Abstracts Vol. 1 [International Conference on AIDS (11th: 1996: Vancouver, Canada)]

Mo.C.212 - Mo.C.220

Monday, July 8, 1996

Mo.C.212
HIVTESTING AND POLICY CHANGE,AN ANALYSIS BASED ON STATEWIDE
HIV TESTING DATA, 1994-1995.
Stevenson, Elaine M.*,Thompson SC*, Crofts N*. for the Victorian Collaborative Group on
HIV and AIDS Surveillance. "Epidemiology and Social Research Unit, Ma a'rane Burnet
Centre for Medical Research, Melbourne, Australia
Objective: In July 1994 Victoria introduced a system of user-pays (U-P) HIV testisng which
aimed to reduce the amount of "unnecessary' HIV testing, whilst maintaining a ccess to free
testing for those at increased risk.This policy was revised in April 1995 to extend the availability of free testing to all individuals with specified risk factors.We aim to describe the
impact of the application the U P policy to HIV testing concentrating on changes in the
populations undergoing HIV testing in during 1994 and 1995.
Methods: Data have routinely been collected from HIV testing laboratories in Victora since
199 I. Data forwarded on computer diskette each quarter to the MacfarIane Burnet Centre
(MBC) by most laboratories includes age, sex, personal category postcode, test reason,
result and test history for each specimen tested. Where laboratories were unable to provide computerised data, information on the number of tests they performed each week
was obtained and used to reconstruct the missing records. Demographic data were unavailable for these reconstructed records.
Results: For Jan-Jun 1994 63,389 HIV tests were conducted in Victoria (pop. 4.2 million).
Most of these tests were undertaken in people for whom a history of male homosexual
contact or injecting drug use (IDU) was not recorded. Following the initial policy change in
July 1994 the total number of tests decreased by 18% while testing in homosexual men and
IDUs decreased by approx. 40%.Testing in relation to a potential sexual, occupational or
blood exposure also decreased.Testing in females decreased more than in males A considerable shift in testing from public to priva te laboratories occurred, while the propotion of
tests without satisfactory demographic information increased. Large increases in testing did
not occur in specially designated sentinel free testing sites.
Conclusions: Despite the continued availability of free HIV testing for those atn ncreased
risk, the introduction of a U-P policy led to a reduction in the amount of testing performed
irrespective of risk indications. Further, increases in test in the private ector, which is less
able to collect HIV specific demographic data, may limit our ability to characterise the population who present for HIV testing in this jurisdiction.
Elaine M. Stevenson, Macfarlane Burnet Centre for Medical Research, PO Box 254, Fa rfeld,
Victoria, 3078, Australia.Tel: [phone] Fax: [phone]
Email: stevenso(@dburnet.nmbcmr:unimelb.edu.au
Mo.C.213
LOCAL RESEARCH FOR LOCAL APPROVAL - A CASE STUDY OF A RAPID HIV
SEROSTRIP ASSAY FOR USE IN DEVELOPING COUNTRIES
Frerichs Ralph R*, Zaw M#, Lwin S, 00 KY, EskesN. *UCLA. Los Angeles, CA USA; #
AIDS Prevention & Control Program, Myanmrar; National Health Laboratory Yangon,
Myanmar,- (Saliva Diagnostic Systems, Vancouve rWA USA
Issue: HIV can be avoided if HIV antibody testing finds blood to be contaminated, current
or future sexual partners to be infected, or pregnant women to be carriers.Yet in developing countries widespread testing wil not occur unless donor nations or agenies greatly
increase the supply of tests, or more realistically if internal markets can be developed for
private sector sales of valid and inexpensive test kits.To ensure that public health interests
are served, such test kits should be evaluated in local settings before approval s granted by
the government.The research should be funded by the manufacturers and conducted in a
representative environment by in country research scientists using local laboratory personnel and facilities.
Methods: To evaluate the SeroStrip HIV I/2~ [Saliva Diagnostic Systems (Singapore) Pte,
Ltd], we designed and conducted a study at four township hospitals (2a' 35 beds) located 20
to 50 kilometers from Yangon, the nation's capital. At the National Health Laboratory 400
true HIV positive and 400 true HIV negative specimens were assembled and divided into
four groups, each with 100 (+) and 100 ( ) sera. All sera were coded only with identification
numbers.The blinded specimens were taken to each township hospital, where the laboratory
technicians recorded their results.The kits were used in settirngs where the daily temperature
varied from 14.4 to 32.6"C (58-91~F) and the average humidity was 68%.The SeroStrip
HIV I1/2~ kits did not need to be continuously refrigerated.The results were decoded arnd
independently analyzed by two of us (RRF and MZ) in our respective countries.
Results: The validity of the Sero*Strip HIV I/2~ test was determined using 800 specimens.
Due to an initial coding error one true (+) was actually a true ( -).Thus 399 true ( +) and
401 true (-) were sent to the four field sites. All of the true (+) were correctly identified by
the township hospital laboratory technicians, as were all but two of the true (- ) The sensitivity (% of true positives that test positive) was 399/399 or I100%, and the specificity (% of
true negatives that test negative) was 399/401 or 99.5%.
Lessons: The research was completed in a rapid and professional manner by in-countiy scientists who are best suited to evaluate the validity of HIV tests in local settirgs. Such
research should become part of the national licensing procedure. In addition, the governmooit should charge the manufacturer s a small licensing fee so that the quashag of test kits in
the often harsh phys cal environments of developing countries can be monitored by local
researchersa over tinae.
Nora Eskes, SDS, Inc., Suite A, I 1719 NE 95th Street, Suite G,Vancouvec WA 98682 USA
Tel: [phone]; FAX: I 360-254-7942; Email: Salivalolaol.com
Mo.C.2 14'
PREDICTORS OF HIV TESTING HISTORY AMONG MEN ATTENDING AN STD CLINIC IN HOLLYWOOD, CALIFORNIA
Pearce, C. Leigh', Katsufraks PJ, Dyer IEF', *rRichwald GA. *0Borenstein LA. am"Courtney
JG"'. *LA Gay and Lesbian Community Services Center LA, CA, USA; "University of Southern
California School of Medicine, LA, CA, USA; *LA County STD Prograr, LA, CA, USA
Objective: lo determine the predictors of HIV testing history among men attending the
Los Angeles Gay and Lesbian Community Services Center Sexually Transmitted Disease
(STD) Clinic in Hollywood, California.
Methods: A total of 438 men who presented for care at the clinic enrolled in the study
from Aug. 7 to Oct. 27, 1995.We collected demographic, behavioral, and clinical data on all

