A demonstration of a multi-method variable selection approach for treatment selection: Recommending cognitive-behavioral versus psychodynamic therapy for mild to moderate adult depression

Objective: Treatment selection could improve outcomes by helping individuals select their optimal intervention. We refine the Personalized Advantage Index approach to generate individual treatment recommendations based on pre-treatment characteristics for adults with depression deciding between cognitive behavioral (CBT) versus psychodynamic therapy (PDT).
Method: Data were drawn from a randomized comparison of CBT versus PDT in a sample of 167 individuals with depression. We introduce a novel method combining four different statistical techniques to identify consistent patient characteristics associated with treatment outcome. We combined these variables to generate predictions indicating the optimal treatment for each patient. We assessed retrospectively the effectiveness of our model by comparing the average treatment outcomes for the patients who received their indicated treatment versus those who did not.
Results: Depression severity, anxiety sensitivity, extraversion, and psychological treatment needs were found to predict differential treatment efficacy. The average post-treatment Hamilton Depression Rating Scale scores was 1.6 points lower (95%CI=[0.5:2.8]; d=0.21) for those who received their indicated treatment compared to non-indicated. Among the 60% of patients with the strongest treatment recommendations, that advantage grew to 2.6 points (95%CI=[1.4:3.7]; d=0.37). Treatment recommendations were improved by combining different statistical techniques to identify moderators of treatment response rather than relying on one method.
Conclusions: Patient characteristics could help individuals choose between CBT and PDT. The small sample and lack of a separate validation sample indicate the need for prospective tests before using this model for treatment selection. These findings contribute to a growing literature on model-guided treatment recommendations in depression.

目标：治疗选择可通过辅助个体选择其最佳干预措施，从而提升治疗结果。本研究对个性化优势指数方法进行优化，旨在根据抑郁症成年患者的治疗前特征生成个体化治疗建议，以认知行为疗法（CBT）与心理动力学疗法（PDT）之间的选择为研究对象。
方法：数据来源于一项包含167名抑郁症患者的CBT与PDT随机比较研究。我们提出一种新颖的方法，融合四种不同的统计技术，以识别与治疗结果一致的病人特征。我们将这些变量相结合，生成针对每位患者的最佳治疗预测。通过回顾性评估，我们将接受其指定治疗的患者与未接受指定治疗的患者平均治疗结果进行比较，以检验我们模型的有效性。
结果：抑郁症严重程度、焦虑敏感性、外向性和心理治疗需求被发现可以预测不同治疗的有效性。与未接受指定治疗的患者相比，接受其指定治疗的患者平均汉密尔顿抑郁量表评分降低1.6分（95%CI=[0.5:2.8]; d=0.21）。在60%治疗建议最为强烈的患者中，这一优势增至2.6分（95%CI=[1.4:3.7]; d=0.37）。通过结合不同的统计技术以识别治疗反应的调节因素，而非依赖单一方法，治疗建议得到改善。
结论：病人特征有助于个体在CBT与PDT之间做出选择。样本量小且缺乏独立的验证样本，表明在使用此模型进行治疗选择之前，需要进行前瞻性测试。这些发现为抑郁症中基于模型的治疗建议的文献研究做出了贡献。