| Outcome | Details |
|---|---|
| A&E attendances | All A&E attendances at all types of emergency care departments and providers were included. For financial year 2019–2020 we used HES A&E data rather than the new Emergency Care Data Set, to maintain consistency across the whole analysis timeframe. Multiple attendances on the same day were only included once. |
| Outpatient attendances | Only outpatient appointments marked as having been attended were included. |
| Inpatient admissions | Inpatient admissions are recorded in HES as a series of ‘Finished Consultant Episodes’ (time spent under a particular consultant’s care). Sometimes a patient’s stay in hospital includes successive periods under the care of different consultants. We linked these episodes to form single admissions using the University of York Centre for Health Economics 'Continuous Inpatient Spell' definition. Because the person identifier for HES changed during the 2019–2020 financial year, we used the mapping files provided by NHS Digital to allow for continuous inpatient spells that started in financial year 2018–2019 but finished in 2019–2020 (the date across files where the person identifier changed). |
| Inpatient admissions with an ambulatory care sensitive condition | Admissions were flagged if they were related to an ambulatory care sensitive condition (ACSC) based on the classification used by Bardsley et al.24 |
| Emergency inpatient admissions | Admissions were defined as emergency admissions based on the HES data classification. |
| Emergency readmissions | We used the University of York Centre for Health Economics definition for readmissions within 30 days of discharge, and included only emergency-coded admissions. |
| Length of stay | Days calculated for continuous inpatient spells, and included based on the date of admission. |
A&E = accident and emergency. HES = Hospital Episode Statistics