The impact of initial depression and anxiety levels on symptom severity change during psychological treatment for common mental disorders: a systematic review

Depression and anxiety are highly comorbid conditions, which may impact treatment response.
We intended to examine how depression and anxiety symptoms change during psychological interventions for adults and how the initial severity is associated with distinct trajectory classes.
We identified papers examining longitudinal changes in depression and anxiety. Trajectory classes were grouped into responders, improvers and non-improvers. Narrative synthesis, alongside meta-analysis, was applied to explore the number and nature of classes, their associations with study characteristics and predictors of trajectory class membership. This study was pre-registered on International Prospective Register of Systematic Reviews (PROSPERO) (CRD42023393155), and the authors followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance.
We identified 18 papers, including 22 models with two to six classes. Responding and non-improving classes prevailed. Meta-regression revealed no characteristics associated with the proportion of these classes. Initial depression and anxiety levels predicted trajectory class membership for each other, but not always non-response. Unemployment, suicidality, social functioning impairment, treatment type and medication use predicted following non-responding trajectory.
Our findings highlight the complex relationship between initial anxiety and depression severity and their changes during psychological interventions. They provide practitioners with insights into factors associated with non-response patterns of symptom change.

抑郁与焦虑为高度共病（comorbid）的病症，二者可对治疗应答产生影响。本研究旨在探讨成人接受心理干预期间，抑郁与焦虑症状的变化规律，以及初始症状严重程度与不同症状轨迹类别（trajectory classes）的关联。
我们检索了考察抑郁与焦虑症状纵向变化的相关文献，将症状轨迹类别划分为应答组（responders）、改善组（improvers）与无改善组（non-improvers）。本研究采用叙事综合（narrative synthesis）结合Meta分析（meta-analysis）的方法，探究症状轨迹类别的数量与特征、其与研究特征的关联，以及轨迹类别归属的预测因素。本研究已在国际前瞻性系统评价注册库（International Prospective Register of Systematic Reviews, PROSPERO）完成预注册，注册号为CRD42023393155，研究团队遵循《系统评价与Meta分析优先报告条目》（Preferred Reporting Items for Systematic Reviews and Meta-Analyses, PRISMA）的规范开展工作。
本研究共纳入18篇文献，包含22个模型，其症状轨迹类别数量为2至6类。其中以应答组与无改善组占比最高。Meta回归分析未发现与上述类别占比相关的研究特征。初始抑郁与焦虑水平可互相预测对方的轨迹类别归属，但并非总能预测无应答情况。失业、自杀意念、社会功能受损、治疗类型与药物使用情况可预测后续出现无应答的症状轨迹。
本研究结果揭示了初始焦虑与抑郁严重程度与其在心理干预期间的症状变化之间存在复杂关联，可为临床从业者提供与症状变化无应答模式相关的影响因素的参考依据。