beach, California U.S.A.
Summuy: Thtrt ii marked debate about whether outcomei of care, particularty mortality, vary at a Ainction of hoipiul and phyiician expfrianct with a
dlteau. ThU litue U etpccially Important with reipeet to AIDS because
>200.000 Indlvlduali have n w bean diaftnottd wiih ihit diuai#. W« analyzed
o
ditcharii datafor3.126 penoni with AIDS w o hid Pneumocystis carinii
h
pneumonia and w o were treated at out of 73 New York City hoiphali in 1987.
h
In-hoipitil mortality wai 25*. Facton attoelated with higher chancet of
ihott-tenn death were older age, being black, not having private health iniur*
ance, and being severely lit. A loglitlc regression model Indicated that after
control]Ing for differences in patient and hospital chanctartstlci, the chancet
of death decreased when care wu given at hoipitali with higher caseloads of
patients with Pneumocystis carinii pneumonia. Ourflndingssuggest that hospital experience m y decrease mortality In this subset of patients with human
a
immunodeficiency vinii disease, although it is unknown whether thU is due to
differences in quality of can. Key Words: Hospital experience—In-hospiial
mortaJity-^Mi/maryi/ii carinii p e m na N w Yoric City.
nu o i — e
The most common reason for hospital admission
and one of the most common causes of death for
persons with AIDS is Pneumocystis carinii pneumonia (PCP) (1-7). Despite the high mortality rate
associated with PCP, patients can be successfully
treated when there ii early recognition of the disease, use of appropriate antimicrobial agents, and
close observation for signs of drug resistance and/or
drug toxicity (3,4,8). The hospital burden due to
Addrt convtpondence a d reprint raqueiu te Dr. C. L. AIDS is spread unevenly, with fewer than 5 of
n
%
B n et at Cuter Cr Health Policy Heteaieh a d Edueaiioa.
e nt
o
n
D k Unlvenlty Medical Cimer, Health Service* Returch a d hospitals nationally providing care for >50% of paue
a
Dmlopment (152). V.A. Medical Center, Durham. M 2770i. tients with human immunodeficiency virus (HIV)
C
U.S.A.
disease (9). There is considerable debate about
Reiults w r preiented in eart at t e Vtlth [ t m t m l Con-whether outcomei are better at hospitals that proee
h
ne ak a
firane* en AIDS In l»| In norence. Italy.
Mumtcrii* rc W d O t b r 31.1991; a c pe Jamwy g, 1 9 . vide can for Urge numbers of persons with AIDS.
t e e co e
ce td
92
w
956
�05/06/93
09:29
©202 233 3594
VCCS
AOR
LIBTECHWLD142D
©003/010
857
HOSPITAL EXPERIENCE AND IN-HOSPITAL MORTALITY IN AIDS PCP
lontini to the tame iodividual throufh the use of a multittep
matchlns procedure thai used all available Individual identifyini
InfKmatloa. The matchlnf process maintaint stroni conMentlallty prauctlon. and no patient ideadfien are present In the
woftln|dau.
W selected padenu ifsi 1S-63 yean w o bed a PCP admite
h
sloe between January 1.19S7 end December 31,1987 •nd who
b d ne prior PCP edmiislens In the 1 months befori the study
a
2
u*ln| the m t o described by Turner et al. (23). W started with
eh d
e
ifileof ill hospital admissions during 1 8 and 1 8 with any
97
98
dlifaoils of HIV (1CD-9 (H2~0M). Admiwioni during 1 8 with
97
toy eugnosU tf PCP (1CD-9 UU) were judpd to b "fint
e
admistioai" and were Included in this study If there wu no PCP
dlaanosit during the preceding 1 - o t period for thst pitlem.
2m nh
PaUcDt-tpedflc dau included age. sex, nee. admission sutu*
(routine versus mgent), hospital admitsioit site («rocrj«ncy d«.
partment, interhospital tnnifer. er home), insunnce covtrege
[private health buunjiee or not (including Medicaid and sslT-pay
la one tubgrowp)]. diiposilion oa dlschargt. specific resource
UM (c.f.. bronchoscopy or intubation), length of stay, outcome
of botpiialisalion (survival versus death), status of individud 3
u
days after admission (died or allva), priACipal dieinotii. secondary diagnosis (up to four), and payor. Inirtvcnou* drug use
ICD-9 o d s were n t included In analyses because of high leveli
oe
o
of underreporting le both dau uu. Death status was obtained
from two sources: (a) the SPARCS or FTF record w e the
hn
hosptullzatloe diseha/ze tutus for Ue marker or subsequtnl
t
admlsilon wu death, a d (b) N w York City and Sute monalliy
n
e
dau from the Department of Health death certlflcau flit. Sutus
at 3 diys after tdmlisloa wu evalleblt In 78* of th* pstlentt.
0
W derived varisblcs repressntlnv the n m e of prior hospitale
u br
iiaUotu within the past 1 moflths aad n m e or non-PCP di2
u br
agnoses coded la Uu marker admlstion (up to four additional
diagnoses were available in tbe dau set).
Severily-of-Ulness variables were derived based on the ichemt
O Tuner et el.torclassifying AIDS patients Into I of 2 dit|f
0
nestle stages (23). There are three mijor diseaic stages: siage l
(longest survival and least morbidity): stags 2 (reduced survival):
and stag* 3 (poor survival and significant morbidity). The tugei
tor patienu with PCP include stage 2.1 (PCP and nonlnftctlou*
edmpllcatlon* such a* Kapo*!'* tarcoma): 2.2 (PCP only): 2.4
(PCP and a noo-AlDS-deflnJn* Infection); stage 2.3 (mor* than
o e opportunliUc Infection): suae 3.0 (non-AlDS-dtHnlni CNS
n
InfecUon): ittge 3.1 (HIV eAcephalopethy); •mgc 1.2 (more then
o e AJD&deflalng lafbctlon and one or men non-AlDS-dcflning
n
iafectiM): stage 3.3 (ttptis): and (Uge 3.4 (organ failure;. Su«et
3.1,3.2, end 3.3 wen collapsed together in the current analysis
because the isdivlduil cells conulned small numbers of individuals aed the subgroups win clinically related.
