The THINC-it digital cognitive screening tool in depressive disorders: a systematic review of psychometric properties and clinical applications.

OBJECTIVE

To evaluate the psychometric properties and clinical utility of the THINC-it digital cognitive screening battery in depressive and mood disorders.

METHODS

Systematic review conducted in accordance with PRISMA 2020 and organized within a COSMIN-informed framework. PsycInfo, PubMed, Scopus, Web of Science, and Europe PMC were searched from inception to September 1, 2025, without language restrictions. Two independent reviewers screened records and extracted data on validity, reliability, measurement error, responsiveness, diagnostic test accuracy, and feasibility.

RESULTS

Of 284 records identified, 17 studies were included. Construct validity was the most consistently supported measurement property: convergence was most consistent between Codebreaker and the Digit Symbol Substitution Test and between Trails and Part B of the Trail Making Test, and more variable for other pairings. Cutoffs with sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were rarely reported. Test-retest reliability was reported in three studies, only one of which assessed patients rather than healthy controls; the standard error of measurement (SEM), minimal detectable change (MDC), and minimal clinically important difference (MCID) were never reported. Heterogeneity across populations, language versions, devices, and administration settings limited comparability and interpretation of longitudinal change.

CONCLUSIONS

THINC-it is a brief digitalized cognitive screening battery (≈10-15 min) with potential utility for continuous, non-cutoff assessment. The current evidence provides the strongest support for its construct validity. Further studies are required to rigorously quantify measurement error, define the minimal clinically important difference (MCID), establish diagnostic accuracy, and evaluate equivalence across devices, settings, and language versions.