Practical guides · Page 19

Diagnostic Accuracy: Choose the Right Denominator

Build and audit a diagnostic two-by-two table before interpreting sensitivity, specificity, or predictive values.

Turn protocol criteria into reproducible screening decisions with a small, documented pilot.

Define a testing family and error-control objective before adjusting a list of p-values.

Define denominators, count missing values, and record reasons before deciding how to handle incomplete research data.

Check event definitions, time at risk, and units before expressing events per person-year or person-month.

Turn a sample-size target into a realistic recruitment discussion, with explicit assumptions for enrollment, loss, and follow-up.

Match your report to a guideline, check extensions, and turn the checklist into a manuscript-location map.

A practical handoff checklist for data definitions, analysis scripts, software versions, and manuscript outputs.

Distinguish reported, derived, and estimated standard deviations and document the assumptions behind each.

Check event definitions, denominators, group order, and sparse cells before calculating an unadjusted risk ratio.

Check scale direction, outcome timing, sample sizes, and standard deviations before calculating Hedges g.

Record database platforms, exact strategies, dates, exports, and decisions throughout a systematic review search.