Comorbidity variables were derived using the m t o of She.
eh d
plro et al. (24). The up-to-fbur non-PCP diagnosesrecordedfor
each admlssloa wen cliuslfled into the 1 "analytic" cetetoriet
7
that the Health Care Financing Adml&lstratloa (HCFA) uses to
METHODS
classify ICD-9 principal diagnoses by atortality me (23): 1
7
d m y variables Indicated the pretence or absence of each of
u m
Patient-Spedfk Data
these ettegoriet a o g tbe recorded noo-PCP diagnoses, W
mn
c
The pitieni-lcvcl data baie tor thli study wu obtained from alio created a modified set of comorbidity variables by excluding
from the 1 HCFA analytic catagorles those diagnoses for which
7
two tourcei: (*) i modjfled versien ef the UaUbrm Hospital
*
Dliehnnie Dau Set trom 70 hoipitali located in tbeflvaboreufhi w believed then wu mora than a trivial poulbllliy thai the
of N w York City (obtalAed from tbe Statewide Plannlna and dlagitoti* was a ceitsequenee of poor care (such u cardiac are
nit).
Rcuarcb Cooperative Syium (SPARCS) maintained by the
N w York State Department af Health), and (fa)talentTreate
m e Fllei (PTPi)torpatient* treated at Ue three Vctcrani Afet
t
fair* Medical Ceoten In New York City. Beth dau MU con.
HospUal-Spedfic Dau
ulned diicharee abstract record*torpatienU hotpitilUed ie
ihort-ttrm tcutc eare, or acute»car* unit* of lonf-term N w
e
HoipiuJ-level ehanetcrlstlcs wen defined wfin information
York hoipitali in 1 1 and 1987. Thete data teu were kmgitu* from the American Hoipiul Association Guide (26). Hospital
96
dlnalized mini methods eimflar te these ef Kaufman et al. (22). bed-sin cattgoriil wen <300, 300-399. 600-899. and 900 or
In brief, a ualqui caie number wu Milined te all records be- aion bedi. Teachleg chaneteristies wan cluiiflcd H mi^or
This debate has been particularly prominent in New
York City, which has the largest number of HIVinfected Individuals of any city, the (argett number
of HIV-infected individuals who are intravenous
drug users, and the largest number of penons with
AIDS who lack private health insurance (10-14).
A previous California study found mortality to be
lower for patients treated at hospitals with a larger
number of PCP cases (1.15). Similar volumeoutcome relationships also have been described for
specific surgical procedures such as open heart surgery, cholecystectomy, and transurethral resection
of the prostate (16-21). However, few studies have
identified similar relationships for nonsurgical diseases. The policy implications of volume-outcome
relationships depend on the functional form of the
relationship. If outcomes continue to improve as
caseload increases because of organizational factors such as "economies of scale" in which the
availability of shared resources helps the medical
staff achieve good patient outcomes, then the goal
should be to limit care to a very few centers. Patienu with liver disease undergoing organ transplantation typify this relationship. If there is a
threshold effect (hospitals that treat more than a
minimum number of persons with a given disease
have better outcomes), then the goal should be to
assure that patients are referred to the presumably
larger number of such hospitals. If hospitals with
larger caseloads have outcomes similar to hospitals
with smaller caseloads, no efforts should be made
lo concentrate care.
The current study was undertaken to determine
whether tbe volume-outcome relationship found for
AIDS-related PCP in California was also true for
patients in New York City and to evaluate whether
such a relationship, if it existed, was of the continuous threshold type,
Icum^tfAti^ni Imwu btlMtty iymimnu. Vol, S. No, >. / » ;
�05/06/93
05/08/93
858
09:30
10:13
©202 233 3594
©419 605 7905
VACOCRS
HDBAL*HDC LIBR
•- LIBTECHWLD142D
*>> VACOCRS
••
12)004/010
121004/010
C. L. BENNETT ET AL.
(priimryludilnikMpltil). minor (other tffilttud heiplul), end
n n (w>nt«ichln« hofpiUI). The prinwy t » » ^ l heipluU ere
oe
thou thet ere owned by er eloiely ded to medial tcheoll end
ire eentldered the "fiefihip" ueehlai hoipiuJ of Hut ta«Utu-
RESULTS
The 73 New York hoipitali varied in bed sire,
ownership, teaching status, and PCP caseload (Ta*
"Tn addition, for each hoipltal. an HIV-ipeeHIc characterjiUc. ble 1). Three were Veterans Affairs Medical CenPCP volume, wti defined from Infonnatloo In the SPARCS and ters, 1 were municipal hospitals, and the remaining
1
PTF nici. PCP volume wu dallnad u tbe total n m e efHIV- 39 were private hospitals. PCP volumes ranged
u br
infected indlvlduali with flret-epitoda PCP admitted lit IM7. In
t m uulytei. hoipitali were elattffled Into high-vohim- PCP from 1 to 355 cases of first-episode PCP admissions,
o e
hospitali if ibey treated »160 patient epIiodM la 1 7 and low- with almost two-thirds of the hospitals treating 50 or
M
volume PCP hoipitali if they treated <IM patient cpUodei in fewer patients.
