Engagement and collaborative learning should continue throughout the implementation of a pragmatic trial, including both pilot testing and full-scale implementation. Pragmatic trial investigators should actively encourage anecdotal feedback from frontline health system staff and health system leaders regarding unexpected characteristics of study participants, barriers to or variability in implementation of study interventions, gaps or discontinuities in study data sources, and concerns regarding integrity of study interventions or outcome assessments. Those anecdotes may lead to adjusting or expanding study data collection to include additional measures of fidelity or additional outcome measures.

Ideally, study teams systematically assess and record processes and outcomes of interest to health system decision-makers and frontline staff throughout trial implementation. As discussed in Sections 1 and 2 of this chapter, systematically assessing those relevant processes and outcomes may require additional primary data collection or use of business and management data resources less familiar to clinical researchers.

Anecdotal experience during a pragmatic trial can also inform a more systematic approach to debriefing (discussed in Section 4 of this chapter) and secondary or exploratory analyses (discussed in Section 5 of this chapter).

Example of Learning Along the Way: Fibromyalgia TENS in Physical Therapy Study (FM-TIPS)

FM-TIPS, an NIH Collaboratory Trial, was a cluster randomized pragmatic trial that evaluated whether the addition of transcutaneous electrical nerve stimulation (TENS) to outpatient physical therapy improved fibromyalgia-associated, movement-evoked pain (Dailey et al 2026). Early review of health record data from 28 outpatient physical therapy clinics showed lower than expected use of fibromyalgia diagnostic codes, higher than expected recording of nonneuromusculoskeletal codes, and variation in diagnostic coding according to pathway to physical therapy (referral vs direct access). Consultation with clinic directors enabled the study team to understand variations in coding and effects of that variation on eligibility, recruitment, and assessment of intervention effects on care processes.