General practice workforce data reveal impact on emergency hospital attendances in England

Background Policy of the English National Health Service recently has encouraged greater diversity in professional roles. Alternative forms of ownership have developed in place of the traditional partnership model. Aim To investigate the impact of primary care structure on emergency hospital attendances. Design & setting Analysis of routine data for 766 general practice members of the English primary care sentinel network, 2019. Method Composite index analysis, using agglomerative hierarchical clustering applied to practice-level variables from the workforce Minimum Data Set, quality indicators, and hospital-linked routine pseudonymised primary care records. Logistic regression assessed whether odds of an emergency hospital attendance for an ambulatory care sensitive condition differed between clusters, controlling for patient-level factors. Results Two distinct clusters were identified. Cluster 2 (97 practices, 1.16 million patients) had 11% higher odds of an emergency hospital attendance than Cluster 1 (281 practices, 3.18 million patients) ( P <0.0001). Cluster 1 reflected a partnership training model having practice partners accounting for a higher proportion of total general practitioner FTE (63.4% vs 29.5%, P <0.001) and more practices providing training (OR 4.35, 95% CI 2.61–7.34). Cluster 2 was typified as multiprofessional, urban and more deprived having more encounters delivered by non-medical patient care staff (15.5% vs 5.7%, P <0.001), more practices in urban settings (OR 2.48, 95% CI 1.16–5.95) and greater deprivation (chi-square P <0.02). Despite workforce and care delivery differences, practice list sizes were similar across clusters. Conclusions These findings have implications for determining the optimal structure of general practices and the future of the partnership model.