Diagnostic impact (test-treatment) studies
A diagnostic impact study asks whether using a test to guide care improves patient outcomes compared with another care strategy.
Why this level?
Specify the research question: intervention, diagnostic accuracy, prognosis, aetiology, or screening intervention. Other question types are outside this hierarchy. NHMRC 2009 sourceDiagnostic impact asks about an intervention strategy, not accuracy alone. Evidence depends on randomization, allocation unit and the actual comparison; the broad label cannot assign one level.
Fit
When to use it, and when not to
Use it when
- The test has plausible accuracy, but its effect on treatment, patient benefit or harm remains uncertain.
- You can define what care follows each test result and compare complete diagnostic strategies.
Not when
- You first need accuracy against a reference standard. Use a diagnostic accuracy study. Read: Diagnostic accuracy study
- You are evaluating testing in people without symptoms as a population program. Consider a screening evaluation. Read: Screening biases
Before you start
What you need
- Both test-and-care pathwaysSpecify test thresholds, result timing and action after positive, negative or uncertain results in each strategy.
- A patient-relevant outcome and time horizonTreatment changes can explain a pathway but are not automatically patient benefit. Include downstream harms and resource use.
- Who receives which strategyRandomize patients where feasible. Randomize clinics when shared practice makes individual allocation contaminate the comparison.
- How staff will receive and act on resultsRecord whether the test was performed, returned in time, understood and acted on. Plan missing-outcome handling.
- The comparison separates complete care strategies.How to check: Track access to the other test and shared staff practice. Contamination can reduce the contrast between groups.
- Inference respects the unit of allocation.How to check: Clinic allocation with individual outcomes needs clustered analysis or an explicit justification. Include small-sample methods when clinics are few.
- Missing outcomes need an explicit assumption.How to check: Missing completely at random (MCAR) means gaps are unrelated to recorded and unrecorded values; this is rarely plausible. Missing at random (MAR) means gaps depend only on recorded factors. Under MAR, consider multiple imputation; it is not automatically better than a justified complete-case analysis. Missing not at random (MNAR) is plausible if unobserved values themselves influence whether records are absent. Under MNAR, test departures with a delta or tipping-point analysis that shifts the unobserved values.
- Outcome measurement is comparable between strategies.How to check: Mask assessors where feasible and use the same follow-up. Extra testing can detect events without changing their true frequency.
Plain language
How it works
Compare the whole pathway from test to care to outcome. A test can change classification without changing treatment, or change treatment without helping patients. Randomization compares strategies while balancing baseline causes on average. Individual and cluster trials need analyses and sizing that match their allocation. An observational strategy comparison instead needs confounding control and an assumption-first missing-data plan.
Measure test uptake, result timing, clinician actions and treatment adherence alongside clinical outcomes. These measures help explain why a strategy worked or failed. They do not replace the patient outcome. Prespecify the main comparison, follow-up and safety measures; include harms from false positives and false negatives.
Worked example
A worked example
- Setting
- An individually randomized trial compares a rapid test-and-treatment strategy with usual care.
- Question
- Does the strategy lower admission risk within the chosen follow-up window?
- Approach
- Analyze participants by assigned strategy. Measure patient outcomes and harms, as well as whether clinicians followed the test advice.
- Result
- Admissions occur in 30 of 300 patients (10 in 100) versus 60 of 300 (20 in 100). The difference is -10 per 100; the risk ratio is 0.5.
- What it means
- The contrast concerns the assigned strategy, including how it was used. It does not isolate the test chemistry from subsequent care.
In practice
Recent studies PubMed search · not vetted
Watch out
Common mistakes
The one to remember
A more accurate test need not improve patient outcomes. Benefit depends on result delivery, treatment decisions, adherence and harms. Sensitivity and specificity do not size or establish the benefit of a test-treatment strategy.
- Using higher sensitivity as proof of clinical benefit.Instead: Measure downstream patient outcomes. False positives, overtreatment, delay and missed diagnoses can change net benefit.
- Analyzing only patients whose clinicians followed the test advice.Instead: Keep the assigned-strategy analysis primary. Adherence-based comparisons can introduce selection and answer a different question.
- Sizing patient benefit with a diagnostic accuracy calculation.Instead: Size the patient outcome contrast, allowing for allocation, clustering, contamination and missing outcomes.
Inside Mallard
How Mallard uses this
Plans do not link to this design yet. Mallard's diagnostic accuracy form sizes sensitivity and specificity, not patient benefit from a test-treatment strategy.
design.diagnostic_impact, the id a plan cue records.How to report it
Use CONSORT for a randomized test-treatment trial, with the relevant allocation-design extension where needed. Describe both diagnostic and treatment pathways, result delivery, thresholds, clinician actions, adherence and contamination. Report the patient outcome contrast by assigned strategy, with uncertainty and downstream harms. Accuracy estimates alone do not report clinical impact. An observational comparison needs reporting guidance matched to that design instead. A reporting guideline is not a design manual: use it to make the study transparent, not as a substitute for careful design. Check the target journal's own requirements as well, which this page cannot know.
Related Learn pages
Related designs and methods
Evidence
Sources & further reading
- Procalcitonin-Guided Use of Antibiotics for Lower Respiratory Tract Infection (opens in a new tab) · N Engl J Med · doi 10.1056/NEJMoa1802670
- High-sensitivity troponin in the evaluation of patients with suspected acute coronary syndrome: a stepped-wedge, cluster-randomised controlled trial (opens in a new tab) · Lancet · doi 10.1016/S0140-6736(18)31923-8
- CONSORT 2025 statement: updated guideline for reporting randomised trials (opens in a new tab) · BMJ · doi 10.1136/bmj-2024-081123
- Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls (opens in a new tab) · BMJ · doi 10.1136/bmj.b2393