Evaluation of Computerized Adaptive Tests (CATs) for longitudinal monitoring of depression, anxiety, and stress reactions

Abstract Background Computerized adaptive testing (CAT) based on Item Response Theory, (IRT) offers an efficient way for accurate measurement of patient reported outcomes. The efficiency lies within a minimal response burden and a high measurement precision over a broad measurement range. The objective of the study was to evaluate and compare the responsiveness of CATs measuring anxiety, depression, and stress reaction to standard static self-assessment tools. Methods Longitudinal data of n =595 psychosomatic inpatients were analyzed for evaluating retest-reliability and sensitivity to change of the CATs compared to static measures (GAD-7, PHQ-9, and PSQ) using correlational and ANOVA statistics. The study hypothesized that CATs are at least as retest-reliable and as sensitive to change as static tools. Results The three CATs show a low burden for patients, administering on average 5–7 (±2–6SD) items with similar retest-reliability compared to the static tools applied (A-CAT: r =.78 vs. GAD-7: r =.75, D-CAT: r =.71 vs. PHQ-9: r =.75, S-CAT: r =.80 vs. PSQworries scale: r =.80). The CATs were overall as sensitive to change as the static tools (Cohen׳s d ranged between .19 and .69). Limitations This is a monocenter, observational, longitudinal study without external clinical criteria; thus generalization to other settings may be limited. Conclusions The tested CATs belong to the first generation of CATs being used in daily routine for more than a decade. They are as retest reliable and sensitive to change as static tools. Newer CATs may provide further practical advantages.