The greatest burden of younger than 5 years mortality is in low- and middle-income nations where education resources are often few. The World Health Organization recommends scale-up of simulation in these settings, but it has been poorly studied. Although there has been an increase of contextualized resuscitation simulation programs designed for these settings, sustaining clinical outcomes and provider skill retention have remained research gaps. Our team designed a study to evaluate skill retention after an initial Helping Babies Breathe training at a rural Kenya referral hospital between randomized learner groups receiving supervised mock codes with debriefing versus just-in-time training with a peer. Although we saw sustained skills retention and some clinical improvements, we were unable to answer our research question because of numerous challenges, mainly that hospital leadership preferred the implementation of 1 arm of the study over another because of lack of protected education time and resources, eliminating differences between randomized study groups. Further challenges included lack of familiarity with simulation and debriefing and lack of protected educational resources and time, cultural differences in giving feedback, undeveloped systems for documentation, and high acuity and clinical volume. Our experience teaches many important lessons in how best to implement and study simulation in low-resource settings. Best practices include long-term partnerships, flexibility, community and staff engagement, mixed methodologies including community-based participatory methods, and careful attention to educational and research capacity building.