Evidence selection for a prescription drug's benefit-harm assessment: Challenges and recommendations

Kevin M. Fain, Tsung Yu, Tianjing Li, Cynthia M. Boyd, Sonal Singh, Milo A. Puhan

Research output: Contribution to journalArticlepeer-review

6 Scopus citations


Objectives To describe challenges and make recommendations for researchers in how they select evidence to quantitatively assess a prescription drug's benefits and harms. Study Design and Setting These challenges and recommendations are based on our recent experience conducting a benefit-harm assessment for the prescription drug roflumilast. We considered the selection of evidence to quantify (1) the drug's treatment effects in patients, (2) the patient population's baseline risks for beneficial and harmful outcomes without treatment, and (3) the patient population's preferences for these beneficial effects and harms. These are fundamental steps for most benefit-harm assessment methods. Results We identify critical issues in selecting evidence for each of these steps. We justify in particular the need to incorporate (1) clinical trials for the drug's specific treatment effect; (2) observational studies with the most valid, precise, and applicable effect estimates for the baseline risk; and (3) flexible weighting approaches for balancing the drug benefits and harms. Conclusion We identify challenges and make recommendations for selecting evidence at the critical steps in a prescription drug's benefit-harm assessment. Our findings should assist other researchers conducting these assessments for prescription drugs, which could help regulators, medical professionals, and patients make better decisions about prescription drug use.

Original languageEnglish (US)
Pages (from-to)151-157
Number of pages7
JournalJournal of Clinical Epidemiology
StatePublished - Jun 1 2016


  • Benefit-harm assessment
  • Patient preferences
  • Pharmacoepidemiology
  • Prescription drug regulation
  • Prescription drugs
  • Roflumilast

ASJC Scopus subject areas

  • Epidemiology


Dive into the research topics of 'Evidence selection for a prescription drug's benefit-harm assessment: Challenges and recommendations'. Together they form a unique fingerprint.

Cite this