FORTH models Sample Clauses

FORTH models. The expected performance of the developed prediction models was evaluated on observed input – output data using statistical measures estimating the accuracy of forecast methods (e.g. root mean squared error, sensitivity, specificity) and approaches (e.g. sensitivity analysis).

FORTH models. ‌ 4.1.1 Implementation of prediction models within the Decision Support Tool (FORTH)‌ Following extensive modeling work on the total sample (D4.3b, first part) and the across-site validation experiments (Section 3.1, present document) the functionalities of the Decision Support Tool (DST) were revised accordingly. The final DST was designed on the following principles: (i) flexibility toward future use in clinical settings, (ii) performance accuracy in predicting key aspects of patient well-being, (iii) robustness in formulating personalized risk profiles of potentially modifiable patient characteristics, and (iv) directly linking personalized needs assessment with concrete suggestions regarding psychological prevention strategies. Regarding the first principle, the DST incorporates models addressing eight distinct clinical outcomes: (a) Prediction of overall mental health status (“good” vs“poor”) at M12 post-diagnosis (b) Prediction of overall mental health status (“good” vs “poor”) at M18 post-diagnosis (c) Prediction of Global QoL status (“good” vs “poor”) at M12 post-diagnosis (d) Prediction of Global QoL status (“good” vs “poor”) at M12 post-diagnosis (e) Prediction of mental health decline (“good” at M0 or M6 🡪 “poor” at M12 vs “good” at M0 or M6 🡪 “good” at M12 (f) Prediction of mental health decline (“good” at M0 or M6 🡪 “poor” at M18 vs “good” at M0 or M6 🡪 “good” at M18 (g) Prediction of QoL decline (“good” at M0 or M6 🡪 “poor” at M12 vs “good” at M0 or M6 🡪 “good” at M12 (h) Prediction of QoL decline (“good” at M0 or M6 🡪 “poor” at M18 vs “good” at M0 or M6 🡪 “good” at M18 Models (a-d) address the need to identify patients at risk of either overall poor mental health (or global QoL, respectively) at a particular end-time. These models were deemed more appropriate in terms of classification performance for patients who report poor mental health (or global QoL, respectively) at the time of diagnosis. Models (e-h) address the need to identify patients at risk of declining mental health (or global QoL, respectively) having displayed adequate classification performance on the subgroup of patients who reported good mental health (or global QoL, respectively) at the time of diagnosis. Added clinical flexibility is afforded by the tool which permits prediction of key end points (overall “poor” vs “good” or “stable good” vs “decline”) based on available psychological and life-style measurements (1) at the time of diagnosis and 3 months later, (2)...