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

940 Abstracts 43479-43482

12th World AIDS Conference

were analyzed using a series of hierarchical planned setwise Cox regression
analyses and Kaplan-Meier survival analyses. The measure of program retention
was the total number of days between enrollment into services and the last day
the client received services. Retention in services was significantly associated
with client characteristics including HIV status, substance abuse, and risky sexual behaviors. Participation in medical, social support, and case management
services appears to incrementally predict retention in the programs above and
beyond client characteristics. In particular, youth who received case management
had significantly higher retention rates than did youth who did not receive case
management.
Lessons Learned: The 10 adolescent care projects have identified a number
of factors that increase the likelihood of engaging HIV-positive and at-risk young
people in services. The provision of sensitive and appropriate HIV services for
youth appears to be associated with retaining the young people in care.
43479 Tackling the impact of the HIV epidemic in Malawi
Owen Kaluwa1, Elizabeth Reid2, A.S. Sy3, Mary Chinyama1,
Angela Trenton-Mbonde4, Lester Chitsulu5, Bill O'Loughlin6. 1 National AIDS
Secretariat, Ministry of Population and Health; 4UNAIDS CPA, Lilongwe, Malawi;
2Senior Adviser BDP, UNDP New York NY, USA; 3UNAIDS Intercountry Team,
Pretoria, South Africa; 5WHO, Geneva, Switzerland; 6ACFOA, Melbourne,
Australia
Issues: How can national capacity be strengthened to analyze the impact of the
epidemic on Malawian society, and to develop effective policies and programmes
to address it.
Approach: A partnership was formed with the National AIDS Secretariat (NAS),
UNDP and UNAIDS, in which each partner made a commitment of time and resources to a long term capacity building process. A working group was created of
senior and experienced people from a wide range of key government institutions,
the private sector, the non-government sector and the media interested in and
prepared to commit themselves to the work. A capacity building process was
designed which was inter alia to identify three areas in which the epidemic would
create severe dysfunctionalities in the next 5 to 7 years.
Results: The three areas identified were: impact on systems and structures,
impact on interpersonal and social relations, andimpact on community support
and household coping systems. The country's capacity to function will be undermined as productive and managerial capacity and growth infrastructure, along
with social and biological reproductive capacity, are adversely affected. An 18
month strategic planning process was designed: to strengthen capacity to discuss the difficult issues raised by the epidemic; to reach a national consensus
on the policies and programmes for the next national strategy; to develop and
have approved the National HIV/AIDS Strategy 1999-2004. The Government has
decided to undertake this extensive and inclusive process of consultations with
communities, technicians and decision-makers leading to the formulation of the
next national HIV/AIDS Strategy for Malawi.
Lessons Learned: HIV sensitive planning processes cannot assume that communities or technicians can identify and discuss HIV related needs and priorities.
The issues are often complex, intimate and threatening. A sense of agency and
of hope must be created, ways found to enter into sustained dialogue, since
proposed actions may benefit from drawing on lessons learned elsewhere, and a
national consensus built on a relevant and sustainable response.
43480 The modification of the AIDS incubation period in
1996-1997: An indirect evidence through a
back-calculation model
Patrizio Pezzotti, M. Balducci, S. Boros, P.A. Napoli, G. Rezza. Istituto Supeiore
Di Sanita', Centro Operativo AIDS, V. Le Regina Elena 299, Roma, Italy
Objectives: To produce an indirect estimate of the effect of the new antiretroviral
treatments based on the expected AIDS cases through a back-calculation (BC)
model.
Methods: Adult PWAs, reported to the AIDS Registry up to 12/31/97, were
included. We adjusted for reporting delay (Heisterkamp method) and for underreporting. BC was used to estimate the HIV incidence curve. Markov Chain Monte
Carlo algorithms were used to perform inference. The epidemic was modelled
using a homogeneous Markov model. Incubation period was divided into three
CD4 count states (i.e., >500, 200-500, <200 cells/mm3), followed by AIDS and
death. From each stage, individuals can progress to the following stage or directly
to death (including a different pre-AIDS mortality rate for HIV exposure category).
We assumed that, beginning in June 1995, therapy reduced only the progression
rate from CD4 < 200 to AIDS, by a factor Q. We also assumed that only a
proportion of individuals in CD4 < 200 has access to effective therapy and that
this proportion increases over time. We estimated both ~ and percentages of
treated individuals, having fixed alternatively the two parameters.
Results: All models incorporating the effect of therapy showed a better fit to the
AIDS data of the basic model without treatment effect. We found a link between o0
and the percentage of treated individuals, with a high fixed value for one yielding a

low estimated value for the other. Fixing the percentage of individuals in treatment
from 5% in 1995 to 90% in 1997, the mean occupancy in CD4 < 200 increased
by a mean of 21 months. Fixing the percentage of individuals in treatment, we
estimated that this effect varies by risk group.
Conclusion: Given that there is no evidence that reporting delay and underreporting changed in the study period, the introduction of the effect of treatment in
the BC model has contributed to better explaining the recent AIDS trend in Italy.

