Long-term outcomes for patients at high risk of death after surgery in the UK: a retrospective cohort study.

Abstract

More than 5 million surgical procedures are performed every year in the UK, but little is known about individuals at high surgical risk who account for most deaths after surgery. We aimed to describe the size, demographics, and long-term outcomes of high-risk surgical patients in the UK. In this retrospective, observational cohort study, we analysed routine National Health Service (NHS) data from England (hospital episode statistics, and admitted patient care), Scotland (Scottish Morbidity Records), and Wales (Patient Episode Database for Wales). We derived death data from linkage to Office for National Statistics (England & Wales), National Records of Scotland (Scotland), and the Welsh Death Service (Wales). We included adults (aged ≥18 years) undergoing elective or emergency, in-patient or day-case, non-obstetric surgery between Jan 1, 2015, and Dec 31, 2019 (Dec 31, 2018, in Wales) using an existing code-based definition. We included patients in sequential surgical spells (from day of surgery to 365 days follow-up) and patients who underwent more than one surgical procedure over the study period could contribute more than one surgical spell. The primary (explanatory) outcome was death within 90 days after the index surgery. Secondary outcomes were duration of hospital stay, 90-day emergency hospital readmission, 1-year mortality, and 5-year survival. Patients were ranked from highest to lowest risk of 90-day death using a logistic regression model. The high-risk group included patients accounting for the first 80% of deaths, moderate-risk group included those accounting for the next 15% of deaths, and the low-risk group included those accounting for the final 5% of deaths. We compared high-risk and low-risk groups with risk ratios (RR; 95% CI). We included 12 905 737 patients (mean age 57·5 [SD 19·2] years; 8 584 593 [53·2%] women, 7 552 101 [46·8%] men; 12 086 325 [74·9%] White patients) who experienced 16 136 694 surgical spells. There were 262 500 (1·6%) deaths within 90 days, and 644 596 (4·0%) deaths within 1 year. The median length of stay was 0 days (IQR 0-1), and there were 1 506 299 (9·4%) emergency readmissions within 90 days after surgery. We identified 1 481 357 (9·2%) patient spells in the high-risk group, who had experienced 206 257 (13·9%) deaths within 90 days (RR vs low 117·15 [115·18-119·14]), a median length of stay of 10 days (IQR 3-21), 426 496 (30·9%) emergency readmissions (RR vs low 5·67 [5·65-5·69]), and 402 033 (27·1%) deaths within 1 year (RR vs low 38·89 [38·61-39·17]). As a proportion of all outcomes in the cohort, high-risk patients accounted for 53·8% of all hospital bed-days used after surgery (21 206 482 of 39 441 408), 28·3% of emergency readmissions (426 496 of 1 506 299), and 62·4% of 1-year deaths (402 033 of 644 596). In England, 94·9% (95% CI 94·8-94·9) of patients in the low-risk group survived by 5 years, compared with 75·1% (75·0-75·2) of patients in the moderate-risk group, and 40·3% (40·1-40·4) of patients in the high-risk group. Data from 2015-19 suggest that around 300 000 high-risk patients undergo surgery each year in the NHS. This group have poor outcomes and very high health-care resource needs, which presents important health policy and delivery challenges in the UK. National Institute for Health and Care Research (NIHR).