Pragmatic Diagnostic Accuracy Study
Also known as: real-world diagnostic accuracy study, pragmatic DAS, routine-care diagnostic study, pragmatic test evaluation
A pragmatic diagnostic accuracy study evaluates how well a diagnostic test performs under real-world clinical conditions — not in idealized, tightly controlled settings. Conducted within routine care workflows, it measures sensitivity, specificity, predictive values, and likelihood ratios for an index test against a reference standard, yielding accuracy estimates directly applicable to clinical practice rather than laboratory benchmarks.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use a pragmatic diagnostic accuracy study when you need accuracy estimates that reflect real clinical practice — after a test has already shown promise in explanatory studies and the question shifts to how it performs in the field. It is appropriate when patients are enrolled consecutively from routine care, when the operator mix and workflow conditions are representative, and when the reference standard is feasible in that setting. Do not use it as a first step for a completely novel, unvalidated test where safety or operator-training requirements have not yet been established; an explanatory design with strict control is more appropriate there. Also avoid it when the reference standard cannot be applied consistently across all enrolled patients, as this creates verification bias.
Strengths & limitations
- Accuracy estimates have high external validity and are directly applicable to everyday clinical decisions.
- Leverages existing clinical workflows, reducing cost and recruitment barriers.
- Captures real heterogeneity in patients, operators, and equipment that explanatory studies exclude.
- STARD guidelines provide a mature, widely accepted reporting framework that facilitates peer review and systematic review inclusion.
- Findings are easier to translate into clinical guidelines and health technology assessments than explanatory results alone.
- Less internal control than explanatory studies — operator variability and protocol deviations can inflate or deflate true accuracy.
- Prevalence of the target condition in routine care directly affects predictive values; estimates may not transfer to settings with very different prevalence.
- Applying a rigorous reference standard to all patients in routine care is often logistically difficult, creating partial verification bias.
- Observational design means that confounding by clinical indication can affect which patients receive the index test.
Frequently asked
What makes a diagnostic accuracy study 'pragmatic' rather than 'explanatory'?
A pragmatic design enrolls a broad, representative patient population from routine clinical care, uses the test exactly as it would be used in practice (including real-world operator variability and equipment), and accepts naturally occurring heterogeneity. An explanatory design uses strict inclusion/exclusion criteria, standardized operators, and controlled conditions to maximize internal validity. Pragmatic studies yield estimates closer to real-world performance; explanatory studies yield estimates closer to ideal performance.
How do I avoid verification bias in a pragmatic setting?
Verification bias occurs when only index-test-positive patients receive the reference standard. To avoid it, apply the reference standard to all enrolled patients — or at minimum to a random subsample of negatives. If full verification is infeasible, use statistical correction methods (e.g., the Begg-Greenes correction) and report the limitation transparently.
What sample size do I need?
Sample size depends on the target sensitivity and specificity, the expected prevalence of the target condition, and the desired confidence interval width. A common rule of thumb is at least 30–50 cases (condition-positive patients) and 30–50 controls, but formal power calculations using methods such as those by Buderer or the PASS software module for diagnostic studies are strongly recommended.
Should I report sensitivity/specificity or predictive values?
Report both. Sensitivity and specificity are intrinsic properties of the test (relatively prevalence-independent) and are useful for comparing tests and for inclusion in meta-analyses. Predictive values are prevalence-dependent and tell clinicians how to interpret a positive or negative result in their specific setting. Likelihood ratios combine both and facilitate Bayesian updating of pre-test probability.
Do I need ethical approval for a pragmatic diagnostic accuracy study?
Usually yes, because the study involves collection and analysis of patient data, and often the pairing of a new or candidate test with a reference standard that may not yet be standard care. Check local regulations; some retrospective designs using anonymized existing records may qualify for expedited review, but prospective enrollment almost always requires full ethics committee approval and informed consent.
Sources
- Bossuyt, P. M., et al. (2015). STARD 2015: An Updated List of Essential Items for Reporting Diagnostic Accuracy Studies. BMJ, 351, h5527. DOI: 10.1136/bmj.h5527 ↗
- Schilling, I., & Burchardt, M. (2018). Pragmatic diagnostic accuracy studies: bridging the gap between explanatory trials and routine practice. Diagnostic and Prognostic Research, 2(1), 14. link ↗
How to cite this page
ScholarGate. (2026, June 3). Pragmatic Diagnostic Accuracy Study. ScholarGate. https://scholargate.app/en/epidemiology/pragmatic-diagnostic-accuracy-study
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Cohort StudyEpidemiology↔ compare
- Cross-sectional epidemiological studyEpidemiology↔ compare
- Diagnostic Accuracy Study DesignClinical Research↔ compare
- Pragmatic randomized clinical trialEpidemiology↔ compare
- Prospective Diagnostic Accuracy StudyEpidemiology↔ compare
- Screening Test EvaluationEpidemiology↔ compare