From Clinical Question to Test

#Define the uncertainty before the measurement

A test question starts with what is uncertain and why knowing more might matter. Is the purpose screening, investigating a suspected condition or following an established condition or treatment over time? MedlinePlus describes these different uses of laboratory tests and explains that results are considered with symptoms, examination findings and health history. A measurement is one piece of that picture, not a complete account of health.

For research, STARD asks authors to describe the intended use and clinical role of the test. A question that names the people, setting, target condition and proposed use is easier to evaluate than a broad claim that a device detects disease. This page explains that framing; it does not decide whether an individual needs testing.

Evidence: MedlinePlus: understanding lab results / STARD 2015: explanation and elaboration

#Place the test in the pathway

The role of a test includes its position relative to information already available. STARD distinguishes a triage test used before an existing test, an add-on used after earlier testing and a replacement for an established test. Evidence for one position cannot automatically be assumed to support another.

Ask what prior assessment participants had, what the test is meant to add and what information was available to the people interpreting it. The same technology may be studied in different roles. A headline sensitivity or specificity does not by itself explain whether it contributes useful information to the whole strategy, or whether the population in that study resembles the proposed setting.

Evidence: STARD 2015: explanation and elaboration / FDA: reporting diagnostic test performance

#Read the comparison and the people together

An accuracy study needs a defined target condition and an appropriate reference standard for establishing its presence or absence. FDA distinguishes this from agreement with another test that is not a reference standard. Two tests agreeing is not proof that either gives the correct answer. Read how the comparison was made before accepting a performance claim.

STARD asks for eligibility, recruitment, setting, participant flow and handling of missing or indeterminate results. A study drawn from obvious cases and healthy controls can differ from a group with uncertain symptoms in ordinary care. These details help readers assess applicability and possible bias. They do not turn a published study into a personal diagnosis.

Evidence: FDA: reporting diagnostic test performance / STARD 2015: explanation and elaboration

#Connect information to a meaningful outcome

Detecting an abnormality is different from demonstrating that a testing strategy helps people. For population screening, WHO and the UK National Screening Committee consider the wider programme, including further assessment, intervention, possible harms and the capacity to provide follow-up. Those criteria concern programme evaluation, not a universal rule for choosing every diagnostic test.

A useful evidence question asks what an answer could change and what burdens follow from it. Avoid claiming benefit merely because a test finds more conditions or produces more detail. This is general health-science education, not a testing algorithm, an eligibility decision, a recommendation to start or stop testing, or a claim that Mynd offers diagnostic services.

Evidence: WHO: screening programme benefits and harms / UK National Screening Committee: programme appraisal criteria / MedlinePlus: understanding lab results

Source note

The sections above were checked against the linked sources. No clinical review has been performed. This is general research education, not a clinical guideline.