Bridging the Gap: Conference Record [Abstract book, International Conference on AIDS (12th: 1998: Geneva, Switzerland)]

12th World AIDS Conference

Abstracts 13226-13230 131

13226 The impact of AIDS on mortality among drug injectors in
Chicago: 1988-1995
Dawn Sievert, Wayne Wiebel, L.J. Ouellet, V. Ramakrishnan, R.C. Hershow,
A.D. Jimenez, W.A. Johnson. University of Illinois at Chicago SPH 2121 W
Taylor St. 552 Chicago IL 60612, USA
Background: This study investigates the relative impact of HIV infection on
mortality among injecting drug users (IDU) in Chicago.
Methods: IDU (n = 1090) who were not in drug treatment were street-recruited
in 1988 and 1992 from three low-income neighborhoods targeted for an HIV
prevention intervention. Survey interviews and serologic samples were collected
at baseline and nine follow-up waves between 1988 and 1995. Cause of death
was obtained through a National Death Index search, medical examiner and
police reports, and interviews with health care providers, caregivers, relatives,
and friends.
Results: HIV seroprevalence at baseline was 26%. We observed 111 HIV
seroconversions among the 805 IDUs who were seronegative at baseline (mean
follow-up = 5.78 years). By the end of 1995, 23% of the cohort were deceased. The
leading causes of death were AIDS-related illnesses (39%), drug overdose (13%)
and homicide (8%). Drug overdose was rare among HIV-positive participants.
The death rate per 100 persons increased from 1.06 in 1988 to 5.11 in 1995,
compared to an average rate for Chicago of 0.99. From 1992-1995, non-AIDS
death rates remained steady while AIDS death rates increased two-fold. The
probability of survival after eight years was 0.83 for HIV negative participants,
0.61 for HIV seroconverters, and 0.46 for those who were HIV positive at baseline.
The log-rank statistic indicates that these survival curves differ significantly from
one another (p -- 0.002).
AIDS-related illness contributed heavily to a high death rate among IDUs in
Chicago. During each year of the study, the relative impact of AIDS on mortality
increased. There was no evidence of increased rates of suicide among IDUs who
were HIV-positive.
13227 Trends in hospital utilization by patients with
symptomatic HIV infection in the United States
Richard M. Selik. CDC, MS-E 47, 1600 Clifton Road Atlanta, GA 30333, USA
Objective: To examine trends in hospital utilization by HIV-infected patients during
1985-1995.
Methods: Data were from the annual National Hospital Discharge Survey,
based on a representative sample of discharges from nonfederal short-stay
hospitals. HIV-associated discharges were defined as those with an ICD-9-CM
diagnosis code of 042, 043, or 044 in 1987-1995 or 279.1 in 1985-1986.
Results: The rate of HIV-associated hospital discharges per 100,000 civilian
population increased from 10 in 1985 to 83 in 1995. The proportional increase
was less among males than females, particularly between 1992 and 1995, when
the rate among males increased 12% (from 113 to 126), while the rate among
females increased 52% (from 27 to 41). Comparing 1985-1987 with 1994-1995,
the percentage of HIV-associated hospitalizations for females increased from 9%
to 25%, the percentage for persons of black race rose from 20% to 43%, that
for persons of white race fell from 68% to 38%, and the percentage for which
government funding was the expected primary source of payment increased from
36% to 62%; the percentage with a diagnosis of pneumocystosis fell from 31%
to 13%, the percentage with Kaposi's sarcoma decreased from 13% to 6%, the
percentage with bacterial or unspecified pneumonia rose from 10% to 18%, the
percentage with severe weight loss increased from 4% to 10%, the percentage
with a disorder of fluid, electrolyte, or acid-base balance rose from 10% to 28%,
and the percentage reported with illicit drug use increased from 7% to 14%.
Conclusions: Among HIV-associated hospitalizations during 1985-1995, the
proportions increased for persons who were black, female, indigent enough to
expect government funding, or used illicit drugs; the proportions decreased for the
AIDS-defining diagnoses of Kaposi's sarcoma and pneumocystosis and increased
for diagnoses that are less specific for underlying HIV infection.
13228 HIV in Canada: National HIV Test Surveillance data,
1985-1996
Nicola P. Wright1, N. Cameron2, J.D. Farley2, D. Sutherland2. 'Bureau of
HIV/AIDS & STD Health Canada Tunneys Pasture Ottawa Kia OK9 PL0900BI;
2Health Canada, Ottawa, Canada
Objectives: To describe national Canadian HIV Test Surveillance data for the
period 1985-1996.
Design: HIV positive test data, without personal identifiers, were received from
all Canadian provinces and territories. Duplicate reports were removed, to the
degree possible, at the provincial/territorial level. National HIV positive test data
were analysed for the period 1985-1994, 1995, and 1996 by gender and exposure
category.
Results: A total of 33,398 positive HIV tests were reported in Canada for the
10 year period 1985-1994; 3,091 HIV positive tests were reported for 1995; and

2,857 for 1996. Data from those test reports that had information on gender
indicated that females increasingly accounted for a greater percentage of the HIV
positive test reports (9.8% for 1985-94; 19.4% for 1995; and 21.5% for 1996).
The distribution of test reports by exposure category also changed over time.
Data for those HIV positive reports that had information on exposure category
are presented in the table.

