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

12th World AIDS Conference

Abstracts 43474-43478 939

43474 Modelling disease progression in HIV-infected patients
without AIDS in the era of antiretroviral triple
combination therapy
Heiner C. Bucher', Peter P. Sendi1, D. Pfluger2, B.A. Craig3, M. Battegay1.
' Medizinische Universitaets-Poliklinik Petersgraben 4, 4031 Basel; 2BMP
Biometrics Pfluger; 3Dep. Statistics Purdue University, Switzerland
Objectives: To compare the progression to AIDS in HIV-infected subjects with a
CD4-cell count below 100/mm3 before and after the introduction of antiretroviral
triple combination therapy in 1996.
Setting: Seven clinical centers within the Swiss HIV cohort study (SHCS)
Methods: We modeled the progression to AIDS with a Markov model consisting of four AIDS-free and one AIDS state: (1) CD4 count 0-49 cells/mm3,
(2) CD4 count 50-74 cells/mm3, (3) CD4 count 75-99 cells/mm3, (4) CD4 count
-100 cells/mm3, and (5) AIDS. Pooled six-month observations contributing to
the two periods 1993-1995 and 1996-1997 were extracted from the SHCS
database to estimate transition probabilities. For each period AIDS-free survival
was calculated over three years by cohort simulation.
Results: Using data from the time-period 1993-1995 and 1996-1997, the mean
expected AIDS-free survival for a hypothetical cohort starting in disease state (3)
was 23.75 and 33.49 months over three years of simulation. The corresponding
survival times for a hypothetical cohort starting in disease states (2) and (1) were
20.66 and 31.39 months, respectively 18.04 and 28.38 months. In a hypothetical
1000-subject cohort starting in disease state (3), 661 respectively 140 subjects
would be expected to have developed AIDS at the end of each simulated 3-year
period. In a hypothetical 1000-subject cohort starting in disease state (2) and
(1) 730 and 208, respectively 785 and 306 subjects would be expected to have
developed AIDS.
Conclusion: AIDS-free survival among HIV-infected patients with a CD4 count
of less than 100 cells/mm3 improved after the introduction of triple combination
therapy.
3475 Estimating HIV prevalence and incidence in Canada:
Using direct and indirect methods with Monte Carlo
simulation
Chris P. Archibald'2, R.S. Remis3, J.D. Farley2, D. Sutherland2. 'Room 0108B
BC Building Tunneys Pasture 0900B1 Ottawa, Ontario; 2Bureau Of HIV/AIDS,
Health Canada, Ottawa; 3University Of Toronto, Toronto, ON, Canada
Background: To monitor the HIV epidemic in Canada and to help evaluate and
guide prevention efforts, estimates of national HIV prevalence and incidence are
required. We describe a technique that combines direct and indirect estimation
methods with Monte-Carlo simulation to obtain national estimates for Canada at
the end of 1996.
Methods: For the direct method, prevalence and incidence estimates (numbers)
for Montreal, Toronto and Vancouver were derived by multiplying prevalence and
incidence observed in available studies by our best estimates of population size
obtained from census and survey data for each exposure category (homo/bisexual
men (MSM), injection drug users (IDU), MSM-IDU, heterosexual-endemic and
heterosexual-other). Estimates were extrapolated from the three cities (representing 62% of recent AIDS cases and HIV diagnoses in Canada) to their respective
provinces and then to the rest of Canada based on trends in recent AIDS cases
and HIV diagnoses for each exposure category. We also used an indirect method
to calculate prevalence for each exposure category and geographic region by
dividing the cumulative number of persons diagnosed with HIV infection, less
cumulative deaths, by the proportion tested for HIV (obtained from population
surveys). We estimated geometric upper and lower limits for each parameter by
statistical and epidemiologic means and carried out Monte-Carlo simulation to
obtain ranges for the final estimates. We also used secondary estimation methods
to validate estimates obtained as above, including back-calculation from AIDS
cases.
Results: Our preliminary estimate for HIV prevalence in Canada at
end-1996 was 38,000 (95% CI: 33,000-44,000) and for annual incidence 4,100
(3,400-4,900). By exposure category, 51% of new infections were among IDU,
MSM 29%, heterosexuals 16%, and MSM-IDU 4%. The estimate for incidence is
higher than an earlier estimate of 2,500-3,000 annual infections during 1989-94,
the increase being primarily among IDU.
Conclusions: The advantages of this combined estimation method are that it
makes maximum use of a wide variety of surveillance and research data, it can
readily incorporate new data and it serves to highlight gaps in existing knowledge.
Though the resulting estimates are necessarily somewhat imprecise, they are
useful for planning and evaluation purposes and in Canada have demonstrated
the need for continued and improved HIV prevention and care programs.
575*/43476 Patterns of participation in mass STD treatment:
Model-aided evaluation of impact on HIV incidence
Eline Korenromp1, C. Van Villet1, A. Gavyole2, H. Grosskurth3, R.H. Hayes3,

