Research, Evidence-Based Practice, and Quality Improvement
Core concepts
Evidence-based practice integrates best available evidence, clinical expertise, and patient values. Know study designs (RCT, cohort, case-control, cross-sectional, systematic review/meta-analysis), measures of association (RR, OR, HR), and measures of effect (ARR, NNT). Understand bias, confounding, validity, reliability, sensitivity, specificity, and PPV/NPV. Be able to appraise a paper: was the question focused, was the design appropriate, were confounders controlled, were results precise and generalizable? Quality improvement uses PDSA cycles, root-cause analysis, and indicators (structure, process, outcome). Understand how to translate guidelines into practice and how to measure whether change improved care.
Traps
Questions test whether you know that correlation is not causation, that a statistically significant result may not be clinically meaningful, and that PPV depends on prevalence. Watch for confusion between sensitivity and specificity, and between absolute and relative risk.
Worked example
A trial reduces infection from 20% to 15%: ARR = 5% (0.05); NNT = 1/0.05 = 20. Relative risk reduction = 25%. Rule of thumb: report absolute effects and NNT for clinical decisions.
Rule of thumb
Ask: compared with what, in whom, how measured, and how precise? Then judge applicability to your patient.