Abstract
OBJECTIVES: This study tested the theoretically grounded conceptual model of a multi-level intervention, Family Health = Family Wealth (FH = FW), by examining FH = FW's effect on intermediate outcomes among couples in rural Uganda. FH = FW is grounded in the social-ecological model and the social psychological theory of transformative communication. DESIGN: A pilot quasi-experimental controlled trial. METHODS: Two matched clusters (communities) were randomly allocated to receive the FH = FW intervention or an attention/time-matched water, sanitation and hygiene intervention (N = 140, 35 couples per arm). Quantitative outcomes were collected through interviewer-administered questionnaires at baseline, 7-months and 10-months follow-up. Focus group discussions (n = 39) and semi-structured interviews (n = 27) were conducted with subsets of FH = FW participants after data collection. Generalized estimated equations tested intervention effects on quantitative outcomes, and qualitative data were analysed through thematic analysis-these data were mixed and are presented by level of the social-ecological model. RESULTS: The findings demonstrated an intervention effect on family planning determinants across social-ecological levels. Improved individual-level family planning knowledge, attitudes and intentions, and reduced inequitable gender attitudes, were observed in intervention versus comparator, corroborated by the qualitative findings. Interpersonal-level changes included improved communication, shared decision-making and equitable relationship dynamics. At the community level, FH = FW increased perceived acceptance of family planning among others (norms), and the qualitative findings highlighted how FH = FW's transformative communication approach reshaped definitions of a successful family to better align with family planning. CONCLUSIONS: This mixed methods pilot evaluation supports FH = FW's theoretically grounded conceptual model and ability to affect multi-level drivers of a high unmet need for family planning.