1W7. Seniitlviiy antlyuitedlcatadthat ifatulw reiulu were obtained If PCP volume wu iplii at my level from 1 0 to 2 0 caiei Prom these 73 hospitals, 3,126 cases of first2
4
of PCP in 19*7,
episode PCP were identified. The average age of the
patienU was 37,1 years. Eighty-two percent of the
patients were male. Forty percent of the patients
Univariate Analyici
were white, 3 % were black, and the remaining 2 %
3
7
In eatet that Involved tlmple eomparlioni between two
were primarily Hispanic. Patients admitted to the
groupi. unpaired > teiu (two uiled) were used on eontlnuout hospital with PCP had an in-hospital mortality rate
viriablei and Fiiher'i Exact Tctt (two tailed) wu uud for eomof 2 % (Table 2). Using the Turner staging system,
5
parlioni between m r than two lubpoupt.
oe
almost one-fifth of the patients were categorized as
stage 3.0 or greater. The average length of hospitalLoglitlc Regreesion Models
ization was 24 days, 1% of the patients were Intu0
bated for respiratory failure, an intensive care unit
The individual patient wit the unit of ualytli. Two «eu ef
%
16
logiitie regratiion model! wire uiipted from our prior report, was used for almost 8 , and 6 9 underwent bronwhich uicd itmilar dau from 1 Califoraia hoipitali. In brief, tn choscopy.
5
the flrtt ttt. the outcome wu ln-hoipiul death end the predlc Patients treated at hospitals with >160 cases of
torn Included patient-level variablei (Iniurence ttetui, age, race,
aei. hoipiul admlitlon source, n m e er prior hotpluUeatieni PCP in 1987 differed from patients who received
u br
within the prior 1 monthi. th* modified HCFA c m r d y care at hospitals with <160 cases of PCP because
2
o oW H
variibl<>. and th* iii uverlty of lllneu itage* ef Turner et el. they were less likely to be categorized as stage 3.0
03). iml hoiplul-ipeclfk futures (PCP velum*, bad tizt. and
or greater in the Turner system, and more likely to
teaching «uiuil.
Becautc petlenti m y die ihortly after dlicharge from a b e
a
o*
PIUJ. logMIcragreulonmodcli bated on in-hoipital death might
TABLE t. Cktrtcteriiiia vfthw 73 hospilaU vuiltd
b biased in favor of boipltals that routinely dlicharge very ill
e
patienu. This pountial bias wu addreised by evaluating ilmflar
No. of
loglitlc resreuion modals with th* outcome variable being duth
hospitals
Within 30 days ef admission. To anew the robustness of the
Charactcrinie
logisticregreisienmodel* with respect to hospita] eaaeload e*
f
fwu. wc evaluited several models that differed in the eoding of Bed siu
1
7
0-299
the variable for hospital PCP velum*: caugorical (<I60 uses of
2
7
300-599
PCP/hoipiul or »160 PCPeatct). linear (with volume being repIS
600-899
resented by the abioluu n m e of cases of PCP in a given
u br
1
4
hoipiul in 1917). er quadratic. Finally, in a models, th* rune- >900
U
tlonal form for comorbidity variablei varied between ineluslon of Teaching itatui*
4
0
th* 1 comorbidity variables, the 1 modified comorbidity vari- Meier teaching affiliation
7
7
1
9
Minor teaching iffiHetion
ibln, or omission of all eomoitoidity variables. Because tbe re*
1
4
lulu were ilmflar Cr any choice about comorbidity variablei, w No uacblag efflliation
o
e
No. of flret-eoisorfe PCP cues In 1 8 for th*
97
repon only thereiulufromthe models that included the 1
7
modified comorbidity variablei.taaS analyses, patient and hot. gives hospital*
0-10
2
1
pital Idanttflen were oetiued.
11-30
2
3
31-100
7
101-130
»
tntrahotpltal Conelatian Analyses
131-200
6
>230
J
Intrehoipiul correlation effecu wen alto evaluaud. A simple
PCP, PHtumocytth r W t pfliumonls.
a ut
calculation ef ihe intrahoipiul correlation yielded a comlauon
value of 0.006. which eorretpondod to • design effect ef 13 o ' Teaching sutu* wu beaed on daureponedby th* Amsricin
. n
the SE teal* for hoipltal-level variablu. ThU did notresultin Aiioelatlon of Medical College* (ref. 26).
any marked change In lUtliUeal significance of the boiplul-level * PCP volume for a given hospital repreunts that toul n m e
u br
variables in all regruilon models. For simplicity, wa report tha of patient episode* of PCP treated in 1987,regardlessof whether
reiulu of tha standard logistic itgreitloe models.
it is a flrsMlme inftetiofl or not.
Jcknui ^At^nM I mnr Dtjktt*rt ipUnmti, V*/. 3, V*. 9,
m tt
�05/06/93
09:31
®202 233 3594
LIBTECHWLD142D
VCCS
AOR
©005/010
HOSPITAL EXPERIENCE AND IN-HOSPITAL MORTALITY IN AIDS PCP
8S9
TAB LB 1 Ckancttrittlct tfttlpulinlt and whttktr pettetu wutohigh- or lowvoliimt Pneumecyilli ctrinli pneumonlu
iKh hospital
Hlgh-PCP'VOlume
Low-PCP-volumt
All paUcnU
(n - 3,126)
23.2*
In>hoipltal maulhy rau'
MUA tgt (yrt) iSDr*
31.1 0.3)
Male*
82.6*
Rice/ethnic group'
40.6*
White
Slack
33.2*
Hlipenlctathir
26.2*
Staga »3.0 la itagtng lyitam of Turner et al. (23r
19.6*
Admitted emergently
70.7*
M a number of prior hotpltalltatloni (SPr
en
0X1.0)
Inturance provider'
62.5*
Gov't ,/telf.pey/oihcr
Rrivate
37.3*
PaiieoU treated in botpltali with teaching aflUlatloiu of'
s.g*
Nn
oe
Minor
22.1*
M*|or
72.1*
Hetpitti reiource ute
InuibtUufl'
10.4*
Bronchoscopy'
60.6*
1.0*
ICLT
24.1 (20.2)
Average length of stay (days) (SD)*
• High-volume PCP hoipltal treated >IW patient-episodes ef PCP In l»B7.
* Uw-voluma PCP hoipltal treated <I60 palieni-epiiodei of PCP In 1917.
'p<0.03.
*p • nontlgnlflcant at 0.03 level of tlgnlflr.tftce uaiat r teit.
