Abstract
OBJECTIVES: To determine the longitudinal development of drug use in very old adults. DESIGN: Longitudinal cohort study with waves in 1998, 2000, 2002, and 2005. SETTING: Nationwide study in Denmark. PARTICIPANTS: All living Danes born in 1905 were approached in 1998; 2,262 responded at baseline. MEASUREMENTS: Self-reported use of regularly taken drugs. Mean and median number of drugs and growth curve models were used to identify the change in number of drugs as the cohort aged from 92 to 100. RESULTS: The within-person use of drugs increased with age for women (0.19 per year; 95% confidence interval (CI) = 0.15-0.24) and men (0.15 per year; 95% CI = 0.06-0.24). Persons leaving the study prematurely had higher baseline values and a steeper increase in their annual use of drugs. The population-level mean number of drugs increased from baseline (3.6 drugs) to the first follow-up (4.1 drugs) but thereafter remained stable at approximately 4 drugs. Women used more drugs than men at all waves. CONCLUSION: In this first longitudinal study of drug use in nonagenarians, individuals used an increasing number of drugs as they aged. This increase is difficult to detect in cross-sectional analyses of the population-level mean. More efforts to understand what is reasonable prescribing at these older ages are needed.