The Influence of Local Service Configuration on All-Cause Mortality: An Age- and Sex-Standardised Analysis

Public Health England mortality data shows that most deaths are associated with underlying longer-term health issues and could potentially be mitigated by more effective preventive and planned healthcare. General practitioner practices (GPPs) are the provider and gatekeeper to most NHS healthcare services and there are no direct publications of deaths at GPP level Our aim here was to determine how estimated mortality rate at GPP level may associate with the effectiveness of local services in delivering healthcare outcomes.

We used Office of National Statistics (ONS) data for January 2018-January 2020 (inclusive). From this an estimated Age Standardised Mortality Rate (SMR) for that Layer Super Output Areas (LSOA) was calculated. A linear link between local LSOA SMR and IMD was established so that the SMR_D could be calculated which assumes the practice was at median deprivation. Risk Factors that were then included into the Analysis these included: demographic, medical conditions, local use of preventive Medication (prescriptions) and impact of local area health services, including diabetes. Stepwise (removing factors with p value > 0.05) multifactorial linear regression was applied to derive the impact of these factors on the SMR_D RESULTS: The analysis covered 5792 GPP where all the data was available, and where more than 2000 patients were on the practice list. There were 22 factors taken into consideration for the model, of which 17 remained significant. If all practices achieved at least the median level there might be as many as 30,000 (7 %) fewer deaths/year. IMD even after adjustment within the LSOA SMR continued to have a strong effect. The association of Cancer QOF registration rates was associated with a lower mortality rate. A higher rate of diabetes case identification did not associate with reduced mortality rate, rather with higher mortality rate. Practice size was not a factor. Practices with higher percentage of older patients had relative lower mortality rates as did those with higher rates of antihypertensive prescribing.

We here describe associations that suggest that there are underlying themes to understanding the drivers to mortality in England. Our model can be applied to local practices to highlight those that have the largest gap. The association of general practice percentage identification of diabetes cases with increased practice mortality rate likely relates to the reality of type 2 diabetes being associated with many other conditions that can shorten life. The association of higher rates of prescription of antihypertensive agents in individuals over 65 years old with lower practice mortality, highlights the importance of effective identification and effective treatment of hypertension in this group.