BACKGROUND: As new female-initiated HIV prevention products enter development, it is crucial to incorporate women's preferences to ensure products will be desired, accepted and used. A discrete choice experiment (DCE) was designed to identify preferred features of a vaginallydelivered HIV prevention product.
METHODS: The DCE survey was c onducted among sexually active women aged 18-30 in Durban, South Africa and Harare, Zimbabwe. Participants were from two samples: productexperienced women from a randomized trial of four vaginal placebo forms (QUATRO) and product-naïve community members. In a tablet-self-administered survey, women were asked to choose between two hypothetical intravaginal HIV prevention products over 8 choice sets. Alternative designs were characterized by six attributes: HIV prevention efficacy (30%, 50%, 80%), pregnancy prevention, dosage (before sex, after sex, daily, monthly), mode of insertion (finger, applicator), wetness (none, some, a lot), and partner awareness during sex. Preference weights and relative importance scores were estimated using random-parameters logit.
RESULTS: In total, 395 women completed the survey (56% product-naïve). Median age w as 24 (IQR 21-26), 97% had a primary partner, and 69% had completed secondary school. Preferences did not vary by sample (p>0.90). HIV protection was the most important attribute, dominating stated choice. Yet, there were identifiable preferences among other features. Women preferred a product which also prevented pregnancy and caused some vaginal wetness (p< 0.001). They disliked a product that needs to be used every day (p=0.003), with no significant preference for other timing. A reusable applicator was favored to insertion by finger (p=0.001). Partner awareness was not a significant factor when choosing a product (p=0.09).
CONCLUSIONS: Efficacy was the most important attribute. However, features such as wetness and dosing may influence preference; there may be interest in a multipurpose product that also prevents pregnancy.
