Machine learning-decision tree classifiers in psychiatric assessment - An application to the diagnosis of major depressive disorder

Aug 5, 2026

This work demonstrates the benefits of machine learning classifiers in psychiatric assessment by introducing machine learning decision trees (ML-DTs) for scoring and interpreting psychodiagnostic test results. The method is applied to real PHQ-9 data from cross-sectional nonclinical and clinical Japanese samples, using ROC analysis, DSM algorithm comparison, and ML-DT modeling. ML-DTs achieved higher diagnostic accuracy (0.71–0.75) than the DSM algorithm (0.69) and ROC curves (0.70–0.71), classifying depression using information from an average of 2.99 of 9 items, supporting efficient, clinically valuable integration with traditional methods.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

Psychological assessment and diagnosis with classical methods and ML classifiers

Psychiatry Research 322 (2023) 115127 Contents lists available at ScienceDirect Psychiatry Research journal [homepage:](homepage: www.elsevier.com/locate/psychres)[ www.elsevier.com/locate/psychres](homepage: www.elsevier.com/locate/psychres) | Machine learning-decision tree classifiers in psychiatric assessment: An application to the diagnosis of major depressive disorder | | | | | --- | --- | --- | --- | | *
Daiana Colledani , Pasquale Anselmi, Egidio Robusto
Department of Philosophy, Sociology, Education and Applied Psychology, University of Padova, Via Venezia 14, 35131, Padova, Italy | | | | | A R T I C L E I N F O | | A B S T R A C T | | | Keywords:
Machine learning Psychodiagnostic test PHQ-9
Sensibility
Specificity | | This work illustrates the advantages of using machine learning classifiers in psychiatric assessment. Machine learning-decision trees (ML-DTs) represent a new approach to scoring and interpreting psychodiagnostic test data that allows for increasing assessment accuracy and efficiency. The approach is outlined in an easy yet detailed way, and its application is illustrated on real psychodiagnostic test data. Specifically, cross-sectional data concerning nonclinical and clinical Japanese populations were taken from a panel registered with an internet survey company. Responses to the Patient Health Questionnaire-9 (PHQ-9) underwent receiver operating characteristic (ROC) curve, DSM algorithm, and ML-DT analyses. The results showed greater diagnostic accuracy for ML-DT (0.71–0.75) compared with the DSM algorithm (0.69) and ROC curves (0.70–0.71). Moreover, ML-DT enabled classifying participants as having or not having a diagnosis of depression using, on average, the information from 2.99 out of 9 items (SD = 1.35). The application showed that ML-DTs can provide information of high clinical value to integrate traditional psychometric methods. The resulting assessments are informative, accurate, and efficient. | | 1. Introduction Formulating timely and accurate diagnoses is crucial to maximizing the efficacy of therapeutic actions and reducing costs. The diagnostic process often requires the administration of multiple questionnaires and the execution of complex, expensive, and sometimes invasive clinical examinations. Understanding which symptoms are the most informative for each patient might not be easy. Medicine and psychology have recently begun to take advantage of machine learning (ML; Battineniet al., 2020; Witten et al., 2016), a subfield of artificial intelligence that aims to learn new pieces of knowledge from a set ofdata to use them for effectively predicting new cases (Witten et al., 2016). This work aims to illustrate how ML can contribute to the diagnostic process in psychiatry. Specifically, the paper outlines the usefulness of ML classifiers as a new approach to interpreting the data from psychodiagnostic tests. ML can improve the efficiency of the diagnostic process while guaranteeing its accuracy. Moreover, it can provide information of high clinical value that supplements that of traditional methods. In the next section, ML is described, and its strengths and weaknesses are compared with those of traditional methods. Then, the results of an application on real psychodiagnostic test data are presented and discussed. 1.1. Psychological assessment and diagnosis with classical methods and ML classifiers Usually, the testing process starts with the administration of a set of items to an individual and ends with the computation of an aggregate score that informs the clinician about the level of the assessed characteristic in that individual. The diagnosis of a certain disease is formulated if the score is above a specific value. Clinical cut-off scores are usually identified using receiver operating characteristic (ROC) curves (Carter et al., 2016). Given a diagnosis obtained with a gold standard (e. g., clinical interview), the ROC curve method identif

FAQs

What is the main goal of using ML decision trees in psychiatric assessment?

To improve scoring and interpretation of psychodiagnostic test data, increasing assessment accuracy and efficiency while supporting clinical decision-making.

How was the approach applied in this paper?

PHQ-9 responses from nonclinical and clinical Japanese populations were analyzed using ROC, a DSM algorithm, and ML-DT analyses.

How did ML-DTs perform compared with traditional methods?

ML-DTs showed greater diagnostic accuracy (0.71–0.75) than the DSM algorithm (0.69) and ROC curves (0.70–0.71), while using on average 2.99 of 9 PHQ-9 items for classification.

Aug 5, 2026

Upload by Olivia Brown