Systematic review of diagnostic test accuracy
How accurately does a test identify a target condition in its intended clinical setting?
When to use this project
Use this starter to synthesize accuracy evidence for a defined index test, target condition and reference standard. Specify whether the test is used for screening, triage, replacement or add-on testing.
How it differs
Diagnostic accuracy synthesis jointly addresses sensitivity and specificity, often across thresholds. It is not an intervention-effect review or a prediction-model development study, and accuracy alone does not show improved patient outcomes.
Decisions before you start
- Extract paired true/false positives and negatives with threshold, participant spectrum and reference-standard details.
- Explain handling of indeterminate tests, partial verification, repeated thresholds and missing cells.
- Choose a bivariate or HSROC approach when justified by the question and data; preserve threshold context and avoid independently pooling sensitivity and specificity as unrelated outcomes.
- Record the diagnostic appraisal tool and version. The QUADAS programme now includes QUADAS-3; do not silently label an older domain list as a completed assessment under a newer tool.
How the flows work together
Each flow below links to its diagram, steps and tables. Flow order organizes the project; it does not imply that all work is strictly sequential. Data assessment, analysis planning and protocol refinement may inform one another.
1. Diagnostic review overview
Overview · Question through test evidence interpretation
Use this overview to state the purpose and link to detailed decisions. Keep a versioned protocol and decision register, and add evidence references for the rationale.
2. Search and study selection
Selection · Identify eligible diagnostic studies
Keep records, reports and studies as separate counting units. The starter partition covers duplicate removal only: add explicit branches for automation or other removals if used. Supplementary searches may need separate identification routes. A final report-to-study mapping is not an exclusion count.
3. Accuracy extraction and synthesis
Statistics · Combine paired test accuracy measures
Record TP, FP, FN and TN separately in the extraction table. Use separate rows for thresholds and retain their dependence in the analysis. The sensitivity/specificity equation is a statistical specification, not an assumption node.
4. Bias, heterogeneity and applicability
Statistics · Assess trustworthiness across clinical settings
The starter names major concerns but does not reproduce a full appraisal instrument. Record the chosen QUADAS version and apply its full instructions. Keep risk of bias, applicability and body-of-evidence certainty distinct.
Equations and statistical interpretation
These explanations describe notation and assumptions. Flow does not calculate these values.
Paired accuracy measures
TP/FN are true-positive/false-negative counts among reference-positive participants; TN/FP are true-negative/false-positive counts among reference-negative participants.
These study-level proportions are not a meta-analysis model. Confidence intervals, paired structure, thresholds and between-study variability require additional methods.
Published example
A diagnostic accuracy review series. The linked page contains the current update, whose scope may differ from earlier versions; record the version you use in your own evidence node.
This publication illustrates a study design, not an endorsement of every method or a reproduction of its results. Follow its protocol, supplement and update history when studying the example.
Hypothetical scenario
A clinic wants to understand the accuracy of a rapid antigen test in asymptomatic people before deciding whether further clinical evaluation is needed.
- Diagnostic review overview: Define the intended use, target condition, population and acceptable reference standard.
- Search and study selection: Search and select diagnostic studies; retain threshold and setting information during selection.
- Accuracy extraction and synthesis: Extract paired 2×2 data, record indeterminate results and specify a suitable synthesis model.
- Bias, heterogeneity and applicability: Assess bias and applicability, investigate threshold and setting differences, and report uncertainty without claiming a clinical-outcome benefit.
Deliver a populated protocol, decision log, data specification, and planned outputs. Enter numerical findings only after analysis and verification. Keep decisions made before seeing results distinguishable from later amendments.
Sources
- Cochrane diagnostic test accuracy handbook
- QUADAS: diagnostic accuracy appraisal tools
- PRISMA-DTA: Diagnostic test accuracy
- PRISMA 2020 flow diagrams: identification and selection
Source check: 2026-09-17. Record the version of each reporting or appraisal tool used in your actual protocol.
Compare all Evidence synthesis project types · Open editor in a new tab