Measurement Property | Quality Criteria |
---|---|
Construct validity (hypothesis testing) | Clear description given of the construct measured by the comparator instrument Measurement properties of the comparator instrument described and adequate Design and statistical methods adequate for the hypothesis to be tested Otherwise free of any important flaws |
Test-retest reliability | Patients stable in the interim period Time interval appropriate Test conditions similar for the measurements Correct statistic used (ICC for continuous data, Îș for dichotomous/ordinal/nominal scores) Otherwise free of important flaws |
Responsiveness (longitudinal construct validity) | Criteria for change considered an adequate gold standard or the construct for change is clear, either as a situation of change or an actual indicator of change Measurement properties of the comparator standard described and adequate Statistical methods appropriate for the testing situations:
Otherwise free of important flaws |
Clinical trial discrimination | Time interval between testing stated and appropriate A proportion of people were expected to change in 1 or both groups A priori hypotheses stated regarding the anticipated mean differences in change scores between subgroups (positive, negative, or no change expected) Statistical methods adequate for the hypotheses tested (relative efficiencies, pooled treatment effect sizes, standardized mean differences) Otherwise free of important flaws |
Thresholds of meaning | Patient group similar to target population Criterion (external anchor, benchmarks, comparable population) selected in a credible manner Analysis done separately for improvement and deterioration or only in direction anticipated in the target application Multiple criteria used and results triangulated Analysis includes either a Youden index threshold from ROC or another cutoff on a ROC approach. If a threshold approach was used, was it tested for diagnostic utility (sensitivity and specificity)? Otherwise free of any flaws |
AUC: area under the curve; COSMIN-OMERACT: COnsensus-based Standards for the selection of health Measurement InstrumentsâOutcome Measures in Rheumatology; ICC: intraclass correlation coefficient; ROC: receiver-operating characteristic curve; SRM: standardized response mean.