Who is responsible? Defining a hospital catchment population in the English National Health Service

Abstract

Abstract Background There is a growing emphasis on National Health Service hospitals in England promoting population health. Patients can access any hospital, making it complex to define the population a hospital is responsible for. Defining this 'catchment' population is fundamental to provide a population denominator from which to evaluate service provision such as unmet need and the effect of prevention initiatives. Using Chelsea and Westminster Hospital NHS Foundation Trust (CWFT), a large hospital in London as a case study, methods to define the population that has potential to attend the hospital were compared. Methods Inpatient, outpatient and emergency attendances were identified using Hospital Episode Statistics from 1st April 2017-31st March 2018. Lower Layer Super Output Areas (LSOAs), consisting of 1,500 people on average, were used as the geographic unit. Catchment populations were constructed using 3 different methods. Under First-Past-The-Post (FPTP), LSOAs were allocated if a greater proportion of patients attended CWFT than any other hospital trust. Under 30% Proportional Flow (30PF), LSOAs were allocated if more than 30% of patients attended CWFT, while under Stratified Proportional Allocation (SPA), patients were assigned in accordance with the proportion from each LSOA that attended CWFT, by gender and 5-year age strata. Results Under FPTP, 30PF and SPA, a total of 303, 326 and 10,636 LSOAs were assigned to CWFT, respectively, with corresponding populations of 530,980, 569,682, and 484,249 and median ages of 36, 36 and 29 years. Under FPTP, the catchment area did not overlap with that of any other hospital, while under 30PF, 13.2% of the LSOAs were also allocated to another hospital catchment. Maps were constructed for FPTP and 30PF. Conclusions The 3 methods produced different catchment populations, with differing characteristics. Understanding the relative merits of each method has implications for hospitals in how they engage in and evaluate population health. Key messages Engagement in and evaluation of public health activities requires knowledge of ‘who’ the baseline population denominator is, but there is no consensus on determining this at the hospital level. Comparing 3 methods to define a hospital catchment, these differ by total population, median age and geography, the choice of which impacts on how hospitals engage in population health.