J.D.F. Habbema'. 'Dept. of Public Health, Erasmus University Rotterdam
Postbus 1738, 3000 DR Rotterdam, The Netherlands; 2African Medical and
Research Foundation Mwanza, Tazmania; 3London School of Hygiene and
Tropical Medicine London, UK
Background: Repeated mass treatment (MT) of STDs is a radical approach to
HIV control, which is a subject of current research (Rakai study). Impact is likely
to depend on levels and patterns of participation, and on migration rates in and

out of the targeted population. We used the simulation model STDSIM to predict
the effects of these factors in annual MT in a typical rural African population.
Methods: STDSIM parameters were fitted using data from rural Tanzania.
Cure rates for gonorrhea, chlamydia, syphilis and chancroid following MT were
assumed to be 95%. Participation at each round of MT was assumed to be 80% or
90%; participation at one round was independent of participation at other rounds.
Two scenarios were assessed: 1) participants randomly distributed in the population; 2) concordance in participation between partners in stable relationships.
Effects of migration rates were also assessed.
Results: Model predictions depended on whether subjects are susceptible to
reinfection immediately after treatment of latent syphilis. For a 1 year non-susceptibility, predicted reductions in annual HIV incidence after 4 rounds, in the
absence of migration, were approx. 21% resp. 31% in scenarios 1 and 2 for 80%
participation, and 43% resp. 50% for 90% participation. If treated subjects were
assumed non-susceptible to reinfection for 2 years, the corresponding reductions
were 31% resp. 42%, and 50% resp. 55%. Similar reductions were found if
participation at one round depended on participation at earlier rounds, because
participants who had STDs probably get reinfected before the next round of
MT. Assuming 5% annual population replacement due to migration reduced the
impact of MT by about half in all cases.
Conclusions: STDSIM simulations indicate that if stable sexual partners behave concordantly in their participation, the impact of MT on HIV incidence is
larger than if participation is partner-independent. The extent of migration appears
as important for the population impact as the overall level of participation. Overall
effects will to a large extent depend on the unknown time till re-susceptibility
after MT of latent syphilis. These factors should be taken into account when
interpreting the results of randomized trials of MT.
S188*/43477 1Estimating the effect of treatments on AIDS
incidence in Australia
John M. Kaldor, M.G. Law, J. Cui. 'National Centre in HIV Epidemiology and
Clinical Research 376 Victoria Street, Darlinghurst 2010, Australia
Objective: To estimate the effect of combination antiretroviral treatments on AIDS
incidence in Australia.
Methods: Statistical analyses used the method of back-projection, which is
based on observed AIDS incidence data, and assumes knowledge of the progression rate from HIV infection to AIDS and the modifying effect of antiretroviral
treatments. Analyses were based on quarterly AIDS cases, adjusted for reporting delays, diagnosed by the end of 1996 and reported by 30 June 1997. The
reduction in the observed number of AIDS cases, and the number of AIDS-free
person-years gained, were obtained by estimating the pattern of AIDS incidence
that would have occurred if no antiretroviral treatments had been available.
Results: Without any reduction in the rate of progression to AIDS, it was
estimated that there would have been 845 cases in 1996, whereas 739 cases
were observed. It was estimated that all antiretroviral treatments in Australia have,
to the end of 1996, decreased AIDS diagnoses by 340 cases (plausible range
140 to 750), corresponding to 1000 AIDS-free person-years (plausible range
370 to 2230). All antiretroviral treatments were estimated to have reduced AIDS
incidence in 1996 by 10%.
Conclusions: Improved combination antiretroviral treatments have reduced
AIDS diagnoses in Australia in 1996. Much of the benefit of combination antiretroviral treatments is yet to occur.
43478 Models of service utilization and retention in treatment
among youth with, and at-risk for HIV
G.J. Huba', L.A. Melchor2, A.T. Panter3, E. Wright4, E.R. Woods5, A. Schneir5,
R. Brady6, A. Tenner7, B. Greenberg8, M. Sturdevant9, M. Wallace1',
S. Tierney", B. Singer6, D.E. Brief2, R. Feudo12, T. Bettencourt13,
G. Remafedi14, E. Goodman5. 15811A Uplander Way Culver City, CA 90230;
2The Measurement Group Culver City CA; 3Univ. of North Carolina at Chapel
Hill Chapel Hill, NC; 4Indiana University Indianapolis, IN; Children's Hospital of
Boston Boston, MA; 6Health Resources & Services Admin. Rockville, MD;
7Youthcare Seattle, WA; 8Walden House, Inc. San Francisco, CA; 9Univ. of
Alabama at Birmingham Birmingham, AL; '1Indiana State Department of Health
Indianapolis, IN; 11 Health Initiative for Youth San Francisco, CA; '2Gtr. Bridgeport
Adol. Pregnancy Program Bridgeport, CT; 13Bay Area Young Positives, Inc. San
Francisco, CA; '4Univ. of Minnesota Youth & AIDS Project Minneapolis, MN, USA
Issues: Young people represent one of the most medically under-served populations in the United States; in addition, adolescents and young adults are
increasingly affected by HIV disease. Adult-focused services may not be accessible or appropriate for youth. This paper examines the retention of young people
who are HIV-positive or at high risk for HIV in service delivery programs designed
specifically for this population.
Project: In 1993, the US Health Resources and Services Administration (HRSA)
Special Projects of National Significance (SPNS) Program funded 10 national
demonstration projects targeting HIV/AIDS services to adolescents and youth. A

key goal of all of the 10 projects has been to engage youth in attractive, relevant,
and quality services and to keep them returning to those services in order to
deal with the multiple issues each brings to the door of the programs in which
they participate. This paper examines various factors in the service episodes that
predict how long young people are retained in these programs.
Results: The service episodes of 1,277 young men (17.6 percent known to
be HIV-positive) and 1,311 young women (7.9 percent known to be HIV-positive)