43481 | Disseminated Mycobacterium avium complex (MAC)
infection prophylaxis in HIV-infected patients with CD4
cells < 75/mm3: An updated decision analytic view
Heiner C. Bucher1, Peter P. Sendi2, D. Pfluger3, G. Meier2, M. Battegay2,
A. Gafni4. 1Medizinische Universitaets-Poliklinik Petersgraben 4, 4031 Basel;
2University Hospital of Basel, 4031 Basel; 3BMP Biometrics Pfluger, Solothurn,
Switzerland; 4 McMaster University, Hamilton, Canada
Objective: To determine the effectiveness of prophylaxis for disseminated MAC
infection in HIV-infected patients with a CD4 count of less than 75 cells/mm3 after
the introduction of antiretroviral triple combination therapy in 1996.
Methods: Data was gathered from randomized controlled trials (effectiveness
and toxicity) and the Swiss HIV Cohort Study (SHCS) database. A Markov model
was developed to evaluate four alternatives: rifabutin (300 mg/d), azithromycin
(1200 mg/wk), clarithromycin (2 x 500 mg/d) and no prophylaxis. Pooled sixmonth observations from the two periods 1993-1995 and 1996-1997 were extracted from the SHCS database to estimate transition probabilities for developing
AIDS (non-MAC) and dying from AIDS (non-MAC) given a specified CD4 cell
count, and for switching from one of three CD4 cell count strata (0-49 cells/mm3,
50-74 cells/mm3, >75 cells/mm3) to another. Rates for MAC-incidence and dying
from MAC were derived from the SHCS database.

-60.0 -1 58.0 -56.0-. 54.0 -S 52.0 -50.0 -48.0 -46.0 -44.0 -S42.0
40.0 -38.0 -S36.0 -0 34.0 -32.0 -
*~a
> E
E
|1
<U

0.0 15.0 30.0 45.0 60.0
Triple combination therapy effect (in months)

00

Results: Based on SHCS data from 1995-1997 we simulated a hypothetical initial cohort of HIV-infected AIDS-free patients with a CD4 count of 50-74
cells/mm3 over five years. Incremental effectiveness of rifabutin, azithromycin and
clarithromycin compared to no prophylaxis were 1.90, 2.17 and 2.24 months of life
saved respectively. Incremental effectiveness of antibiotic prophylaxis increased
with decreasing duration of the effect of antiretroviral triple combination therapy
in the study population (Figure) when we used data from the period 1993-1995
to reflect loss of its effect.
Conclusion: Clarithromycin (2 x 500 mg/day) is the most effective strategy to
prevent disseminated MAC infection in our analysis. However, after the introduction of antiretroviral triple combination therapy, the impact of antibiotic prophylaxis
for disseminated MAC infection on survival has declined.
S43482 Use of HIV-antibody counseling and testing data to
estimate time trends in HIV incidence in San Francisco
Timothy Kellogg1, W. McFarland2, M. Katz2. 1AIDS Office 25 Vanness Avenue
Suite 500, San Francisco, California 94102-9082; 2San Francisco Health
Department, San Francisco, CA, USA
Background: Information on trends in HIV incidence are needed at the local level
to better target populations in need of prevention interventions and to assess
the impact of community-wide prevention efforts. We provide a method of using
counseling and testing data to assess trends in HIV incidence among repeat
testers at anonymous and confidential sites in San Francisco (SF), California
(USA).
Methods: Data consisted of information collected from all persons testing for
HIV antibodies at publicly-funded sites in San Francisco between 1/95 and 6/97.
A repeat tester was defined as a person reporting at least one prior HIV-negative
test. A seroconverter was defined as a person whose current test was positive
and who reported their previous test as negative. HIV incidence was estimated
as the number of seroconverters divided by the number of repeat testers. Seroconverters and repeat testers were assigned to particular years based on the
midpoint date between tests. The x2 test for trend was used to assess the change
in HIV incidence by year.
Results: Among 12,467 anonymous repeat testers, 221 seroconversions were
identified (cumulative incidence 1.77%/yr). Among 9,342 confidential repeat
testers, 191 seroconversions were identified (cumulative incidence 2.05%/yr).
HIV incidence decreased significantly among anonymous testers, from 2.2%/yr
in 93-94 to 0.95%/yr in 96-97 (p < 0.001) but not among confidential testers (p =
0.68). By risk populations, the largest decrease in HIV incidence was among men
who have sex with men (MSM) testing anonymously, from 4.25%/yr in 93-94 to
1.40% /yr in 96-97 (p < 0.001).
Conclusions: Our data indicate a significant temporal decrease in HIV incidence among anonymous repeat testers in SF, due primarily to reductions in
seroconversions among MSM. Although not without limitations, our method provides an inexpensive means for monitoring local trends in the HIV epidemic using
data readily available in many health jurisdictions.