[Overcoming the limitations of the descriptive and categorical approaches in psychiatric diagnosis: a proposal based on bayesian networks].

[Overcoming the limitations of the descriptive and categorical approaches in psychiatric diagnosis: a proposal based on bayesian networks]. Turk Psikiyatri Derg. 2015;26(1):1-12 Authors: Sorias S Abstract Efforts to overcome the problems of descriptive and categorical approaches have not yielded results. In the present article, psychiatric diagnosis using Bayesian networks is proposed. Instead of a yes/no decision, Bayesian networks give the probability of diagnostic category inclusion, thereby yielding both a graded, i.e., dimensional diagnosis, and a value of the certainty of the diagnosis. With the use of Bayesian networks in the diagnosis of mental disorders, information about etiology, associated features, treatment outcome, and laboratory results may be used in addition to clinical signs and symptoms, with each of these factors contributing proportionally to their own specificity and sensitivity. Furthermore, a diagnosis (albeit one with a lower probability) can be made even with incomplete, uncertain, or partially erroneous information, and patients whose symptoms are below the diagnostic threshold can be evaluated. Lastly, there is no need of NOS or "unspecified" categories, and comorbid disorders become different dimensions of the diagnostic evaluation. Bayesian diagnoses allow the preservation of current categories and assessment methods, and may be used concurrently with criteria-based diagnoses. Users need not put in extr...
Source: Turkish Journal of Psychiatry - Category: Psychiatry Authors: Tags: Turk Psikiyatri Derg Source Type: research