Detecting channeling bias after launch – implications for comparative effectiveness studies: a case study in antihypertensive medicines


Early evaluation of the effects of emerging therapies is often challenging because of confounding factors caused by channeling bias (when a newly marketed medicine and an established medicine with similar therapeutic indications are prescribed to different patients according to their prognostic characteristics). This case study aimed to explore time trends in differences in confounder distributions between users of antihypertensive medicines.

What was examined in this case study?

A focused literature review on PUBMED was conducted on observational comparative effectiveness studies of antihypertensive medicines. The review included studies comparing treatment with angiotensin converting enzyme (ACE) inhibitors and calcium channel blockers vs. diuretics and beta-blockers since the launch of ACE inhibitors and calcium channel blockers. For each study, information was extracted on baseline characteristics, duration of follow-up, exposure and outcome.

Differences in patient characteristics between exposure groups were then assessed over calendar years.

What were the findings and conclusions?

  • Forty studies published between 1996 and 2013 were included for the analysis.
  • Major patient characteristics often reported in the studies were age, gender, body mass index (BMI), baseline systolic and diastolic pressure, smoking, diabetes, dyslipidaemia, stroke, ischaemic heart disease, and heart rate.
  • The mean difference in baseline systolic and diastolic blood pressure and smoking status between ACE inhibitor/calcium channel blocker users and diuretic/beta-blocker users decreased overtime.
  • No pattern was observed for age, gender, diabetes, and BMI.
  • Groups of antihypertensive medicine users become more similar for some patient characteristics over time after launch. However, this was not observed for all characteristics, and no time period could be identified that was optimal for observational comparative effectiveness research.

Key contributor

Sanni Ali, UMCU