participantm ' (corcp!ete data n-426). Logistic regression was used to evaluate predictors of
whether dient' citr,'r ever or never previously tested for HIV. We assessed the effects of
the followingr self r-ep: t: ivariables sexual orientation, STD history, age, race/ethnaity, and
number of sexual pair tnrs in the previous 30 days.
Results: Sex'ual orientat r of study participants was 72%I homosexual, 13% bisexual, and
I 5% heterosexual. I pe,lation was 52% white, 29% Hispanic, 12% Afican-American, 4%
Asian American, % Ntivr' Armerican, and 2% other with a mean age of 32 and a mean
number of sexual partner s of 3 (median=2). Percentages previously tested for HIV were
94%, 68%, and 82% for homosexuals, heterosexuals, and bisexuals, respectively In the logistic
model, only sexua! orientation of homosexuali was an important determinant of HIV testing
history OR=-5.5 (95% CI 2.7-1I.0), with heterosexual men as the reference. Bisexual men
were 1.8 times (95 CI.75- 4.39) more likely to have tested for HIV than heterosexual
raen. History of hepatitis B (OR=3.2), human papilloma virus (OR- 1.4), herpes (OR-2.1),
and syphilis (OR-= 1.6) were included in the model, as were 3 categorical age variables of
<25 years (OR=.81), 26 30 years (OR= 1.7), and >30 years (reference)
Conclusions: The rates for ever testing for -HIV in this study population were high, even
given these individuals were seeking care at an STD clinic. Having tested for HIV was
strongly associated with sexual orientation, with hormosexual rmen 5.5 times more likely to
have ever tested than heterosexual men. This suggests that this homosexual male popula
tion overwhelmingly perceives the need for HIV testing. However almost a third of the heterosexual male clinic population may still consider themselves not to be at risk for HIV
infection, despite presenting for STD treatment. In view of the increasing level of HIV infection in the heterosexual population, specific educational messages must be developed to
encourage testing in this group.
C. Leigh Pearce, I 625 N. Schrader Blvd., Los Angeles. CA 90028, USA
Telephone: [phone], Fax: [phone], emarl: pear,087avondeerm.cdc.gov
Mo.C.215
HIV TESTING BEHAVIOR AND BELIEFS OF PHYSICIANS TREATING HIV+ INJECTED
DRUG USERS (IDUS) IN SOUTHERN FRANCE
Cavailler P. I, Cassuto 2, Lepeu 3, GambyT. 4, Moatti JP MayneT 6, Rotily Michel. ISouthEastem French Centre for Disease Control, Marsele, Fr',ne; 2 Department of internal
medicine. H6pital Cirmiez, Nice, France; 3 Dept. of hematoloy, H6pital H. Duffaut, Avignon,
France; 4Dept. of dermatology H6pital St Joseph, Marse lie;5 INSERM Unit 379. Institut P
Calmettes, Marseille: 6CAPS San Franisco, USA.
Objective: To examine HIV testing behavior and the beliefs of physicians treating IDUs with
HIV infection in Southern France.
Methods: This was a consecutive convmience sample of the first 100 HIV+ patients with
IDU as origin of infection seeking services at outpatient clinics in spefic hospitals in
Marseille, Nice, and Avignon between July and Decembe, I1995. We assessed demographic
variables via interview; and HIV testing and drug use behavio, and physician beliefs, via selfreport questionnaires
Results: Demographics- The sample was 67% male, mean age= 34 (SD= 4.5). Average year
of first IDU was 1980 (SD-4): 37% were currently active users; I % had stopped using less
