General practice characteristics associated with emergency hospital attendances for ambulatory care sensitive conditions

BACKGROUND: Ambulatory care sensitive conditions (ACSC) can be managed in primary care and high-quality care should reduce emergency hospital attendances (EHA) due to ACSC. AIM: To identify general practice characteristics associated with different levels of EHA for ACSC. METHOD: Composite index analysis, including agglomerative hierarchical clustering. Data (2019) from the English general practice workforce minimum data set (wMDS), openly available practice indicators, and hospital-linked routine pseudonymised primary care records. Clustering was performed at practice level. Logistic regression, controlling for patient-level factors, explored if the odds of EHA differed across clusters. RESULTS: Two clusters were identified. Cluster 1, n = 281 practices, 3 175 300 patients - the partnership training model, had: more practices with a GMS contract (70.1% versus 45.4%, P<0.001) and providing GP training (73.0% versus 38.0% P<0.001), partners accounting for a higher proportion of total GP FTE (63.4% versus 29.5%, P<0.001), fewer encounters per patient per year (10 versus 20), higher proportions of patients reporting good experiences (85.9% versus 79.2%, p<0.001). Cluster 2, n = 97 practices, 1 163 030 patients - the multiprofessional urban deprived model had: lower proportions of practice FTE performed by GPs (19.4% versus 23.9%, P<0.001), higher proportions of encounters delivered by other healthcare professionals (15.5% versus 5.7%, P<0.001), more urban practices (90.7% versus 80.0%, P<0.01, greater deprivation (P = 0.02). The odds of EHA in Cluster 1 was 11% lower than Cluster 2 (P<0.0001). CONCLUSION: These findings have implications for policies about the partnership model, and current trends that encourage more encounters with a multiprofessional team.