Teaching and training practices: in the wrong places and with the wrong patients?

The problem

Whether practices which teach undergraduate medical students (teaching practices) or train postgraduate GP registrars (training practices) are representative of practices across England is uncertain. The associations between practices teaching and training status and their practice level outcomes are also unclear.

The approach

Practice level analyses (n=8,207) were conducted between English general practices’ undergraduate teaching and postgraduate training status and routinely collected and reported data on their socio-economic deprivation (Index of Multiple Deprivation), rurality (urban-rural indicator), workforce (number of GPs, full time equivalents, mean full time equivalence (FTE)), patient list sizes (including number of patients per FTE), patients’ health status (EQ-5D-5L), patient satisfaction (from the GP Patient Survey), and quality of care (from the Quality and Outcomes Framework).

Chi-square, one-way analysis of variances, univariate logistic regression analyses and multivariate linear regression analyses were used on practice level measures.

Findings

Of 8,207 English GP practices, 2,834 (34.5%) teach undergraduates and 2,301 (28.0%) train postgraduates. Practices which serve deprived populations are less likely to teach (Odds Ratio (OR): 0.78, 95%CI: 0.69, 0.88, p<0.001) and train (OR: 0.55, 95%CI: 0.48, 0.63, p<0.001). Rural and urban practices are equally likely to teach (OR: 1.02, 95%CI: 0.89, 1.15, p=0.82) but rural practices are more likely to train (OR: 1.55, 95%CI: 1.37, 1.77, p<0.001). Teaching and training practices have more GPs (F(3, 7772)=966.38, P<0.001) working less per week (p=0.002), with a smaller proportion of male GPs (F(3, 7772)=177.91, p<0.001). Teaching and training practices have more patients (F(3, 7700)=758.01, p<0.001), but fewer per FTE GP (F(3, 7700)=34.70, p<0.001). All five EQ-5D domain scores for patients of training practices reflect better health (all p<0.001) while scores for patients of teaching practices show no differences except for anxiety / depression which are lower than for non-teaching practices (F(3, 7764)=5.56, p=0.001). Teaching (B=.341, 95%CI: .233, .449, p<0.001) and training status (B=.601, 95%CI: .474, .729, p<0.001) were both associated with increased patient satisfaction. Teaching and training practices have higher scores on all QOF domains (p<0.001). Teaching status (B=.578, 95%CI: .282, .874, p<0.001) and training status (B=1.004, 95%CI: 0.656, 1.352, p<0.001) are both associated with total QOF score.

Consequences

Teaching and training practices are not representative of English general practices: they are more rural, and care for patients who are less deprived with better health status than practices which don’t. Training GPs in leafy suburb and bucolic rural practices has implications for recruitment to deprived urban practices. Urgent investment is needed to increase numbers of deprived urban area training practices. Teaching and training status are associated with better quality of care and patient satisfaction: providing more training in deprived urban practices may have positive consequences for care.