The Power of Data: Assessing Primary Care Performance Using Routinely Collected Emergency Department Data

This paper develops and validates an approach to assess the impact of individual primary care practices (PCPs) on demand for emergency department (ED) services. Using operational data from two geographically adjacent EDs in England, we focus on attendances for ambulatory care-sensitive (ACS) conditions, which are more appropriately and cost-effectively treated outside the ED. We employ a variance-decomposition approach that controls for observed and unobserved heterogeneity and explicitly accounts for socioeconomic differences. The analysis shows that approximately 52% of the variation in ACS attendance proportions is attributable to systematic differences across PCPs and that these differences are operationally meaningful. Using this methodology, we construct a PCP-level measure of systematic variation in ACS attendances and show that it correlates with patient feedback from National Health Service surveys, Quality and Outcomes Framework scores, and PCP staffing decisions. This measure outperforms existing metrics based on ACS admissions in identifying PCP-level variation and is robust to alternative modeling choices. The modular nature of the approach allows it to be applied to data from one or multiple hospitals, providing policymakers with a practical tool to evaluate PCP-level interventions and their downstream effects on ED demand. This paper was accepted by Carri Chan, healthcare management. Funding: The authors acknowledge financial support from the London Business School Research and Materials Development Fund and from the Erasmus School of Health Policy & Management Career Development Research Fund. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2023.03234 .