Index and reference testing

Specify thresholds blinding and timing

The index test is the test under evaluation; the reference standard determines target-condition status. Neither is automatically error-free. Specify thresholds before examining accuracy and distinguish masking test readers from concealing randomized allocation.

Flow diagram

Index and reference testingF1.S1Specify index testF1.S2Specify reference standardF1.S3Plan blinding and test orderF1.S4Define test timing and handlingF1.S5Handle indeterminate and missingresultsF1.N1Test specificationTestThresholdProcedureReaderBlindingTimingF1.N2Incomplete result rulesStatusDefinitionActionAnalysishandlingRationaleF1.N3Reference standard evidenceF1.S0 Start flowIndex and reference testingF1.S999 End flowEnd flow

Steps

Specify index test

Specify the index test, specimen and interpretation threshold before assessing accuracy. Label any threshold selected using the study data as exploratory.

Specify reference standard

Define the reference standard for the target condition, its limitations and any composite or adjudication procedure. Prespecify whether readers know the index-test result.

Plan blinding and test order

Control knowledge of other tests and clinical information.

Define test timing and handling

Specify intervals, intervening treatment and specimen procedures.

Handle indeterminate and missing results

Prespecify repeat testing, adjudication and verification rules.

Index and reference testing

End flow

Tables and supporting information

Test specification

Test specificationTestThresholdProcedureReaderBlindingTiming

Avoid data-driven thresholds presented as prespecified.

Incomplete result rules

Incomplete result rulesStatusDefinitionActionAnalysishandlingRationale

Keep nonbinary outcomes outside the 2×2 table until a justified rule applies.

Reference standard evidence

Document the reference standard and known limitations.

Sources