The screening accuracy of the PROMIS® Anxiety measures in adults - A systematic review and multiple-thresholds meta-analysis.

AIMS: To assess the test accuracy of PROMIS® Anxiety measures for screening any anxiety disorder (AAD) and generalised anxiety disorder (GAD), and how it varies by cut-offs.

METHODS

We searched the electronic databases from inception until 15 July 2025 (Embase, MEDLINE, PubMed-not-MEDLINE-subset, PsycINFO). We included test accuracy studies using any PROMIS® anxiety measure and (semi-)structured interviews in adults, and assessed the study quality with the QUADAS-2 instrument. We applied a multiple thresholds model to estimate the summary sensitivities and specificities across cut-offs and the area under the curve (AUC)

RESULTS

Six studies, each on a different disease, met the inclusion criteria, with a total of 1121 participants. For PROMIS® Short Form-(SF)-Anxiety(A)-8a, we found five studies, two for PROMIS®-CAT-A and one for PROMIS®-SF-A-7a. Meta-analyses for screening AAD and GAD were performed only for PROMIS®-SF-A-8a, identifying the maximised sum of sensitivity and specificity at T-score cut-offs 55.4 and 58.4, respectively. At these cut-offs, summary sensitivity and specificity were 0.76 (95% CI 0.71 to 0.80) and 0.74 (95% CI 0.70 to 0.78) for AAD, and 0.75 (95% CI 0.69 to 0.81) and 0.79 (95% CI 0.77 to 0.80) for GAD. Corresponding AUCs were 0.81 (95% CI 0.79 to 0.84) and 0.84 (95% CI 0.82 to 0.85). Only minor between-study heterogeneity was found.

CONCLUSIONS

The PROMIS®-SF-A-8a results were homogeneous, showing good screening accuracy. However, the limited number of studies - especially for other PROMIS® Anxiety measures - and the lack of replication in settings like primary care highlight the need for further research.