FROM MY PT SCHOOL NOTES · EVIDENCE-BASED PHYSICAL THERAPY · CHAPTER 10
How Do You Decide if an Intervention Study Is Worth Using?
The useful question is not simply whether a treatment had a significant result, but whether its benefit is credible and meaningful.
Look at the comparison
The intervention lecture asks whether a treatment is worth offering: is the observed change credible, meaningful, practical, and acceptable to the patient? It compares randomized, quasi-experimental and observational evidence while noting the trade-offs between control and everyday clinical relevance.
Study appraisal topics from class include random allocation, allocation concealment, baseline group comparability, masking of participants/clinicians/assessors when possible, adherence, other treatments, dropout, consistent outcome measurement and intention-to-treat analysis. Allocation concealment before enrollment is distinct from blinding after assignment.
Examine the effect and uncertainty
The course distinguishes continuous outcomes (for example, mean functional scores) from dichotomous outcomes (for example, whether an injury occurred). It introduces raw and standardized effect sizes, confidence intervals, absolute/relative risk reduction or benefit increase, and number needed to treat (NNT), which depends on baseline risk and follow-up period.
Consider the patient’s priorities
Minimal clinically important difference (MCID) is discussed as a change considered worthwhile by patients or other stakeholders. Measurement error and meaningfulness are related but different questions. A statistically significant difference may still be too small or inconvenient to influence care.
Check your reasoning: Can a statistically significant difference be too small to matter to the patient?
Yes. Compare magnitude, uncertainty, burden and potential harms with a clinically meaningful change and the person’s goals.
Updated-standard note
Fixed Cohen’s-d labels and claims that a treatment effect is ‘big enough to see with the naked eye’ are historical slide shortcuts. Interpret effect sizes in clinical context, with confidence intervals, harms, feasibility, patient values and outcome importance. An MCID is not simply the same as standard error of measurement.
Put it into practice
Scenario: For a hypothetical adverse event over one year, 20 of 100 controls experience the event and 10 of 100 treated participants experience it.
What should you conclude?
Reasoning: The absolute risk reduction is 0.20 − 0.10 = 0.10 (10 percentage points), corresponding to an NNT of 10 over one year, assuming the comparison is credible. Consider harms, burden and applicability too.
Common trap: Quoting a relative percentage without its baseline risk or follow-up period. Remember: Absolute effects and timeframe make treatment claims interpretable.
Current reference (separate from my class notes): CONSORT 2025 reporting guideline. Use current transparent-trial reporting guidance when critically appraising randomized studies. Reporting completeness and study validity are related but not identical.
Course basis: my Evidence-Based PT school notes, source lecture Chapter 13. New questions and headings are teaching aids, not official exam items. For full definitions and examples, use the class notes and assigned readings.
