Rethinking morbidity compression

Studies of morbidity compression routinely report the average number of years spent in an unhealthy state but do not report variation in age at morbidity onset. Variation was highlighted by Fries (1980) as crucial for identifying disease postponement. Using incidence of first hospitalization after age 60, as one working example, we estimate variation in morbidity onset over a 27-year period in Denmark. Annual estimates of first hospitalization and the population at risk for 1987 to 2014 were identified using population-based registers. Sex-specific life tables were constructed, and the average age, the threshold age, and the coefficient of variation in age at first hospitalization were calculated. On average, first admissions lasting two or more days shifted towards older ages between 1987 and 2014. The average age at hospitalization increased from 67.8 years (95% CI 67.7–67.9) to 69.5 years (95% CI 69.4–69.6) in men, and 69.1 (95% CI 69.1–69.2) to 70.5 years (95% CI 70.4–70.6) in women. Variation in age at first admission increased slightly as the coefficient of variation increased from 9.1 (95% CI 9.0–9.1) to 9.9% (95% CI 9.8–10.0) among men, and from 10.3% (95% CI 10.2–10.4) to 10.6% (95% CI 10.5–10.6) among women. Our results suggest populations are ageing with better health today than in the past, but experience increasing diversity in healthy ageing. Pensions, social care, and health services will have to adapt to increasingly heterogeneous ageing populations, a phenomenon that average measures of morbidity do not capture. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s10654-020-00642-3) contains supplementary material, which is available to authorized users.