Abstract Objective To support implementation of effective treatments for back pain that can be delivered to a range of people, we summarize learnings from our process evaluation of the MATCH trial’s implementation of an adaptation of the STarT Back risk-stratified care model. Design Our logic model–driven evaluation focused primarily on qualitative data sources. Setting This study took place in a US-based health care delivery system that had adapted and implemented the STarT Back stratified care approach. This was the first formal test of the strategy in a US setting. Methods Data collection included observation of implementation activities, staff/provider interviews, and post-training evaluation questionnaires. Data were analyzed using thematic analysis of qualitative data and descriptive statistics for questionnaire data. Results We found that both primary care teams and physical therapists at intervention clinics gave the training high scores on evaluation questionnaires and reported in the interviews that they found the training engaging and useful. However, there was significant variation in the extent to which the risk stratification strategy was incorporated into care. Some primary care providers reported that the intervention changed their conversations with patients and increased their confidence in working with patients with back pain. Providers using the STarT Back tool did not change referral rates for recommended matched treatments. Conclusions These insights provide guidance for future efforts to adapt and implement the STarT Back strategy and other complex practice change interventions. They emphasize the need for primary care–based interventions to minimize complexity and the need for ongoing monitoring and feedback.