A study finding no effect isn't proof the effect doesn't exist.
This distinguishes a null result caused by a true lack of effect from one caused by insufficient power, poor measurement, or a short follow-up period. Clinicians should check a study's power, sample size, and confidence intervals before concluding an intervention 'doesn't work' — a wide confidence interval spanning both harm and benefit is absence of evidence, not evidence of absence. Genuine evidence of absence requires an adequately powered study with a tight confidence interval excluding clinically meaningful effects.
This guide was auto-drafted and is pending editorial review.