In this paper we set out to examine by means of microsimulation the effect of patterns of divorce — specifically high levels of divorce — and remarriage, and of spousal fidelity, on the lifetime risk of women’s acquiring HIV in rural southern Malawi. Although our investigation is undeniably incomplete, some interesting findings have already emerged.
When extramarital sexual activity is both common (although by no means universal) and sometimes precipitates divorce, the resulting divorce rates may be substantial, just as they are in the actual population. Nevertheless, even if strategic divorce separates partners one of whom believes that the other’s illicit sexual behaviour may bring HIV infection to the marriage, the reduction in women’s lifetime risk of HIV acquisition may be rather modest. The problem is that women may be already infected when they divorce and, once divorced, may have affairs with men of various disease statuses, and once remarried they will be exposed to any infections possessed by their new, possibly philandering, husband, and in some additional cases to the infections of their own extramural sexual partners.
Given the levels of extramarital sexual activity and marital infidelity assumed in the models, and the same level of divorce we achieved a 20 per cent reduction in the lifetime risk of HIV (from 34 per cent to 27 per cent) when spouses divorced on the basis of AIDS or infidelity. In the absence of remarriage after divorce or widowhood we achieved a reduction of about one-half (34 per cent to 16 per cent). This is an encouraging result but the scenario, in a society that places so much importance on the institution of marriage, is unrealistic.
The picture changes when we impose marital fidelity on our simulated couples. In this case, a prevalence of 17 per cent is the highest the simulations can achieve. It occurs with regional background divorce, with no additional divorce if a spouse develops AIDS, and with remarriage. Once again, however, the scenario is unrealistic because the strategy creates as much divorce as in the real population where divorces are undoubtedly sometimes brought on by marital infidelity, which does not exist in our models. This time, however, an unrealistic scenario creates lifetime risk that is too high rather than too low. Lifetime risk, given uniformly faithful couples, is likely to be lower than this 17 per cent, although no lower than about eleven per cent.
We can comment only on the HIV-avoidance strategies we have tested. Of these, divorcing an unfaithful spouse is a reasonable strategy. Not remarrying thereafter, or after any other marriage has ended, for whatever reason, is better still. But marital fidelity is the best.
