Research and Evidence-Based Practice
What to know
Evidence-based practice integrates best available research, clinical expertise, and client values and preferences. Know study designs and their trade-offs: randomized controlled trials (causal inference, internal validity), quasi-experimental, cohort, case-control, and case studies. Understand bias types (selection, attrition, measurement, publication), confounding, randomization, blinding, and effect size versus statistical significance. Be able to read basic statistics: p-values, confidence intervals, number needed to treat, and power. Understand levels of evidence, systematic reviews and meta-analyses, and the limits of generalizing findings to clients unlike study samples. Know fidelity, implementation, and how to appraise whether a treatment is evidence-based for a specific presentation.
Traps in questions
- Equating statistical significance with clinical importance; effect size and CI matter.
- Treating a single study or anecdote as definitive.
- Ignoring sample composition and applicability to your client.
- Confusing correlation with causation.
- Assuming a manualized treatment works identically across cultures and settings.
Rule of thumb
Ask three questions: Is the evidence valid? Is the effect meaningful? Does it apply to this client? A wide confidence interval crossing zero means uncertainty about direction.
Worked example: A trial reports a small effect with a 95% CI from 0.02 to 0.30. It is statistically significant but the lower bound is near zero, so clinical importance is modest and should be discussed with the client.