-
hoipitali*
(a - 1.596)
hoipitali*
In - 1.528)
20.7*
37.2 (1.5)
69.0*
30.1*
36.918.1)
76.2*
31.9*
24J*
23.8*
70.1*
0.4 (0.9)
26.9*
42.9*
31.2*
22.7*
71.2*
0.3(1.1)
34.3*
43.7*
72.1*
27.9*
g.6*
0.0*
91,4*
2.7*
45.4*
51.9*
6.3*
64.6*
7.8*
24.0(19.3)
I2.3K
56.2*
8.4*
24.3 (21.1)
im
be male, have private hedth insurance, and be
mortality rate. Resource use differed between surwhite (Table 2). Patterns of resource use differed
vivors and nonsurvivors, with survivors having a
for patients treated in the two groups of hospitals, lower frequency of use of intensive care units (4
with patients at higher PCP-caseioad hospitals more versus 20, p < 0.001) and intubation (4 versus 3 %
0 .
likely to have undergone a diagnostic bronchoscopy p < 0.001), but having a higher rate of bronchos(65 versus 56%, p < 0.001), but less likely to have copy (64 versus 49%, p < 0.001). Hospital-level
been treated with intubation (8 versus 1 % p < characteristics associated with higher chances of in3 ,
0.001). Patients treated at hospitals with higher hospital death included having a teaching affiliation
caseloads of PCP were more likely to survive in the and having treated <160 patients with PCP in the
hospital relative to those with lower caseloads (21 prior year (Table 3).
versus 3 % p < 0.001).
0 .
Logistic regression analyses were used to assess
Patient and hospital-level variables associated
the odds of in-hospital death and death within 30
with higher In-hospital mortality Included soclode- days of admission after simultaneous adjustment for
mographic characteristics (being black, being feother factors (Table 4). With atyustmenl. receiving
male, not having private medical insurance, and
can at bigher-PCP caseload hospitals was signifiolder age), severity of illness (admission of the pa- cantly associated with lower odds of death. For extient from an emergency department), higher num- ample, patients treated at hospitals with 160-239
ber of prior hospitalizations in the past year, higher cases of PCP in 1 8 had an odds ratio of 0.64 for
97
severity of illness [as measured by the staging sys- in-hospital death relative to those with 80 or fewer
tem of Turner et al. (23)], and higher comorbidity eases of PCP ( 5 confidence interval. 0.46-O.88).
9%
(as measured by the 1 modified comorbidity vari- and for death at 30 days from admission the odds
7
ables). Patients with the lowest severity of illness in ratio was 0.51 ( 3 confidence Interval, 0.36-O.73).
9%
the Turner system (stage 2.1) had a 2 % in-hospital Older age, being black, being admitted emergently.
1
rtality rite, whereas the patients with the highest having a higher severity of illness in the Turner
rity of illness (stage 3.4) had a 6 % in-hospital staging system, and having had prior hospitaliza5
tfAturirtUmmm* DtflrtMCy Jj*****. VW. J. No. >. 1982
�05/06/93
05/06/93
09:31
ie:14
VACOCRS
HDBAL*HDC LIBR
© 2 0 2 233 3594
0 4 1 0 606 7906
- - - LIBTECHWLD142D
VACOCRS
©006/010
8008/010
C. L BENNETT ETAL,
TABLE J.
CharaeHrittUtofpeit**ttvlik]{rthap\to4i
Pneuroocynlc urlnJJ p*tumo*l» (Kh h whtthur tkty
turvivtd iktlr New York City hotptteitnth*
No. of
IndlvWiiAli
Pulenl-kvd chwwtcriitk
Stx'
Mill
F«md«
Rice*
Whhc
Blsck
Hiipwiic/othtr
Sute in tha tyaicm of Turner
ei
(«r
2.[
2.2
2.4
2.5
1.0
3.1-1.3
3.4
Admiulon iourea'
Emerteney department
Heme or phyiician ofTlce
Iniuramre provider*
Gov'Utcir-pay/olher
Privete
Patienti treated In hoipitali with PCP
cauload* of*
<80
80-139
l60-2)>
»340
Patients treated In hoipitali with
teaching leveli o f
None
Minor
Mt^or
Hoipltal resource uie
Intubated'
Not Intubated
ICUICU not uicd
Bronchoicoptd*
Not bronchoieopad
«
Sunrivlni
2.382
544
76.3
67J
1.268
1,038
820
10.4
67.7
74.9
295
1.047
953
221
16
330
264
794
63.7
79.1
73.1
50.0
63.9
34.7
2.223
903
72.3
80.7
1.954
1.172
7IJ
80.5
719
759
701
697
71.5
68.5
61.8
77.3
180
693
2.231
80.6
70.3
73.7
325
2,101
250
2.176
1.894
1,232
27.4
80.3
37J
78.0
79.3
67J
death as the dependent variable. The scatter points
represent actual in-hospital death rates for hospitals
with at least 30 first-episode PCP cases during 1987.
Hospitals with <30 cases are lumped together into
aggregates with at least 30 cases, and the average
death rate and PCP volume are plotted for the aggregates. The area of the plotted circles is proportions] to the number of PCP cases included in the
regressionforeach scatter point. Predicted probability of death as a function of hospital PCP volume
are graphically shown based on regression models,
which include representation for PCP caseloads as
either a continuous variable, a categorical variable,
or with both a linear and a quadratic term, whereas
the other patient-level and hospital-level variables
* p < 0.05 for eomparitan of survival ratal far the given chtr*
acterlsUc.
TABLE 4. OMl wlot of the adjust td tffect ofhoiptlul
Pneumocyitit eariali pneumonia (PCP) vottime end elinlcml
end demognphle etumcttrUtlrt on death In ptnons with
ftriheplsode PCP in 1917
VirlaMe
PitiscMavtJ •harutarlstUi
Aplyn)
11-44 yean
iS-JPyurt
«0-<4y«an
While
Black
Htipiak/etber
Femalste*
No prior admhi
One prior admit
Emerieiwy admit
Sivamy-er-tHneu itaae
2.1
2J
U
l.J
1.0
J.W.I
u
tions in the prior 12 months were slso usoeieted
with poorer odds of short-term survival (p < 0.005).
