A study estimating later fall risk may deliberately exclude the earliest falls, changing the population to which its results apply.
The Medicare study included 102,750 older adults with depression initiating antidepressants, with or without benzodiazepines. Researchers modeled daily medication patterns during the first eighty-four days, then examined fall-related injuries during the following year. Twelve trajectories represented different dose levels and trends.
People who died or experienced a fall-related injury during the trajectory-measurement window were excluded. This allowed enough exposure information to assign a pattern, but it means the subsequent risk estimates do not describe all people starting treatment. They also do not directly estimate injuries occurring immediately after initiation.
Some higher or sustained combined-exposure patterns were associated with greater later injury risk, while very-low or declining benzodiazepine patterns did not clearly add risk in selected comparisons. These are observational associations based on prescriptions. Actual ingestion, symptoms, frailty and treatment selection can differ despite adjustment.
First locate the individual event relative to treatment initiation. A fall within the initial exposure window cannot be directly characterized by an estimate derived after excluding such events. Conversely, a nonsignificant trajectory comparison does not establish that a medicine could not contribute to a particular fall. Review the complete medication timeline, function, environment and alternative causes before drawing an individualized conclusion.
Source: PMC12366386, supplied full-text paper. Findings are paraphrased; forensic implications are an interpretation.
Related: A Hip-Fracture Surgery Cohort Could Not Estimate Risk Among People Without Fractures.