Trends in hospital deaths in Denmark from 1980 to 2014, at ages 50 and older
, Calduch Verdiell, N., Hale, J. M.
, Andersen-Ranberg, K., Lindahl-Jacobsen, R., Oksuzyan, A.
Journal of the American Geriatrics Society, 67:3, 471–476 (2019)
Objectives: To explore temporal trends and individual‐level determinants of hospital deaths at ages 50 and older in Denmark from 1980 to 2014.
Design: Individual‐level register‐based retrospective study.
Setting: Denmark, 1980 to 2014.
Participants: All deaths that occurred in Denmark from 1980 to 2014 among individuals 50 years or older (N = 1 834 437), extracted from population registers.
Measurements: A death was defined as a hospital death if the individual was admitted to the hospital as an inpatient and the date of discharge from the hospital is equal to the date of death.
Results: The percentage of hospital deaths decreased in both sexes (all ages combined, men: 56% to 44%; women: 49% to 39%) and at ages 50 to 79, remained almost unchanged at ages 80 to 89, and increased in the oldest age group (90+ men: 27% to 32%; women: 18% to 24%). We observed increasing trends of hospital deaths for three groups, people 90 years and older, dying from respiratory diseases, and who had terminal hospitalizations lasting 1 to 3 days. Subanalysis of all hospital deaths according to length of the terminal hospitalizations suggests that the overall reduction of hospital deaths might be driven by a reduction in hospitalizations that were longer than 1 week. Persons who are married, have middle or high income, have a history of hospitalizations in the year before death, or die because of respiratory diseases have higher odds of dying in a hospital.
Conclusion: Results provide evidence that Danes 50 years and older are increasingly dying outside the hospital context. We find three age‐specific patterns in the proportion of hospital deaths. Changes in healthcare and social systems implemented in Denmark during the observation period may underlie the broader reduction in hospital deaths in the country.