An examination of the quality and performance of the Alda scale for classifying lithium response phenotypes.

摘要: Objectives The Retrospective Assessment of the Lithium Response Phenotype Scale (Alda scale) is most widely used clinical measure lithium response phenotypes. We assess its performance against recommended psychometric and clinimetric standards. Methods data from Consortium for Genetics a French study phenotypes (combined sample >2500) to reproducibility, responsiveness, validity, interpretability A scale (assessing change in illness activity), B scale, items confounders response) previously established categories derived Total Score Alda scale. Results key findings are that vulnerable error measurement. For example, some contribute little overall (eg, B2) does not reliably single construct (uncertainty response). Machine learning models indicate it may be more useful employ an algorithm combining ratings individual sequence clarifies noise signal ratio instead using composite score. Conclusions This highlights three important topics. First, empirical approaches can help determine which aspects any improved. Second, best applied as multidimensional index (identifying several independent assessment Third, integrated science approach precision psychiatry vital, otherwise phenotypic misclassifications will undermine reliability validity genetics biomarker studies.