Objectives: To identify the causes and future trends underpinning improvements in life expectancy in Scotland and quantify the relative contributions of disease incidence and survival.
Design: Population-based study.
Setting: Linked secondary care and mortality records across Scotland.
Participants: 1,967,130 individuals born between 1905 and 1965, and resident in Scotland throughout 2001-2016.
Main outcome measures: Hospital admission rates and survival in the five years following admission for 28 diseases, stratified by sex and socioeconomic status.
Results: The five hospital admission diagnoses associated with the greatest burden of death were "Influenza and pneumonia", "Symptoms and signs involving the circulatory and respiratory systems", "Malignant neoplasm of respiratory and intrathoracic organs", "Symptoms and signs involving the digestive system and abdomen", and "General symptoms and signs". Using trends in disease, we modelled a mean mortality hazard ratio of 0.737 (95% CI 0.730-0.745) across decades of birth, equivalent to a life extension of ~3 years per decade. This improvement was 61% (30%-93%) accounted for by improvements in disease survival after hospitalisation (principally cancer) with the remainder accounted for by a fall in hospitalisation incidence (principally heart disease and cancer). In contrast, deteriorations in the incidence and survival of infectious diseases reduced mortality improvements by 9% (~3.3 months per decade). Overall, health-driven mortality improvements were slightly greater for men than women (due to greater falls in disease incidence), and generally similar across socioeconomic deciles. We project mortality improvements will continue over the next decade but will slow down by 21% because much of the progress in disease survival has already been achieved.
Conclusion: Morbidity improvements broadly explain observed improvements in overall mortality, with progress on the prevention and treatment of heart disease and cancer making the most significant contributions. The gaps between men and women's morbidity and mortality are closing, but the gap between socioeconomic groups is not. A slowing trend in improvements in morbidity may explain the stalling in improvements of period life expectancies observed in recent studies in the UK. However, our modelled slowing of improvements could be offset if we achieve even faster improvements in the major diseases contributing to the burden of death, or if we improve prevention and survival of diseases which have deteriorated recently, such as infectious disease, in the future.