Utilisation of a cocreation methodology to develop claims-based indicators for feedback on implementation of comparative effectiveness research results into practice.

Introduction: Comparative effectiveness research (CER) often fails to create quality improvement since implementation of CER results in clinical practice is lacking. Claims-based Audit & Feedback (A&F) provides a resource efficient tool to stimulate implementation, but it is unknown whether medical professionals accept claims-based A&F in the context of CER. Therefore, in this study, we developed claims-based indicators using a cocreation approach and evaluated medical professionals' perception regarding the validity and acceptability of these indicators.

Methods: Between July 2019 and November 2021, we used a cocreation approach with medical experts to develop claims-based indicators for six CER trials. The aim is to use the indicators for group level feedback on implementation of CER results to medical professionals across all healthcare providers in the Netherlands. To build the indicators, we used the most recent available Dutch national healthcare-related claims data of the year 2017. The cocreation process consisted of the following steps: (1) defining the target indicator, (2) selecting relevant claims codes, (3) testing feasibility of the indicators using Dutch claims data, (4) discussing results of feasibility testing and (5) defining the final indicators and reflecting on the acceptability of the indicators for feedback on implementation of CER results by the experts.

Results: Claims-based indicators could not perfectly reflect the CER population for any of the six CER trials. However, the cocreation process did lead to a final indicator that medical experts found acceptable in four of six cases. Recommendations of medical experts for improving claims-based indicators included: select patients with minimal over- or underestimation of the CER population, use proxies to identify patients, determine incidence rather than prevalence for chronic conditions and use data linkage with diagnostic test results.

Conclusion: A cocreation approach was a successful way to develop claims-based indicators on implementation of CER results, which were imperfect, but in some cases still acceptable as feedback to medical experts. Thus, for certain topics, claims data may provide a resource efficient data source for A&F interventions aiming to implement CER trials.