Objective Evaluate the effectiveness of home talks (HTs), a novel model of health education delivered by village health workers (VHWs) with primary-level education to rural African mothers. Talk recipients were assessed by health census to be at risk for ill-health in one of 5 ways: malnutrition, diarrhea, respiratory disease, HIV, and poverty due to family size. Methods Each participant received a pre-test, immediate post-test and delayed post-test on their assigned HT topic and a pre-test and delayed post-test on a randomly assigned control topic. Differences in scoring were examined against controls and over time using paired t-tests and general linear regression analysis, respectively. Results Subjects lost knowledge gained from the HTs over time, but what they retained at 3 months was far greater than what they learned about the control topics (p-values Conclusion Targeted HTs to people with health census-identified risk factors resulted in learning and significant retention of knowledge. Practice implications Positive behavioral change resulting from health education has been shown in diverse contexts. This personal model of home talk education by VHWs targeting vulnerable families is flexible and effective and may be used to improve community health in other impoverished settings worldwide.