Variation in patient experience between general practice contract types: multilevel analysis of a national cross-sectional survey

Abstract

BackgroundRecent health-care reforms in England have opened up the provision of general practice to new models of care, and primary care can now be provided under various contractual mechanisms. Previous research has highlighted population-level variation in quality indicators across practices under different contract types. We present differences in patient-reported experience of general practice services across three contract types in England. MethodsData come from the national General Practice Patient Survey 2013–14 (903 357 responders from 7949 general practices in England; 34% response rate, range 29–44). The primary outcome was overall patient experience of a general practice (on a five-point scale, rescaled to 0–100 for all measures). Sociodemographic data included patient age, sex, ethnicity, deprivation, and self-reported health, and the profile of each practice's registered population. We used mixed-effects linear regression, using a random intercept for each general practice to estimate case-mix-adjusted associations between contract type and patient experience. FindingsThe mean score for overall experience across all practices was 83·3 (SD 20·3), indicating a fairly good experience nationally. Practices run as limited companies were uncommon (n=118, 1·5%) but overall their levels of patient experience were lower. Relative to general medical services (GMS) practices (mean 83·7), the adjusted means for overall experience for limited company-owned alternative provider of medical services (APMS) and personal medical services (PMS) practices were 3·0 (p<0·0001) and 3·6 (p=0·0176) lower. Results were similar for the other indicators: ability to see a preferred doctor (GMS mean 70·3; mean difference limited company APMS −12·8 [p<0·0001], mean difference limited company PMS −11·6 [p=0·0003]), appointment convenience (78·5; −3·0 [p<0·0001], −2·8 [p=0·1330]), doctor communication (84·1; −2·9 [p<0·0001], −3·5 [p=0·0052]), and ease of telephone contact (69·6; −0·2 [p=0·5527], −6·2 [p=0·0465]). InterpretationGeneral practices run as limited companies had worse scores for overall patient experience. Strengths of this study include the large, nationally representative sample. Limitations include the possibility for residual confounding, and the cross-sectional nature of available data. The results support continued monitoring of quality of care across different forms of general practice service provision. FundingThis research was supported by the National Institute for Health Research (NIHR) (Doctoral Research Fellowship for TEC, DRF-2013-06-142).