DEVELOPING A METHOD OF HIV/AIDS ESTIMATION AND PROJECTION USING TRANSMISSION MODELS IN HOCHIMINH CITY

 

Nguyen Thi Thanh Thuy, Tran Phuc Hau,
 Nguyen Vu Thuong, Do Hong Ngoc*,
Nguyen Van Thuc**, Truong Xuan Lieu***, Ha ba Khiem

 

Pasteur Institute in HoChiMinh City and
*Information-Education-Communication Center HCMC,
 ** Dermato-Venereal Center HCMC, *** AIDS Committee HCMC

Background

Efforts on HIV/AIDS estimation and projection were developed in Vietnam. This pilot study applied the method called NESDB (National Social and Economic Development Board) which was established in Thailand to estimate new HIV infections with regard to behavioral components, and to explore the impact of behavioral changes on the epidemic evolution. It would be used to help target prevention measures.

Method

A simple model simulated the processes of HIV transmission (sexual intercourse, sharing of needles, and mother to child passage), to calculate the number of new HIV infections in 1997 in HoChiMinh City (HCMC). The incidence was entered into EPIMODEL used as a second component to calculate the current HIV infections. Three scenarios of projections have been run representing the impacts of three levels of action on behavior: baseline, medium intervention, and high intervention.

New infections were separately presented among populations not infected yet. New male infections were calculated among 15–49 year old males who had sex with infected female sex workers (FSW) without condom use and among male intravenous drug users (IDUs) who shared needles with other infected IDUs. New female infections were among FSW who had sex with infected clients without condom use, women who had sex with infected spouses and female IDUs who shared needles with other infected IDUs. Newly infected infants were born by newly infected females. We used transmission probability rates of other studies (mostly in Thailand) for calculations. Existing data and results of additional surveys were used as inputs.

Results

In 1997, 3,953 new HIV infections and 14,470 current infections in adults were estimated. Baseline data of major risk behaviors were used for these estimates: 26% of men visited FSW, 42% men and 58% FSW used condom inconsistently, 78% IDUs shared needles. New infections were more likely to occur in heterosexuals (64%), that is attributed to a large sexual population. Among infected heterosexuals, males accounted for 62.2%. In male heterosexuals, if interventions of medium level were to be applied to reduce the frequency of FSW visits to 20%, the inconsistent condom use to 30%, with assumptions that other factors evolve in the year 2000, we could theoretically obtain a reduction of 45% of projected new infections comparing to this year’s baseline scenario. New infections in women (non-FSW) depend mainly on the HIV seroprevalence of their male partners. New infections among FSW reflect their condom use and HIV seroprevalence among male clients. Sharing needle behaviors and HIV seroprevalence of IDUs mainly contributed to new infections among IDUs. The 782 reported new infections in 1997 in HCMC represented 20% (ratio: 1/5) of 3,957 new infections estimated from this study. Compared to the previous estimate of the national level on current HIV infections in HCMC (12,261; medium scenario), data from this study are higher.

Conclusions

As with any HIV/AIDS model, this method has its limitations, because it is based on many estimates and assumptions, mixing patterns of risk behaviors could not be estimated, and some data are not available or less reliable. Further behavior studies or studies on transmission probability need to be addressed. In spite of these problems, this study showed the crucial roles played by some high risk behaviors in the epidemic evolution. Furthermore it allowed us to identify those types of behavior which can be dealt with to reduce new infections in the future, such as FSW visits, inconsistent condom use among clients and FSW, and sharing needles among IDUs. Prevention activities should focus on these behaviors.