Predicting Internet-Based Treatment Response for Major Depressive Disorder

预测重度抑郁症基于互联网的治疗反应

• 批准号: 9314157
• 负责人: RANDY PATRICK AUERBACH
• 金额: $ 36.47万
• 依托单位: MCLEAN HOSPITAL
• 依托单位国家: 美国
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-08-08 至 2018-07-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9314157
• 关键词: Acute; Address; Aftercare; Algorithms; Boston; Characteristics; Clinical; Cognitive Therapy; Comorbidity; Complex; Data; Data Collection; Depressed mood; Depression screen; Development; Dropout; Effectiveness; Enrollment; Failure; Foundations; Frequencies; Functional Magnetic Resonance Imaging; Goals; Healthcare; Heterogeneity; Individual; Internet; Intervention; Machine Learning; Major Depressive Disorder; Measures; Mental Depression; Mental Health; Mental Health Services; Methods; Modeling; Negative Valence; Neurocognitive; Online Systems; Participant; Patient Self-Report; Patients; Phase; Policy Maker; Positioning Attribute; Positive Valence; Predictive Analytics; Probability; Randomized; Reaction; Recruitment Activity; Recurrence; Reporting; Research; Research Domain Criteria; Research Personnel; Resources; Risk Behaviors; Sample Size; Sampling; Schools; Services; Students; Suicide; Surveys; Symptoms; System; Technology; Testing; Training; Treatment outcome; Universities; Validation; base; brain circuitry; cognitive system; cohort; college; control trial; cost; cost effective; data integration; depressive symptoms; design; effective intervention; effective therapy; ethnic diversity; experience; improved; indexing; innovation; learning strategy; member; minor depressive disorder; minority student; neurocognitive test; neuromechanism; predicting response; prediction algorithm; predictive modeling; programs; reduce symptoms; relating to nervous system; responders and non-responders; response; reward expectancy; screening; suicidal; support tools; treatment as usual; treatment effect; treatment responders; treatment response; university student

As many as 53% of students report experiencing depression during college, and these depressive episodes are associated with a higher frequency of academic problems, comorbidity, and suicide. Although there are effective options for treatment, the majority of individuals (>70%) do not pursue services, and even for those who do, response rates remain modest (~40-50%). As a means of increasing accessibility to treatment, internet-based interventions for depression have been developed and tested. Despite increased availability, response to internet-based interventions continues to vary substantially, and failed treatment often contributes to persistence and worsening of symptoms. Therefore, identifying individuals with a high likelihood of responding to internet-based treatment would represent a major advance and address a critical unmet need. In recent years, promising approaches for testing the heterogeneity of the treatment effects – delineating which individuals are likely to respond to a given treatment – have been developed. However, their use for identifying predictors of treatment response in depression remains unclear. To address this unmet need, the proposed study will test a new, cost-effective, and feasibly-scaled method of predicting differential treatment response following internet-based cognitive behavioral therapy (iCBT) for depression in a large, representative college sample (Boston Consortium of Colleges and Universities which includes 7 schools). Members of the consortium have committed to screening all incoming freshmen (N = ~14,000) through a rigorous online assessment and to offer iCBT to students with elevated levels of depressive symptoms (i.e., minor or major depression). The following steps will be pursued. First, in the initial phase of the study, freshmen students from the Boston Consortium will be screened for depression through an online survey, and they also will complete web-based neurocognitive tasks probing key mechanisms underpinning depression. Depressed students will be invited to enroll in iCBT, and a predictive algorithm will be developed based on assessments across different units of analysis (i.e., clinical characteristics, neurocognitive indices) to identify iCBT responders. Second, after the development phase, an independent sample of college students will be recruited to validate the predictive algorithm. This validation phase will use clinical indicators and neurocognitive data. Additionally, functional magnetic resonance imaging (fMRI) data targeting key mechanisms within the Research Domain Criteria (RDoC) will be acquired from a subset of participants. Neural data will be integrated to determine whether they improve the predictive algorithm. Third, data across independent samples will be combined, which will increase power to refine our predictive model for both acute and sustained response. In summary, there are alarming rates of depression among college students, and only a minority of students utilize mental health services. The proposed research will personalize our approach to depression treatment, which, ultimately, will improve effectiveness and better inform mental health care across college campuses.

多达53%的学生报告在大学期间经历过抑郁症，这些抑郁症患者 发作与更高频率的学业问题、合并症和自杀相关。虽然 有有效的治疗选择，大多数人（>70%）不寻求服务，甚至 对于那些这样做的人，答复率仍然不高（约40-50%）。作为一种增加获得 在治疗方面，已经开发和测试了基于互联网的抑郁症干预措施。尽管增加了 可用性，对基于互联网的干预措施的反应仍然存在很大差异，失败的治疗往往 有助于症状的持续和恶化。因此，确定具有高可能性的个体 对基于互联网的治疗作出反应将是一个重大进步，并解决了一个尚未满足的关键需求。 近年来，有希望的方法来测试治疗效果的异质性-描绘 哪些个体可能对给定的治疗有反应-已经被开发出来。然而，它们的用途 确定抑郁症治疗反应的预测因子仍不清楚。为了满足这一未满足的需求， 一项拟议的研究将测试一种新的、具有成本效益的、规模可行的预测差别待遇的方法 在一个大的，有代表性的，基于互联网的认知行为疗法（iCBT）治疗抑郁症后的反应， 大学样本（波士顿大学联盟，包括7所学校）。成员 联盟已承诺通过严格的在线筛选所有新生（N = ~ 14，000） 评估并为抑郁症状水平升高的学生提供iCBT（即，轻微或重大 抑郁症）。将采取以下步骤。首先，在研究的初始阶段， 波士顿联盟将通过在线调查进行抑郁症筛查，他们还将完成 基于网络的神经认知任务，探索抑郁症的关键机制。抑郁的学生将 被邀请参加iCBT，并将根据评估开发预测算法， 不同的分析单元（即，临床特征、神经认知指数）以鉴定iCBT应答者。 第二，开发阶段结束后，将招募独立的大学生样本进行验证 预测算法。该验证阶段将使用临床指标和神经认知数据。此外，本发明还 针对研究领域内关键机制的功能性磁共振成像（fMRI）数据 标准（RDoC）将从参与者的子集中获取。神经数据将被整合以确定 他们是否改进了预测算法。第三，独立样本的数据将被合并， 这将提高我们改进急性和持续反应预测模型的能力。总的来说， 在大学生中，抑郁症的发病率令人担忧，只有少数学生利用心理治疗。 保健服务这项拟议中的研究将使我们的抑郁症治疗方法个性化， 最终，将提高效率，并更好地为整个大学校园的精神卫生保健提供信息。