Listen to the data when results are not significant

BMJ. 2008 Jan 5;336(7634):23-5. doi: 10.1136/bmj.39379.359560.AD.

Abstract

Unexpected non-significant results from randomised trials can be difficult to accept. Catherine Hewitt, Natasha Mitchell, and David Torgerson find that some authors continue to support interventions despite evidence that they might be harmful

MeSH terms

  • Abortifacient Agents, Nonsteroidal / administration & dosage
  • Administration, Inhalation
  • Administration, Oral
  • Adolescent
  • Anesthetics, Local / administration & dosage
  • Bias
  • Data Interpretation, Statistical
  • Equipment Safety
  • Evidence-Based Medicine
  • Female
  • Humans
  • Infant
  • Labor, Induced / methods
  • Lidocaine / administration & dosage
  • Misoprostol / administration & dosage
  • Pain / prevention & control
  • Pregnancy
  • Randomized Controlled Trials as Topic / statistics & numerical data*
  • Sex Education / statistics & numerical data
  • Wounds and Injuries / prevention & control

Substances

  • Abortifacient Agents, Nonsteroidal
  • Anesthetics, Local
  • Misoprostol
  • Lidocaine