Hay J, Osmond D; International Conference on AIDS.
Int Conf AIDS. 1990 Jun 20-23; 6: 250 (abstract no. F.C.676).
Hoover Institution, Stanford University, Stanford, California, USA
OBJECTIVE: Back calculation statistical techniques are used to estimate the time pattern of HIV infection in each of six major regions and seven adult/adolescent HIV transmission risk groups in the state of California through 1988. The results are used to develop AIDS incidence projections through 1992 by region and risk group. METHODS: Using regression with linear inequality restrictions, AIDS incidence is fitted to progression to AIDS among HIV seropositives to estimate the time pattern of HIV infections using several alternative estimates of the HIV incubation distribution. This methodology is applied to 1979-1989 California AIDS incidence data, aggregated by region and risk group. Statistical specification tests (i.e., parameter restrictions and heteroskedasticity adjustments) are performed. Adjustments for therapeutic intervention, case reporting delays and underreporting are made. Bootstrap techniques are used to establish confidence bounds on model estimates and projections. RESULTS: A median estimate of 82,900 cumulative HIV infections occurred in California (98% CI: 71,000 - 155,500) prior to 1989, with 80% or more of these infections occurring in 1983 or earlier. HIV incidence began relatively earlier in Los Angeles and San Francisco metropolitan areas with a major peaking in 1981-1982. The other metro/rural regions of California had their major waves of infection 12 to 24 months later. California annual AIDS incidence is projected to increase gradually from 5,750 in 1988 to 7,425 in 1992 (98% CI: 6,227 - 10,218).
Publication Types:
Keywords:
- Acquired Immunodeficiency Syndrome
- Adolescent
- Adult
- California
- Forecasting
- HIV Infections
- HIV Seropositivity
- Humans
- Incidence
- Los Angeles
- Probability
- San Francisco
- methods
- statistics & numerical data
Other ID:
UI: 102184612
From Meeting Abstracts