After adjustment for severity of Illness and other
patient and hospital-level factors, patients treated at
hospitals with a teaching affiliation did not have tugniflcantly different odds of short-term survival than
did those patienu treated at hospitals without a
teaching affiliation.
In Fig. I, we show the relationship between hospital PCP volume and in-hospital survival, based on
the logistic regression analysis with in-hospital
Journal tf Atentnd t
Oefktentf Syndrom*t, Vel,), No, $, IM
Privat* tafurane*
HoipluMeval chameurlukt
PCP caseload IN7
<80 eaaaa
10-139
ISO-SM
•246
Tcaehldi ifflMoa
Nona
Miner
Miter
Hug tal bed alia
1,200
Oddintle
Odds
far
death
ratio far
within 30
la-hoipiul
days of
duth
idmliiioa
{ft! 9 eMfidaiwe
1
(93* teefltfiBH
Interval)
taltrvaU
1.00
I.IKO.M-I.O)
1.27 CO.17-!.M)
4.l»a.ll-S.2l)*
1^0
1.14 (0.92-1,41)
1.22 (0.93-UI)
1.J9(I.70^.17)*
1.00
1.33 n.03-l.6»)»
1.10 (0.S6-M2)
1.13 (0.M-I.44)
\XO
1,10 (1.01-1.19J'
l.49(i.l)-|.IO*
1.00
I.23I0.9S-I.60)
I.M(0.II-1.M>
0.91 (0.71-1.17)
1,00
).l3(IJ0-2.»)'
1.42 (l.ll-1.7»)'
1X0
I.IO(0.7S-1.SS)
1.14 (0.79-1.«)
1.14 (043-1.11)
(1,20-10. i r
1.93 (l.JO-LKr
3.31 f3.70-t.4))»
OM WJ3-0.U)'
1.00
1.00 (0.6S.1.43)
0.97 (0.47-1.41)
0,M (0.3S-I.S5)
1.7a(0.37-S.3«)•
t.»(0,S5-i:96)
1 JO Q.51-3.76)0.70 (0.36-0.19)*
1.00
6.11 (0.69-1.14)
0,«4 (0.<4-0.Mr
0.67 (0.4M.W)r
1.00
0. MtO.te-i.D)
fiJI (0J^.73)»
0J6S (0.30-0.13)*
1.00
].20(0.7M.06)
1,57(0.94-2.62)
1.00
i.l7(0.U-J.O9)
1.34(0.77^.16)
1.00
CM (0.67-1.10)
1.00
1.W (O.TS-IJI)
All eomperisena had p values ihst wen net lUtlillcaJly dUTaranl fran
IA Si lh» p - 645 llsnlllCMM level unlaw aeeompaalad by i a ' .
�05/06/93
09:32
©202 233 3594
VCCS
AOR
LIBTECHWLD142D
©007/010
HOSPITAL EXPERIENCE AND IN-HOSPITAL MORTALITY IN AIDS PCP
861
3-
3-
o gJ'r-^-e^U^._
0
o
o
f"'-\
C
O
3100
800
-r
300
1
1
1
ttnn
rm
n
H
f ^' - I f . ^ ! i } . ^•«" ov» «
p a m na (PCP) UMJoidi u ot r a e n n o i viritblt (-^). citaqornu o i
tM o U u u
lo«i vtriabio •• •)• «> with b t a iiniar a d a quadratic t r ( -), Tha aeattar oolnu raoriatnt actual In-hoioitai t/aath
oh
n
om —
for hoipitali with atliar 3 ftrifaplioda P P euaa d rn 1 7 Hoipitali3<3 S S S S S S S m aggli!
0
C
ui g M ,
wt at least 3 oasas, a d tht avtragt d ah rata a d PCP v l m are po a for tha eggrtgitaa. Tht art. oflht plottttfilrdit
ih
0
n
at
n
ou e
lnd
r
ll proportional to tht numbtr of PCP oaaaa Included In tha wgroselOA for tact) tcitter pelnt.
are the same as those identified in Table 4. As
DISCUSSION
shown, the predicted probability of the in-bospltal
deathratedecreases in all three regression models
This study demonstrated that patients with HIVas the hospital volume of PCP cases increues to
related PCP were twice as likely to die in the hos1 0 cases of first-episode PCP in 1987. However, pital when they received eare at hospitals that
6
the figure also indicates that the data do not allow treated <I60 cases of first-episode PCP in 1987,
us to distinguish between a threshold volume of 1 0 after a4Justlng for differences in patient and hospi6
cases of PCP associated with good outcomes, con- tal-level characteristics. Based on logistic regresstantly decreasing chances of adverse outcomes as sion models, seven patient-level factors were also
volume increases, and chances of adverse outmortality: (a) having
comes that reach a minimum at -240 eases of first- associated with higher odds ofbefore admission, (b)
been hospitalized in tbe year
episode PCP and then increase at higher volumes. being admitted through an emergencyroom,(c)
Similar relationships between volume and outcome
syswere obtained when the outcome was death within having higher severity of illness In the stagingbeing
tem of Turner et al. (23), (d) being older, (e)
30 days of admission and the same patient-level
iharacteristics were included in the logistic regres- black, (f) not having private health insurance, and
(g) having more serious comorbid conditions at the
' n models (dau not shown).
time of hospitalization.
tfA*i*« Immmm Otfrifury ty^romii, Va/. i. No. 1.1992
�05/06/93 09:33
05/06/93 10:15
862
®202 233 3594
0410 605 7905
VCCS
AOR
— LIBTECHWLD142D
H B U D L1BR
DAHC
VCCS
AOR
©008/010
8608/010
C. L. BENNETT ET AL.
Our resuJu ut in ngreement with prior studici in
that increasing age and being black were associated
with greater short-term mortality for PCP (1,5,6).
