Abstract
Numerous NGOs and donors from high-income countries (HICs) offer diverse funding to assist LMICs. A collaborative effort in nutrition programs in urban communities was conducted by an international NGO with a local university, representing the NGO's first mode of partnership with academia. This study used realist evaluation to understand how and why Positive Deviance/Hearth intervention conducted by NGOs collaborated with university work or failed to work in urban population. We uncovered the underlying mechanisms that lead to certain outcomes in different situations of academia-NGO partnership. The initial program theory was formulated after discussions with main program officers from both NGO and academia. We conducted stakeholder interviews and FGD, along with validation and reviewing secondary data. The initial program theory underwent testing and refinement through a series of Focus Group Discussions (FGDs) involving program officers and local government health offices. We adhered to the RAMESES II reporting standards for realist evaluations when presenting the findings. Five Context Mechanism-Outcome (CMO) configurations are identified from three program theories. Our findings illustrate the importance of considering the local context of the population when implementing standardized international NGO interventions. All the CMOs indicate that at the very least, collaborative programs between academia and NGOs should involve more stakeholder involvement and build clearer expectations between stakeholders.