Collaboration with health system partners should begin at the earliest stages of planning. Those early discussions should focus on refining questions to maximize relevance to health system partners and identifying outcomes most relevant to health system leaders and frontline clinicians. For health system leaders, decisions regarding adopting or implementing a new treatment or program depend on a range of outcomes beyond clinical efficacy, including total cost of care, member/patient satisfaction, staff satisfaction, impact on “report card” quality metrics, alignment with organizational goals, and community reputation (Bodenheimer and Sinsky 2014).

Early engagement with health system partners will allow pragmatic trial investigators to both identify relevant outcomes and evaluate specific strategies for assessing or ascertaining those outcomes. Some relevant outcomes can be efficiently and accurately assessed using routinely collected business or management information. Examples include use of billing information to assess cost of care, use of insurance enrollment information to assess member retention, and use of human resources data to assess staff turnover. In some cases, investigators may adjust or supplement patient-reported outcome measures to more closely match measures familiar to health system leaders. Examples include use of Medicare Consumer Assessment of Healthcare Providers and Systems (CAHPS) questions (Crofton et al 1999) to assess patient experience rather than patient satisfaction particular to a specific intervention or health condition, or use of staff engagement or satisfaction measures already routinely completed by health system employees. In some cases, investigators may plan additional analyses of health record data to address outcomes of interest to health system leaders. Examples include assessment of study interventions on publicly reported Healthcare Effectiveness Data and Information Set (HEDIS) quality metrics or on Merit-based Incentive Payment System (MIPS) metrics used for incentive payments.

While health system leaders’ decisions regarding adoption or implementation typically consider outcomes such as staff satisfaction and member or engagement satisfaction, those outcomes are often not systematically measured or thoroughly analyzed. In some cases, leaders’ assessment of staff or patient satisfaction may be limited to anecdotal praise or isolated complaints. Ideally, pragmatic trial investigators engage with health system leaders to identify and understand outcomes of interest and then contribute specific expertise regarding design, measurement, and analysis. In that mutually beneficial collaboration, researchers contribute to the rigor of health system decisions while increasing the relevance and likely impact of trial results.

Example of Early Engagement: Adapting and Implementing a Nurse Care Management Model to Care for Rural Patients with Chronic Pain (AIM-CP)

AIM-CP, an NIH Collaboratory Trial, is a 2-arm randomized controlled trial testing whether implementation of a nurse care manager model that provides 6 months of care coordination, brief cognitive behavioral therapy, and referrals to a virtual exercise program could reduce pain interference in rural patients with chronic pain (Tong et al 2026). In the trial feasibility phase, investigators received feedback from participating and interested health systems that many rural communities lacked a sufficient nursing workforce to participate since their nurses were needed to focus on core, day-to-day clinical activities. Obtaining ideas from interviews and informal conversations, investigators broadened the eligibility for serving as a care manager to any clinical staff who clinics felt could be trained to conduct the core functions of the care manager role. Instead of directly using nurses, investigators focused on ensuring that the nursing metaparadigm domains and whole-person health concepts behind nursing theory were covered in training and in the care manager model, allowing more rural health systems to participate in the study.