AVERT: Validation of AVERT Estimates

The accuracy of the estimates of annual HIV incidence generated by Family Health International's AVERT model rely on the validity of the underlying probabilities of transmission per sexual exposure used in the model. Therefore, it was important to compare the model's estimates with the number of HIV seroconversions observed in a real-life situation.

The validation exercise was performed using data from a recently completed randomized controlled trial evaluating the effect of a commonly used spermicide on HIV transmission among female sex workers in two cities in Cameroon from March 1995 to December 1996.¹ This study provided detailed data on the number of sex acts per year with clients and non-clients, corresponding levels of sex protected by condoms, and estimates of the prevalence of sexually transmitted diseases during the study period (Table 1). The HIV seroprevalence among the partners of female sex workers was estimated from 1994 data on male blood donors from the towns of Douala, Yaoundé and Ebolowa (HIV-positive = 11.5 percent, n = 7,148) and from 1996 data on male members of the military collected from 11 army bases (HIV-positive = 14.6 percent, n = 1,052).

To ensure that the validation was objective, the investigators from the Cameroon trial were asked to provide specific data without revealing the results of their study. Those performing the validation exercise did not know the results of the Cameroon study until they had completed their own estimates with the AVERT model.

The analysis shows that the AVERT estimate of the total number of HIV infections averted was very close to the results of the Cameroon study: 73 estimated new infections compared to 78 actual new infections. The estimated annual incidence rate and observed seroconversion rate were virtually identical (Table 2).

It was not possible to determine how many of the infections observed in the study population resulted from sexual intercourse with clients rather than non-clients. However, estimates derived from AVERT suggest that almost 60 percent of the total infections in this cohort of female sex workers were the result of sexual activity with non-clients. This result most likely reflects different levels of condom use: the reported number of unprotected sex acts was 2.9 times higher with non-clients than it was with clients.

Table 1. Estimated and observed seroconversion in a female sex worker cohort

Parameters used for AVERT:

1. Target population: 1,170 female sex workers, HIV-seronegative at enrollment.

2. Partner population: clients and non-clients. Estimated HIV level: 13%.

3. Average number of sex acts per year in target population and reported condom use:

4. Estimated prevalence of genital ulcer disease (GUD) and non-GUD during the study period:

genital ulcer lesions: 7.5%

inflammatory lesions: 15%

The results of this validation exercise were also published in the journal AIDS.²

References

1. R Roddy, L Zekeng, K Ryan, U Tamoufé, S Weir and E Wong. A controlled trial of nonoxynol 9 film to reduce male-to-female transmission of sexually transmitted diseases. N Engl J Med 1998;339(8):504-510.
2. T Rehle, T Saidel, S Hassig, P Bouey, E Gaillard and D Sokal. Avert: a user-friendly model to estimate the impact of HIV/sexually transmitted disease prevention interventions in HIV transmission. AIDS 1998; 12(suppl. 2):S27-S35.