Hispanics did not have significantly higher mortality rates than whites did in this study and in the
reports of both Lemp et al. and Harris (5,6). Patients admitted through an emergency department
are likely to be more severely ill. Consequently, it is
not surprising that a higher proportion of these individuals died in the hospital. Similarly, Indlviduais
with admissions in the year before the development
of PCP are likely to have major impairment in their
immune system and/or serious Infections such as
cytomegalovinis infection, and would be less likely
to survive the hospitalization for PCP. Our findings
of a threefold difference in mortality between subjects In New York City in 1 8 with first-episode
97
PCP in the lowest stage (21% in-hospital mortality
rate) and highest stage (64% in-hospital mortality
rate) in Turner's classification scheme are very similar to the finding of an 1% in-hospital mortality for
8
patients in New York City in 1985 with the lowest
stage and 6 % for patienu with the highest stage
0
f23.27).
Kelly et al. have evaluated the relationship between hospital characteristics and in-hospital mortality rates for patients with AIDS in New York
(27). In their study and in the current study, patients
who received care at hospitals without teaching affiliations had lower in-hospital mortality rates.
However, in this study, after adjustment for other
patient-level and hospllaJ-level characteristics, the
odds of death did not vary significantly as a function
of the level of hospital teaching affiliatioa.
Results from the current study and the previous
analysis of data from Califomia for the same year
(I) are qualitatively similar but quantitatively quite
different. In the Califomia study, the contrast was
between hospitals with <~12 first-episode PCP patients (33% mortality) and hospitals with >I2 (12%
mortality); none of the hospitals studied in California had > 100 first-episode PCP patients. In New
York, hospitals with <I60 first-episode PCP pg.
tients had 3 % mortality, and hospitals with >160
0
had 2 % mortality.
1
Some factors that may have contributed (and that
should be explored) deserve note. First, it is likely,
given the history of the epidemic, that some hospitals in California made an eariy commitment to providing high-quality care to patients with HTV disease, and even sought admissions. In an environment where the caseload was not large at that time,
tfAcqulrtd i
v Dtflrfracy *y«*«MM. Ve/. S, Nt. 9, /Ml
that may have m d an Important difference, beae
cause hospitals with little interest in the disease
may also have avoided large numbers of admissions
as a result.
In New Yoric, there were almost no hospitals
with very small numbers of eases. At the same time.
New Yoric City health care institutions have been
strapped for resources. Improving expertise of the
nursing staff and categorical units may come into
play as factors associated with improved outcomes
at high-caseload hospitals. Second, New York City
hospitals tend to be much larger than Califomia
hospitals and their larger staff would require more
PCP patienu to gain the same average experience
per staff member. Third, the relative importance of
Intravenous drug use is greater in New York, which
may affect the efficiency of various intervention efforts.
Although our study provided additional empirical
evidence in support of care for persons with HIV
disease at hospitals in New York City with large numbers of AIDS patients, we were unable to find any
empirical evidence to support the policy option of a
Single-disease hospital. The Task Force on Single
Disease Hospitals in New York City debated the
same policy options (10) and concluded that the best
strategy would be to mainstream individuals with HIV
disease Into exiating health care facilities. They believed that single-disease hospitals would have difficulty attracting patients, would be vulnerable to budget cuts, and would have difficulties with staff recruitment. The Task Force thought that a special hospital
for HIV Infection was an "ill-conceived solution [that
was likely to] quickly become part of the problem
(10)" and proposed other options such as designated
AIDS uniu within hospitals.
in contrast, Weinberg and Murray had proposed
in 1987 that New York dedicate one large hospital
to the treatment of HIV-infected individuals, believing that otherwise "the care for other acute lifethreatening Illnesses could be profoundly compromised (12)." The Mayor's Task Force, assembled
in 1989, proposed a similar solution—the development of two 200-patlent subacute care facilities exclusively for the treatment of HIV infection (11), In
this study, our analyses indicated that the 3,126 patienu withfirst-episodePCP had better odds of survival when treated at hospitals with higher PCP volumes. However, we were unable tofindany evidence that mortality continuously improves with
increasing volume, the presumption that underlies a
single-disease AIDS facility.
�05/06/93
09:34
©202 233 3594
LIBTECHWLD142D
VCCS
AOR
©009/010
HOSPITAL EXPERIENCE AND IN-HOSPITAL MORTALITY IN AIDS PCP
86J
An important limitation of our data ii tha lack of tients, of the impact on accessibility of hospital care
clinical and laboratory information needed to iden- for patients, and of the actual outcomes of hospitaltify accurately the severity of illnen at the time of ization under such a policy.
hospital admission, These variables Include alveoAeknowledgmanti We acknowledge support from the
larto-arterial gradient for oxygen, carbon monoxide dlffuilon capacity, lactic dehydrogenase levels, Agency for Health Care Policy and Research
Western Region Special Studies
and hemoglobin levels at the time of admission (7. ()ROlHS06494-0l) theAdministration, and Ihe New York
Group of the Veterans
2&-30). Il also would be helpful to have data on risk City Health and Hospitals Corporation fConUci CAMgroup membership, which are not reliably recorded 733). Dr. Bennett Is a recipient of a Career Development
Awardfromthe Department of Health Service* Research
in diicharge abstracts. Finally, only 7 % of cases
8
had data on discharge itatui 30 days after admis- of the Veterans Administration. We acknowledge the edsion. If high-volume hospitals discharged patients itorial and statistical comments of Robert Elathofl*. Barwho died out of the hospital, but who were not re- bara J. Turner, Steven Miles, Jack A. DeHovitz. Phyllis
A. Guza, and Ronald Mltsuyasu. Joan Keesey and Russri
corded In the dau set, this may account for some of Bennett provided invaluable assistance with creation of
the differences In outcome. Chart review studies
dau analyticflies.Nancy Elwood provided imponant edare currently under way to address these questions. itorial and secretarial atsiitance in the preparation of the
manuscript.
