Evidence in Practice

Evidence in practice makes up roughly 10% of the AKT and is one of the areas trainees most commonly neglect. The questions are often formula-based and follow predictable patterns, making them some of the easiest marks on the paper if you learn the core concepts.

Study design is fundamental. You need to recognise randomised controlled trials, cohort studies, case-control studies, cross-sectional surveys, and systematic reviews. Each has specific strengths and limitations. RCTs are the gold standard for intervention studies but cannot answer every clinical question. Case-control studies work backwards from outcome to exposure and are useful for rare diseases.

Statistical concepts tested include sensitivity (how well a test detects disease), specificity (how well it rules out disease), positive predictive value (PPV), negative predictive value (NPV), number needed to treat (NNT), number needed to harm (NNH), absolute risk reduction (ARR), relative risk, odds ratios, and confidence intervals.

NNT is calculated as 1 divided by the absolute risk reduction. AKT questions often give you event rates in two groups and ask you to calculate the NNT. For example, if a drug reduces events from 20% to 15%, the ARR is 5% and the NNT is 20. This means you need to treat 20 patients for one additional patient to benefit.

Sensitivity and specificity require a 2x2 contingency table. Sensitivity equals true positives divided by all positives (TP + FN). Specificity equals true negatives divided by all negatives (TN + FP). A highly sensitive test is good for ruling out disease (SnNOut), while a highly specific test is good for ruling in disease (SpPIn).

Critical appraisal involves evaluating whether a study's conclusions are valid, applicable, and clinically meaningful. CONSORT guidelines apply to trials, STROBE to observational studies, and PRISMA to systematic reviews.

NICE guideline development uses the GRADE framework to rate evidence quality from high to very low. Understanding how NICE interprets evidence helps you answer questions about guideline recommendations and their strength.