Evidence-Based Practice and Research Methodology
Domain 1 asks you to turn a clinical uncertainty into an answerable question, find the best available evidence, judge its trustworthiness, and apply it to one patient. Expect items on study design, bias, effect measures, and the difference between statistical and clinical significance.
The five-step cycle
Ask → acquire → appraise → apply → assess. The question is built with PICO: Population, Intervention, Comparator, Outcome (add T for timeframe or preferred design). A good PICO names the population precisely ("adults with type 2 diabetes, HbA1c ≥7%, not on insulin") because generalisability is judged against it.
Design and hierarchy
For therapy questions: systematic review/meta-analysis of RCTs > RCT > cohort > case-control > cross-sectional > case series/expert opinion. The hierarchy is question-specific - for rare harms, prognosis, or diagnostic accuracy, a well-conducted cohort or cross-sectional study may be the best available evidence, and an RCT may be unethical.
Effect measures
RR, OR, HR, ARR, and NNT (=1/ARR), each with a 95% CI. A p-value is the probability of data at least this extreme if the null is true; it is not the probability that the null is true, and p<0.05 says nothing about magnitude.
Appraisal and GRADE
Examine randomisation and allocation concealment, blinding, intention-to-treat analysis, attrition, and confounding. GRADE downgrades for risk of bias, inconsistency (e.g. I²>50%), indirectness, imprecision, and publication bias, yielding high/moderate/low/very low certainty.
Worked example
PICO: in adults with type 2 diabetes, does a carbohydrate-restricted diet (<130 g/day) versus a balanced-carbohydrate diet lower HbA1c at 6 months? A PubMed/Cochrane search (MeSH "Diabetes Mellitus, Type 2" AND "Diet, Carbohydrate-Restricted") retrieves a meta-analysis of 12 RCTs (n=1,200): mean HbA1c difference −0.3% (95% CI −0.5 to −0.1), I²=60%. Reasoning: the result is statistically significant and the CI excludes no effect, but 0.3% is below the ~0.5% usually taken as clinically meaningful and smaller than first-line pharmacotherapy; heterogeneity downgrades certainty to moderate. Apply with patient preference and adherence, then reassess.
Traps
Equating p<0.05 with clinical importance; swapping OR for RR; discarding a systematic review for one memorable trial; ignoring attrition; reading "no evidence of effect" as "evidence of no effect"; skipping the patient-values step.