BackgroundCommunity health workers (CHWs) provide critical services to underserved populations in low and middle-income countries, but maintaining CHW’s clinical knowledge through formal continuing medical education (CME) activities is challenging and rarely occurs. We tested whether a Short Message Service (SMS)-based mobile CME (mCME) intervention could improve medical knowledge among a cadre of Vietnamese CHWs (Community Based Physician’s Assistants–CBPAs) who are the leading providers of primary medical care for rural underserved populations.MethodsThe mCME Project was a three arm randomized controlled trial. Group 1 served as controls while Groups 2 and 3 experienced two models of the mCME intervention. Group 2 (passive model) participants received a daily SMS bullet point, and were required to reply to the text to acknowledge receipt; Group 3 (interactive model) participants received an SMS in multiple choice question format addressing the same thematic area as Group 2, entering an answer (A, B, C or D) in their response. The server provided feedback immediately informing the participant whether the answer was correct. Effectiveness was based on standardized examination scores measured at baseline and endline (six months later). Secondary outcomes included job satisfaction and self-efficacy.Results638 CBPAs were enrolled, randomized, and tested at baseline, with 592 returning at endline (93.7%). Baseline scores were similar across all three groups. Over the next six months, participation of Groups 2 and 3 remained high; they responded to >75% of messages. Group 3 participants answered 43% of the daily SMS questions correctly, but their performance did not improve over time. At endline, the CBPAs reported high satisfaction with the mCME intervention, and deemed the SMS messages highly relevant. However, endline exam scores did not increase over baseline, and did not differ between the three groups. Job satisfaction and self-efficacy scores also did not improve. Average times spent on self-study per week did not increase, and the kinds of knowledge resources used by the CBPAs did not differ between the three groups; textbooks, while widely available, were seldom used.ConclusionsThe SMS-based mCME intervention, while feasible and acceptable, did not result in increased medical knowledge. We hypothesize that this was because the intervention failed to stimulate lateral learning. For an intervention of this kind to be effective, it will be essential to find more effective ways to couple SMS as a stimulus to promote increased self-study behaviors.Trial RegistrationClinicalTrials.gov NCT02381743