Immeasurable Time Bias in Self-controlled Designs: Case-crossover, Case-time-control, and Case-case-time-control...

Area of Science:

• Pharmacoepidemiology
• Biostatistics
• Geriatric Medicine

Background:

• Immeasurable time bias (IMTB) is a potential confounder in self-controlled study designs.
• The impact of IMTB has not been previously evaluated in self-controlled epidemiological analyses.

Purpose of the Study:

• To assess the impact of IMTB on drug-effect estimates in self-controlled designs.
• To compare the influence of IMTB on acute events versus mortality outcomes.

Main Methods:

• Utilized case-crossover, case-time-control, and case-case-time-control designs with Korean healthcare data.
• Examined two examples: benzodiazepine use and hip fracture, and benzodiazepine use and mortality in elderly patients.
• Compared results from non-missing exposure and pseudo-outpatient settings to quantify IMTB.

Main Results:

• IMTB had negligible effects on the association between benzodiazepines and hip fracture.
• IMTB significantly biased the estimated effect of benzodiazepines on mortality, with varying magnitudes across designs.
• The pseudo-outpatient setting showed a notable reduction in bias for the mortality outcome compared to the non-missing exposure setting.

Conclusions:

• IMTB has minimal impact on estimating drug effects for acute events unlikely to involve prodromal hospitalizations.
• IMTB can substantially bias drug-effect estimates for mortality outcomes, which often have prodromal phases.
• Findings underscore the importance of addressing IMTB in self-controlled studies, particularly for chronic or mortality-related outcomes.