HIV Surveillance and Epidemiology

    HIV Surveillance Office > Archives

Revised Projection of HIV Infection and AIDS Cases in Hong Kong - 28 Feb 1999

Up to about 1995, it was generally believed that extensive spread of HIV among heterosexuals with high-risk behaviors would occur in almost all Asian-Pacific countries. HIV projections were generally made for Asian-Pacific countries using different HIV epidemic models. In Hong Kong, a scenario/modeling approach was used. HIV prevalence in Hong Kong up to 1995 was estimated using a back calculation method after adjusting for incomplete reporting and reporting delays. As of 1995, it was estimated that the cumulative number of HIV infections in Hong Kong was close to 3,000. In retrospect, this estimate of the cumulative incidence of HIV infection in Hong Kong may have been a bit high, but was a reasonable working estimate, based on the available data at that time.

Using an HIV/AIDS epidemic model (EPIMODEL) several different HIV scenarios were constructed for Hong Kong from 1994 to the year 2000. The low HIV scenario projected that the cumulative total of HIV infections in Hong Kong in the year 2000 would range from 8,000 to 12,000. Since 1995, HIV/AIDS surveillance data from Hong Kong have not indicated any significant increase in HIV prevalence. Based on the limited number of HIV infections detected in HIV serosurveys carried out over the past few years, the relatively low numbers of annual AIDS cases reported since 1995, and the total number of HIV-infected persons detected to date (about 1,000), the best estimate of HIV prevalence in Hong Kong, as of early 1999, remains at about two to three thousand. Thus, projection of HIV prevalence using an epidemic model is inappropriate where epidemic spread of HIV is not present.

The development of HIV/AIDS policy and control programs in Asian-pacific countries has been driven primarily by the belief that extensive HIV spread, in countries with current low HIV prevalence rates, is inevitable. This was because risk factors such as commercial sex, "high" STD prevalence rates, and low condom usage rates, are all present in most Asian-Pacific populations, including Hong Kong. These factors are indeed present in virtually all Asian-Pacific countries, but they are present in significantly varying degrees. In the few Asian countries where high (>2% of adults) HIV prevalence rates are present, the percent of males who regularly visit sex workers and the average sexual partner exchange rates of sex workers, were found to be much higher (several fold greater for each parameter), compared to Asian countries with low (

Epidemiologic studies and observations throughout the world during the 1990s have shown that HIV epidemics in men who have multiple male sex partners (MSM), injecting drug users who routinely share injection equipment (IDU), and heterosexuals with multiple sex partners (HET), are basically independent of each other. The presence of HIV in any of these "Risk Behavior Groups" (RBG) can serve as the introduction or spark to ignite or start an epidemic in another RBG, but any subsequent epidemic spread will depend on the pattern and prevalence of risk behaviors in that group. All areas of the world, including Hong Kong, have had at least a decade or more for HIV to spread in one or more of their RBG. It is therefore unlikely that the very low HIV prevalence levels found in some RBG in different parts of the world are due to insufficient time for HIV to have spread.

It is becoming increasingly clear that HIV prevalence rates present in any population is directly related to the pattern and general prevalence of HIV-risk behaviors in that population. The limited behavioral data available suggests that the patterns and prevalence of HIV risk behaviors in Hong Kong are not sufficient to fuel a self-sustaining epidemic of HIV in the heterosexual population, even among heterosexuals with the highest risk behaviors. Thus, unless a significant increase in heterosexual HIV risk behaviors occurs, the prevalence of HIV in Hong Kong can be projected remain relatively low.

Continued quantitative behavioral surveillance is needed for MSM, IDU and HET "groups" at highest risk of HIV infection.

James Chin

Stockton, California
February 28, 1999