Background: Several mass media health behavior campaigns (HBCs) were conducted in Swaziland between 2010-2015 to avert new HIV infections. We describe the association between HBCs and specific behaviors among a prospectively followed, nationally representative, HIV-negative cohort from the 2011 Swaziland HIV Incidence Measurement Survey (SHIMS).
Methods: Survey staff conducted household-based HIV testing and collected information from adults, 18-49 y, about risk behaviors and exposure to HBCs in the previous 6 months. HIV-negative participants repeated the questionnaire and HIV testing 6 months later. Survey data were weighted for sampling design and multivariable logistic regression analysis assessed associations between HBC exposure and risk behaviors among those reporting sexual activity at both baseline and follow-up. Assessed behaviors were ever HIV testing (at baseline); number of sexual partners; condom use; and male circumcision (MC) or having a circumcised primary partner.
Results: Among 18,172 adults, 12,368 tested HIV-negative at baseline; of these, 11,232 (91%) completed a 6-month follow-up visit and 121 incident infections were observed. Among women, there was a baseline association between reporting ever HIV testing and exposure to HIV testing HBC messages (adjusted OR [aOR] 3.84, 95% CI [1.18-12.55]); and reporting fewer (i.e., 1 versus 2, or 2 versus ≥3) sexual partners and exposure to partner reduction HBC messages (aOR = 3.02, 95% CI [1.38-6.62]). Among men, there was a baseline association between reporting fewer partners and exposure to partner reduction messages (aOR = 2.26, 95% CI [1.49-3.44]). In the longitudinal assessment, men who reported baseline exposure to partner reduction HBCs were more likely to report fewer sexual partners 6 months later (aOR = 1.95, 95% CI [1.26-3.00]). No significant association was found for men or women between exposure to HBCs about condoms and condom use, nor about MC and the MC outcome [Table]. Exposure to HBCs did not predict incident HIV infections.
Conclusions: These findings suggest a minimal impact of mass media HBCs on promoting low-risk behaviors among adults in Swaziland, the country with the highest global national HIV prevalence. Although both men and women appeared to respond positively to partner reduction HBC messages, averting risky behaviors at a population level appears to require additional interventions.



[Weighted Logistic Regression Modeling Low-Risk Behavior for HBC Exposures at Baseline and Risk Behavior Outcomes at Baseline and Follow-Up]

Download the e-Poster