A multilevel machine learning algorithm to predict session-by-session outcome for patients receiving cognitive-behavioural therapy

IF 4.5 2区心理学 Q1 PSYCHOLOGY, CLINICAL

Behaviour Research and Therapy Pub Date : 2025-09-03 DOI:10.1016/j.brat.2025.104848

Juan Martín Gómez Penedo , Alice E. Coyne , Manuel Meglio , Marjolein Fokkema , Rebekka Wassmann , Wolfgang Lutz , Julian Rubel

{"title":"A multilevel machine learning algorithm to predict session-by-session outcome for patients receiving cognitive-behavioural therapy","authors":"Juan Martín Gómez Penedo , Alice E. Coyne , Manuel Meglio , Marjolein Fokkema , Rebekka Wassmann , Wolfgang Lutz , Julian Rubel","doi":"10.1016/j.brat.2025.104848","DOIUrl":null,"url":null,"abstract":"<div><h3>Aims</h3><div>New innovations in predictive models, such as machine learning, could enhance the effectiveness of measurement-based care systems by generating more accurate session-by-session psychotherapy outcome predictions. In this study, we developed a tree-based model that integrates the strengths of multilevel and machine learning models to predict patients’ trajectories of clinical improvement during cognitive-behavioural therapy (CBT).</div></div><div><h3>Methods</h3><div>We used a sample of 1008 outpatients who were treated at a CBT university clinic in Germany. The total sample was randomly divided into a training (2/3 of the sample) and a test (remaining 1/3) set. Grounded on patient demographic and clinical information at baseline, we developed a generalized linear mixed model tree algorithm to predict patients' session-by-session outcome change during the first ten sessions. <strong>Results</strong>: The best-fitting model in the training set identified 10 groups of patients based on their presenting characteristics and improvement trajectories. In the test set, the algorithm resulted in a correlation of 0.65 between the observed and predicted values for the outcome variable (cross-validation <em>R</em><sup>2</sup> = 0.42). Developing failure boundaries based on the tree-based approach allowed us to correctly identify 67.6 % of the test set patients who did not reliably improve within the first 15 sessions of treatment. <strong>Discussion</strong>: This study provides preliminary support for the integration of multilevel and machine learning models via generalized linear mixed model trees. The algorithms developed could help support routine implementation of precision mental health care strategies by informing therapists’ treatment planning and session-by-session responsiveness for different patient subgroups.</div></div>","PeriodicalId":48457,"journal":{"name":"Behaviour Research and Therapy","volume":"193 ","pages":"Article 104848"},"PeriodicalIF":4.5000,"publicationDate":"2025-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Behaviour Research and Therapy","FirstCategoryId":"102","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0005796725001706","RegionNum":2,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PSYCHOLOGY, CLINICAL","Score":null,"Total":0}

引用次数: 0

Abstract

Aims

New innovations in predictive models, such as machine learning, could enhance the effectiveness of measurement-based care systems by generating more accurate session-by-session psychotherapy outcome predictions. In this study, we developed a tree-based model that integrates the strengths of multilevel and machine learning models to predict patients’ trajectories of clinical improvement during cognitive-behavioural therapy (CBT).

Methods

We used a sample of 1008 outpatients who were treated at a CBT university clinic in Germany. The total sample was randomly divided into a training (2/3 of the sample) and a test (remaining 1/3) set. Grounded on patient demographic and clinical information at baseline, we developed a generalized linear mixed model tree algorithm to predict patients' session-by-session outcome change during the first ten sessions. Results: The best-fitting model in the training set identified 10 groups of patients based on their presenting characteristics and improvement trajectories. In the test set, the algorithm resulted in a correlation of 0.65 between the observed and predicted values for the outcome variable (cross-validation R² = 0.42). Developing failure boundaries based on the tree-based approach allowed us to correctly identify 67.6 % of the test set patients who did not reliably improve within the first 15 sessions of treatment. Discussion: This study provides preliminary support for the integration of multilevel and machine learning models via generalized linear mixed model trees. The algorithms developed could help support routine implementation of precision mental health care strategies by informing therapists’ treatment planning and session-by-session responsiveness for different patient subgroups.

查看原文本刊更多论文

一种多层机器学习算法，用于预测接受认知行为治疗的患者的每次治疗结果

预测模型的新创新，如机器学习，可以通过产生更准确的心理治疗结果预测来提高基于测量的护理系统的有效性。在这项研究中，我们开发了一个基于树的模型，该模型整合了多层次和机器学习模型的优势，以预测患者在认知行为治疗（CBT）期间的临床改善轨迹。方法采用1008例在德国一所CBT大学诊所接受治疗的门诊患者作为样本。总样本随机分为训练集（样本的2/3）和测试集（剩余的1/3）。基于基线的患者人口统计和临床信息，我们开发了一种广义线性混合模型树算法来预测患者在前10次治疗期间每次治疗的结果变化。结果：训练集中的最佳拟合模型根据患者的表现特征和改善轨迹确定了10组患者。在测试集中，该算法得出结果变量的观测值与预测值之间的相关系数为0.65（交叉验证R2 = 0.42）。基于基于树的方法开发的失败边界使我们能够正确识别在前15次治疗中没有可靠改善的67.6%的测试集患者。讨论：本研究通过广义线性混合模型树为多层模型和机器学习模型的集成提供了初步支持。所开发的算法可以通过告知治疗师的治疗计划和对不同患者亚组的逐次响应，帮助支持精确精神卫生保健策略的常规实施。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Behaviour Research and Therapy PSYCHOLOGY, CLINICAL-

CiteScore

7.50

自引率

7.30%

发文量

148

期刊介绍： The major focus of Behaviour Research and Therapy is an experimental psychopathology approach to understanding emotional and behavioral disorders and their prevention and treatment, using cognitive, behavioral, and psychophysiological (including neural) methods and models. This includes laboratory-based experimental studies with healthy, at risk and subclinical individuals that inform clinical application as well as studies with clinically severe samples. The following types of submissions are encouraged: theoretical reviews of mechanisms that contribute to psychopathology and that offer new treatment targets; tests of novel, mechanistically focused psychological interventions, especially ones that include theory-driven or experimentally-derived predictors, moderators and mediators; and innovations in dissemination and implementation of evidence-based practices into clinical practice in psychology and associated fields, especially those that target underlying mechanisms or focus on novel approaches to treatment delivery. In addition to traditional psychological disorders, the scope of the journal includes behavioural medicine (e.g., chronic pain). The journal will not consider manuscripts dealing primarily with measurement, psychometric analyses, and personality assessment.