However, if the high mortality rates that we observed at low-volume hospitals are confirmed by
clinical studies to be due to lower quality of care,
REFERENCES
our dau suggest that policy options other than a
1. Jeiuell CL. Gaifinklc JB. Crecnfleld S. et al. The relaiiun
large central hospital for AIDS patients should be
betwece hoipltal experience and ln*hoipital aionaliiy for
considered (1.15,31). These options Include initiapatienu with AlDS-rclaled PCP. MMA 19*9.261:297*4.
tion of educational and organizational interventions 2. Undctman SH. Qlnzberi H. Welti SH. The AIDS epidemic. N Enfl J liti 1»3 J12:32l-J.
at low-volume hospitals, increasing the number of
3. Oram IH, Armitronf D. M n g m n of Infeclloui eompll
a ae e t
patients at low-volume hospitals, and/or transfereatloni ia the eeeuircd immunodcnclency lyndroma. A J
m
ring patients at low-volume hospitals to nearby hos- Mid l966;81:S»-72.
4. Amitroni D, Q M JWM, Dryanikl T. et al. Traaimcm of
o
pitals with higher volumes of AIDS patients.
Infection* in patienu with tha acquired unmunodeficlcncv
lyndrome. Ann Intern Med I963;10J:718-I3.
An unanswered question about the volume5. Lemp OF. Payne SF, Neal D, Temelao T. Rutherford Cw.
outcome relationship for persons with AIDS is
Survival trendi for petlenti with AIDS. JAMA 1990:263:
whether high hospital volumes are due lo relbrrals
4(0-6.
6. Kanii JE. Improved ahon-lirm wrvlval of AIDS patlenm
to hospitals with good outcomes and reputations
Initially dlajnoicd with rntumocyttli carinii pneumonia.
(the selective referral hypothesis) or whether hos19S4 through 1967. JAMA 1990:263:397-401.
pitals with good outcomes have both learned and
7. Matbewi WC, Ftrdoa E. Bennett CL. Kanouse D. Evaluatlnt lutltutlonal performance In AIDS-aisociated PCP: a
developed organizational structures that are associ1989:42:421-5.
ated with better care ("learning makes perfect" hy- 6. riik-adlusuneni approach. J Clin tptdtmMD. Reiults
Bennett CL. Matbewi WC. Kotecoff J. Kanouu
pothesis) (19). High-volume hospitals in New York
ef a con sens us pane) on proceu of care for patients with
PntnmoeyitU wlnll pneumonia (Abttml]. In: tocttdini*
had a higher proportion of patients with low severO tht Vth InlemnllomJ Conftrenve on AtDS. IV8V;2S4A.
f
ity of Illness,fewernonwhite patients, and more
9. Aadrulii DP, WctiowsU VB. Oafc LS. The I9S7 US Hospatients with private health insurance, suggesting
pital AIDS Survey. JAMA l9t9:262i784-94.
that there is some degree of selective referral. Re- 10. Rothmae DJ, Tynan EA. Adventafci and dliadvafltajii of
special hoipitali for pailents with HlV-mficllofl, N En \J
gardless of the reasons behind low-volume hospitals
Med 1990;323:764-4.
having poorer outcomes for persons with AIDS,
11. Mayer'i Taik Forec on AIDS. Auurinf tare for New Y r
ok
Cto'i AIDS population. New York: Mayor'i Talk Force on
more detailed clinical daU, such as can be obtained
AIDS, March
from chart review, arc needed to determine whether12. Weinberg DS, 1969. HW. Coping with AIDS: the ipeclal
Mturay
directing patients with severe opportunistic infecpreblemi of New York City. N Engl J Med i9S7 J17:146971.
tions trom lower-volume AIDS hospitals to higher13. Impemo Pi. The health and health care ehaJIcnsci for New
volume AIDS hospitals is likely to result In ImI990I.
Stair J Mid 1990:90:31-2.
proved chances of short-term survival for these in- 14. York Oty ie thePascalNYCvttenlc MA, at al. The medical eotl>
BeaitettCL.
A,
dividuals. In addition, any interventions to direct
of AIDS In intravenoui drus users In Brooklyn. J Atq ImmtinttUfSyndimpniiU
patienu to higher-volume hospitals should not be
relaiion between
undertaken without careful evaluation of the capac- 15. Bennett CL. Qenler P. Quu PA. et al. The patienti with
raioure* uie and Inhospital mortality for
ity of the hospitals to handle laiger numbers of paAIDS-related PCP. Are* Intirn Mid 1990;130:1447-32.
l
t
Jmntl tfActulrte Immune Defriwmcy Sfndrtmn. Vol. 3. No. f, IW
�05/06/93 09:35
05/06/93 10:17
VACOCRS
MDBAUHDC LIBR
©202 233 3594
©410 606 7905
LIBTECHWLD142D
• • • VACOCRS
»>>
©010/010
121010/010
C. L. BENNETT ET AL.
864
fertnett kttwii* puhlk and vobtntary hoipitali fa Ntw
16. hrbtr BF, Kiiwr DL, WMMI RP. lUUtloit b W n •u^
M M
York CUy, R-3992-NYCHHC. SaaU Monica, Califbnla:
Ilea] voliuiw and htddaue ef pMtopmdva w u d Uteon
Rand (ie pfMi).
lion. NEntlJMH IMI JOS JOO-i.
23. Salllvea LW, Kayi LB. MtdledM hotpttal mortality Mbr17. Flood AB.ScoaWB.EwiyW.DMiprtctle«Bnk«pcrr«a7
matloa, im, Hit, IM. HCFA publication auAbar O T t
O O.