than I year, and 52% had been abstinent an average of 6 years (SD=4.4).Virtually 100% of
the sample had used heroin Mean CD4 cell count at first interview was 435 (SD= 179).
HIV Testing - 78% of the sample reported testing HIV+ on their first test. 40% reported
never suspecting they were HIV+ before test ng, while 32% reported waiting a year or
more between first suspecting they were pos tive and testing. 32% reported testing on the
advice of their primary physician, 56% on their own initiative, and 12% on the advice of others. 32% were tested by their physician, 37% in a hospital, and I % in an anonymous testing
center. Physician/Care-related variables - Though 37% or the sample reported current drug
injection, only 8% of physicians reported detecting signs of active drug use during physical
examination.When IDU report was compared to physican beliefs about current drug
activity 60% of subjects were correctly identified as non-active, and I 2% as currently active.
However, 22% of subjects were falsely identified as non-active, while 5% were falsely identifed as active (X2= 18.5, p<.01).
Conclusions: Primary care physicians were biased against recognizing active drug use
among their patients. Only one third of the sample reported HIV testing on the advice of
their physician, and many either never suspected, or delayed testing significantly after suspecting they were HIV seropositive.This study suggests tae need to heighten physician
awareness of signs of IDU, as well as the need for physic irians to counsel suspected IDUs
regarding the need to test for HIV
DocteurThiern Cry Gamby, Service Derm atologie, Hpita Saint Joseph, Bd de Louvain I13008
Marseille, France Telephone: [phone] / Fax: [phone]
Mo.C.220
A PROFILE OF FEMALE INJECTING DRUG USERS IN ZURICH BASED ON A
META-ANALYSIS OF VARIOUSLY RECRUITED DRUG-USER SAMPLES
Wang, Ien*, Lanz A**, Seidenberg A"'-, Zellweger U*, Somaini B*. Institut fur Sozlal- und
PD;-ventivmedizin der Unuvensit t Zmmclv, Swirtzerland: * Institut fur Suchtforschung. Zurich,
Switzerland *'" ZoacL2 -ARUD, Zurich, Switzerland.
Objective:To estahlsh a proile of the female njecting drug user (IDU) population in the
oty of Zurich, Switzerland.
Methods: Taking data from drug user studies carrred out Juring 1989 92, we perdormed a
meta analysis of 13 datasets collected at Needle Park (ar.,uably one the most diverse samples ever collected anywhere), low threshold social serwce programs targeting drug users,
and various drug tieatment modahties. Orginal datasets were cleaned, and key variables cov
ering demographics, drug use, treatment, and HIV were srandardized. Statistical tests involved
blocking, testing for heteogenety and calculatng combined effect size and p-values where
appropriate.The 3391 cases who reported a history of IDU were included in this analysis
Results: Vomen comprised 29% of the pooled sample and had an average age of 25.4
years (I.8 years younger than the men, p<.0O!). In g'eneral, women were more likely to be
Swiss nationals (85% vs 77' p3.002) and unemployed (52% vs. 46% p-.00002), but these
gender differences do not hold among HIV+ IDs. A higher percentage of women overall
reported an illegal source of income (58% vs. 48%, p D.00002), and female IDUs consistently
distinguished themselves from thfer mate counterparts r their cohabitation patterns: 37%
vs. 18% live with a partner: 7% vs. 28 hve witia parents, and 29% vs. 36% Irve alone

33