Factors related to abstinence among rural Malawian adults

Background: A peer group intervention promoted abstaining as one effective HIV prevention strategy, but abstaining increased only slightly. It is important to know who abstains and what factors foster abstaining.

Objective: To examine impacts of demographics, the intervention, and possible mediators on abstaining in the last two months. Methods: Data come from a quasi-experimental study comparing an intervention and a control district. Demographic factors include: relationship status (living with spouse vs. not currently married or spouse not currently present), gender, age, and education. Potential mediators of the intervention's effects are positive condom attitudes, safer sex self-efficacy, and HIV knowledge. Logistic regression estimates the explained variance and effects of individual predictors.

Results: Relationship status has a substantial impact on abstinence: among adults living with their spouses (n=584) only 10.4% abstained, compared to 72.5% among adults not currently married or with spouse not present (n=241). For adults not currently married or with spouse not present, the intervention and demographic factors explained 11.2% of the variance in abstinence; adding the mediating variables explained a total of 14.4%. Statistically significant (p<0.05) individual predictors were: age > 50 years and HIV knowledge. For adults living with their spouses, the intervention and demographic factors explained 5.3% variance in abstinence; the mediating variables explained a total of 7.1%. Significant predictors were: the intervention and age > 50 years. HIV knowledge was negatively related to abstinence.

Discussion: Interventions need to consider the relationship status of adults when designing content focused on abstinence in African societies like Malawi.