I; the reliliae b w e hoipiul VOUMM end ouieomei Ibr
M en
Health Can Flnaeet Admlalitratlon, Unlud Statei DepartMlcctcd dlMMiik eeutoritl. Mtd Cefr M4S8ilfr-lJ4.
menl of Heatth tad H m n Servlcei, 1989.
u a
II. Flood AB, Scott WR, Bwry W. Doci practice nuke peiffct?
9
1 : the relitloe between hotpttal volume end outcome* end 26. American Hoipiul Aiioelatlon. 1 M AHA Annual Survty of
1
ff^frcb. Olcafo, minols: Hospital lafonoaiioit Servicsi,
other hoepital eharacteriitloi. Med Ca* 1 1 1 21 IMS.
942:
19. Luft KS. Hunt SS. Maerki SC. The volume-Outcome rela27. Rally JV, Ball JK, TWner BJ. Duration and eotu of AIDS
tlomhlp: practice makai paitet er ulectlve paftml palhoepltilkatkmi in New Yoric Variations by patient severity
temif Hnilk Stn Pit l9S7d2il57-S2.
eruinen and hoepital type. Mtd Can 1989^7)1065^6.
30. Maerki SC. Luft HS. Hunt SS. Solacdng catcforiet of ea.
uenu for iiglonaUaillon: Implicalioni of the rcUtioaihiP be- 26. Justice AC, Falmteln AR. Walli CX. A new proinoitle ita*ini system fbr the acquired Immunodefldeacy tyndrame.
tween volume and outcome, Uii Can 1986^4:146-56.
N wtollJ Mtd 1969^20:1311-92.
e
21. HaimanEL.O'Dennelin'.KilbuDH.etal.IevettlsatioBor
tha reUtlonihlp between volume and mortality Ibr turjiul 29. Kalei CP, Munvn JR. Torru RA, at al. Early predleton of
bhhotpttai m n Uy tat Pnenmocyitli carinii pneumonia In
oa t
proeaduru performed in New Yeck State hoipitali. JAMA
the acquired Immunodeficiency lyndrome. Arch Inttrn Mtd
I9«9;262:503-I0.
l967iU7!l4l3-7.
, ,
22. KauAnan OI, Ortbau JC. Schmidt EM, Han Y. HIV Infected
30. Oany SM, Greene J. Prognoitlc indieatori in the Initial prehoipltal patienu in N w York Sute: tht development of
a
lentatlon of Pneumocyitit carinii pneumonia. Chttt 1989;
longitudinal Intormation from a hoipltal diieharfa dau tyn
93:769-72.
Um. NY Siatt J Mti 1990^00)1-42,
m u o e dD
2). Turner BJ. Kelly JV. Ball IK. A icverlty lyitcm for AIDS 31. Cotton DJ. Improving survival in acquired I m n d n e ey syndrome: is experience everything JAMA 19(9^61:
hotpHalitailOM. Mtd Care 1969^7:423-17.
34. Shapiro MF, Park RE, Keeacy j. Brook RH. Mortality3016-7.
Of-
ef*
Otfidemytt
t.Vei.S.Nt.t,
�
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
Health Care Reform
Identifier
An unambiguous reference to the resource within a given context
2006-0810-F
Description
An account of the resource
<p>This collection consists of records related to Hillary Rodham Clinton's Health Care Reform Files, 1993-1996. First Lady Hillary Rodham Clinton served as the Chair of the President's Task Force on National Health Care Reform. The files contain reports, memoranda, correspondence, schedules, and news clippings. These materials discuss topics such as the proposed health care plan, the need for health care reform, benefits packages, Medicare, Medicaid, events in support of the Administration's plan, and other health care reform proposals. Furthermore, this material includes draft reports from the White House Health Care Interdepartmental Working Group, formed to advise the Health Care Task Force on the reform plan.</p>
<p>This collection is divided into two seperate segments. Click here for records from:<br /><a href="http://clinton.presidentiallibraries.us/items/browse?advanced%5B0%5D%5Belement_id%5D=43&advanced%5B0%5D%5Btype%5D=is+exactly&advanced%5B0%5D%5Bterms%5D=2006-0810-F+Segment+1"><strong>Segment One</strong></a> <br /><a href="http://clinton.presidentiallibraries.us/items/browse?advanced%5B0%5D%5Belement_id%5D=43&advanced%5B0%5D%5Btype%5D=is+exactly&advanced%5B0%5D%5Bterms%5D=2006-0810-F+Segment+2"><strong>Segment Two</strong></a></p>
Provenance
A statement of any changes in ownership and custody of the resource since its creation that are significant for its authenticity, integrity, and interpretation. The statement may include a description of any changes successive custodians made to the resource.
Clinton Presidential Records
Publisher
An entity responsible for making the resource available
William J. Clinton Presidential Library & Museum
Text
A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text.
Original Format
The type of object, such as painting, sculpture, paper, photo, and additional data
Paper
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
Briefing Book on AIDS [2]
Creator
An entity primarily responsible for making the resource
Health Care Task Force
General Files
Identifier
An unambiguous reference to the resource within a given context
2006-0810-F Segment 1
Is Part Of
A related resource in which the described resource is physically or logically included.
Box 53
<a href="http://clinton.presidentiallibraries.us/items/show/36144" target="_blank">Collection Finding Aid</a>
<a href="https://catalog.archives.gov/id/12090749" target="_blank">National Archives Catalog Description</a>
Provenance
A statement of any changes in ownership and custody of the resource since its creation that are significant for its authenticity, integrity, and interpretation. The statement may include a description of any changes successive custodians made to the resource.
Clinton Presidential Records: White House Staff and Office Files
Publisher
An entity responsible for making the resource available
William J. Clinton Presidential Library & Museum
Format
The file format, physical medium, or dimensions of the resource
Adobe Acrobat Document
Medium
The material or physical carrier of the resource.
Preservation-Reproduction-Reference
Date Created
Date of creation of the resource.
5/5/2015
Source
A related resource from which the described resource is derived
42-t-2194630-20060810F-Seg1-053-006-2015
12090749