BackgroundIt is unclear whether a relationship exists between extreme heat and hospital length of stay in the UK. Using healthcare and climate data this study explores the effects of extreme heat on hospital length of stay for individuals hospitalised with cardiovascular disease, a common heat-sensitive condition, in Greater London. MethodsThis longitudinal study included patients registered to General Practices in Greater London (8·908 million inhabitants), hospitalised with cardiovascular disease during summer months of 2007–19. Administrative healthcare data was derived from Clinical Practice Research Datalink and Hospital Episode Statistics. Climate data from the Meteorological Office. The study exposure was extreme heat on day of hospital admission, defined using maximum daily temperature as per the UK Health Security Agency (low-level [28–31·9°C]; medium-level [32–39·9°C]; high-level [≥40°C]). The study outcome was hospital length of stay (days). In main analyses, linear fixed-effects models were used with fixed-effects for practice-by-calendar-day and practice-by-year and inclusion of covariates (daily rainfall, age, sex, deprivation, primary diagnosis, comorbidity). Findings24 435 hospital admissions were included. There were 72 low-level heat days, 11 medium-level heat days and 0 high-level heat days. There was no evidence of association between extreme heat and length of stay in unadjusted (0·03; 95% CI –0·10 to 0·15), and fully adjusted models (0·03; 95% CI –0·09 to 0·16). During medium-level heat, there was a significant association in unadjusted models (–0·27; 95% CI –0·50 to –0·03), and models adjusted by rainfall (–0·26; 95% CI –0·50 to –0·03), age (–0·25; 95% CI –0·45 to –0·04), sex (–0·25; 95% CI –0·46 to –0·05) and deprivation (–0·25; 95% CI –0·45 to –0·04). However, accounting for primary diagnosis (–0·15; 95% CI –0·44 to 0·14) and comorbidity (–0·17; 95% CI –0·46 to 0·12), there was no longer a significant association. InterpretationThis first UK-based study did not identify an association between low-level or medium-level heat on day of hospital admission and hospital length of stay. Further work is needed to evaluate the effects of extreme heat on the UK health system. FundingRP discloses funding from the Oxford Martin School as part of the Future of Cooling Programme. This funding allowed access to the data and paid for research assistant time to support the data management and partial analysis. The funder had no role in designing the question, the analysis or the interpretation of the results.
Read full abstract