Exposure Category
1985-94
1995
1996

MSM IDU Hetero. MSM/IDU

74.9% 8.4% 6.6%
46.8% 23.9% 15.5%
39.9% 31.4% 17.6%

2.0%
4.2%
2.8%

MSM Men who have Sex with Men; Hetero. Heterosexual Contact/Endemic; IDU Injection
Drug Use; MSM/IDU; Men who have Sex with Men/Injection Drug Use
Conclusion: The analyses of HIV data was based only on the tested population. The analyses indicated that over time females were accounting for a greater
percentage of positive HIV test reports. In addition, the percentage of reports
accounted for by the exposure categories IDU and heterosexual contact/endemic
increased in contrast to MSM which showed a decrease. These data indicate that
the demographics of those testing positive for HIV in Canada is changing which
may reflect the dynamic nature of the HIV epidemics in Canada. Implications and
limitations of the data will be discussed as well as the importance of ongoing
HIV/AIDS surveillance for effective targeting of interventions.
13229 Characteristics of HIV-infected patients receiving regular
care in the US: Results from the HIV cost and services
utilization study (HCSUS)
Samuel Bozzettel, S.H. Berry', N. Duan', M.R. Frankel2, J. Keesey1,
D. Lefkowitz3, M.F. Shapiro'. 1RAND, 1700 Main Street M-18, Santa Monica CA
904407; 2Baruch College, New York NY; 3AHCPR Rockville MD, USA
Background: Unbiased information on the HIV-infected population receiving regular care in the contiguous United States has been unavailable because reporting
is limited and previous studies have relied on convenience samples.
Methods: The HCSUS is a nationally representative cohort of HIV-infected
persons assembled using a multistage design that randomly sampled areas then
providers then subjects seen in Jan or Feb of 1996. We report on the final sample
of 2865 respondents and preliminary weights to adjust for sampling, multiplicity,
and nonresponse. Estimates of the population under regular care (> = 1 visit
every 6 mo) were developed using various assumptions regarding the clustering
of visits.
Results: In the US, >250,000 HIV-infected persons received regular medical
care outside of prisons and the military in Jan and Feb 1996, implying that
<400,000 are in regular care. 1/5 were women. 1/3 were African-Americans, 1/3
were seen by smaller providers of HIV care, and 1/3 had a previous AIDS-defining
condition. 1/2 had <200 cells, 1/2 were men having sex with men without IV drug
use, 1/2 were unemployed, and 1/2 had households incomes of <$10,000/yr.
1/10 had >500 CD4+ and 1/10th were >50 yrs old. Private insurance was twice as
common in the Midwest and West relative to the Northeast and South and among
women. Women were twice as likely to be African-American. Gay men were twice
as likely to be employed than those in other risk groups. Other characteristics
also varied strikingly by region, gender, risk, insurance and other factors. Those
represented by the HCSUS used substantial amounts of medical care in the 6
months prior to their interviews including >550,000 hospital nights, >1.8 million
outpatient visits, and >30 million days of antiretroviral therapy.
Conclusions: Nearly 1/2 the estimated 750,000 HIV-infected Americans are
not in regular care, and many receive care from smaller providers. These patients'
aggregate consumption of medical care is substantial even though many are in
disadvantaged populations and dependent on public assistance. New initiatives
to improve financing, access, and quality of HIV care in the US need on need not
rely on biased information from convenience samples as representative samples
are achievable.
113230 Trends in AIDS incidence among black persons - United
States, 1990-1996
Hazel Dean-Gaitor, P.L.F. Fleming, J.W.W. Ward. Centers for Disease Control
1600 Clifton RD MSE 47 Atlanta GA, USA
Objective: To describe national trends in estimated AIDS incidence rates among
US blacks from 1990 through 1996.
Methods: We examined trends in the estimated number of black persons for whom AIDS-opportunistic illnesses (AIDS-OI's) were diagnosed during
1990-1996, based on AIDS cases reported to CDC through September 30, 1997.
Data were adjusted for reporting delays, unreported risks, and AIDS-OI diagnosis dates were estimated for cases reported based only on the immunologic
criteria. The 1990 and 1996 estimated census were used to estimate AIDS-OI
incidence rates per 100,000 black population for 1990-1996, by gender, age, and
transmission category.
Results: From 1990 to 1996, estimated AIDS-OI incidence rates declined
among whites (from 12 to 11 per 100,000), increased among blacks (from 50 to
76 cases per 100,000) and Hispanics (from 28 to 32 cases per 100,000). During
1995 to 1996, the AIDS-OC rate among blacks declined from 77 to 76 cases per
100,000 black population; AIDS-OI rates decreased among black men (from 119
to 116) but increased among black women (from 44 to 46). Increases differed by
age groups: AIDS-OI rates increased for the 40-49 age group (from 95 to 98)
and the 50 and over age group (from 36 to 37) but remained stable for the 13-19
age group (2) and for the 30-39 age group (141). Rates declined slightly for other
age groups. AIDS-OI rates increased for persons infected through heterosexual
contact (from 22 to 24) and declined for pediatric exposure groups (from 6 to 5),

men who have sex with men (from 65 to 62), and injecting-drug users (from 40 to
39).