Do general practices in England influence life expectancy?

Problem

Life expectancy at birth in England steadily increased throughout the twentieth century, but the rate of improvement has slowed and, before the pandemic, even reversed in some patient groups. Office for Health Improvement and Disparities data show life expectancy reductions of females in 43.0% and of males in 39.7% of English practices between the 2013-17 and 2015-19 quinquennia used to calculate practice-level life expectancy. Such changes probably have several causes, but the role of primary care is seldom studied. For some years, there has been a growing mismatch between populations’ health care needs and practices’ capacity to respond.

Approach

We undertook an ecological cross-sectional study. The outcomes, life expectancy at birth for males and females of general practice populations in England for the quinquennium 2015-2019, were obtained for 6553 general practices (all with a list size of 750 or greater). Selection of independent variables was guided by frameworks of mechanisms through which general practice influences health outcomes. We eliminated highly correlated variables and those failing to predict life expectancy at p<0.05 in univariate regression. The 11 retained time-matched variables were entered into weighted regression models in blocks: population (practice Index of multiple Deprivation scores, NHS Region, % white ethnicity, morbidity [% on diabetes register]); practice organization (full time equivalent [FTE) GPs, nurses and receptionists per unit of population); access (% seen on the same day of requesting an appointment); clinical performance (two quality and outcomes framework indicators: % of people aged 45 or older whose BPs were checked and % with diabetes whose BP was controlled): and the therapeutic relationship (% relationship continuity, from the general practice patient survey).

Findings

Mean life expectancy was 83.3 years (females) and 79.4 years (males). In both models, deprivation was a strong and significant negative predictor. Regions outside London and white ethnicity were also significant negative predictors. Number of FTE GPs (coefficient, 95% confidence intervals and p values: females 0.76 [0.60 -0.92], <0.001; and males 0.76 [[0.61 – 0.92], <0.001; and continuity: females 0.50 [0.28 – 0.72], <0.001; males 0.52 [0.30 – 0.74], <0.001, were significant positive predictors, as was % with diabetes with controlled BP. However, % over 45s with a BP check was a significant negative predictor only in females, and % on diabetes registers and % being seen on same day were significant positive predictors only in males. The remaining variables were not statistically significant. The adjusted R2 in the final models was 0.70 for females and 0.77 for males.

Consequences

These findings suggest that declines in the numbers of FTE GPs and in relationship continuity may be contributing to stalling or falling life expectancy in England. These deteriorating general practice factors are serious and potentially remediable.