Comparing medication exposure before an admission with exposure a year earlier can remove some individual differences. It can also capture broader changes in prescribing. A Taiwanese claims study used a second comparison to address that problem.
What the case–time–control design added
The researchers defined unplanned hospitalization as admission immediately after an emergency department visit. They compared medication exposure one to 14 days before the visit with a window 366 to 379 days before the visit. They made the same comparison for matched reference visits. The second comparison was designed to account for changes in exposure over time unrelated to the admission itself. For benzodiazepines and Z-hypnotics grouped together, the adjusted odds ratio was 1.23, with a 95% confidence interval from 1.16 to 1.31.
Which changes can still affect the estimate?
The design compares windows with different exposure status and cannot fully account for changing symptoms that prompted both the prescription and the emergency visit. Claims also cannot confirm ingestion, and an unplanned admission can have many causes.
In expert review, identify the outcome and the precise exposure definition before citing the estimate. The study’s combined category cannot assign the same risk to each drug or show that a specific dose caused one admission.
Leave a Reply