生成式AI在数字心理干预中的疗效因子:整合共同因素与AI特异性机制

注册号:

Registration number:

ChiCTR2600122417 

最近更新日期:

Date of Last Refreshed on:

2026-04-13 18:04:21 

注册时间:

Date of Registration:

2026-04-13 00:00:00 

注册号状态:

预注册

Registration Status:

Prospective registration

注册题目:

生成式AI在数字心理干预中的疗效因子:整合共同因素与AI特异性机制

Public title:

Therapeutic Factors of Generative AI in Digital Psychological Interventions: Integrating Common Factors and AI-Specific Mechanisms

注册题目简写:

English Acronym:

研究课题的正式科学名称:

生成式AI在数字心理干预中的疗效因子:整合共同因素与AI特异性机制

Scientific title:

Therapeutic Factors of Generative AI in Digital Psychological Interventions: Integrating Common Factors and AI-Specific Mechanisms

研究课题代号(代码):

Study subject ID:

在二级注册机构或其它机构的注册号:

The registration number of the Partner Registry or other register:

申请注册联系人:

姚炫钰 

研究负责人:

岸本鹏子 

Applicant:

Yao Xuanyu 

Study leader:

An Benpengzi 

申请注册联系人电话:

Applicant telephone:

+86 183 8570 2498

研究负责人电话:

Study leader's
telephone:

+86 159 1066 5207

申请注册联系人传真 :

Applicant Fax:

研究负责人传真:

Study leader's fax:

申请注册联系人电子邮件:

Applicant E-mail:

yaoxuanyu@foxmail.com

研究负责人电子邮件:

Study leader's E-mail:

k.kishimoto@bnu.edu.cn

申请单位网址(自愿提供):

Applicant website(voluntary supply):

研究负责人网址(自愿提供):

Study leader's website(voluntary supply):

申请注册联系人通讯地址:

中国陕西省西安市雁塔区长安南路199号

研究负责人通讯地址:

中国北京市海淀区新街口外大街19号

Applicant address:

199 South Chang'an Rd, Yanta District, Xi'an, Shaanxi, China

Study leader's address:

19 Xinjiekouwai St, Haidian District, Beijing, China

申请注册联系人邮政编码:

Applicant postcode:

710062

研究负责人邮政编码:

Study leader's postcode:

100875

申请人所在单位:

陕西师范大学心理学院

Applicant's institution:

School of Psychology, Shaanxi Normal University

研究负责人所在单位:

北京师范大学心理学部

Affiliation of the Leader:

Faculty of Psychology, Beijing Normal University

是否获伦理委员会批准:

Approved by ethic committee:

Yes

伦理委员会批件文号:

Approved No. of ethic committee:

BNU202601120017

伦理委员会批件附件:

Approved file of Ethical Committee:

查看附件View

批准本研究的伦理委员会名称:

北京师范大学心理学部伦理委员会

Name of the ethic committee:

Institutional Review Board of the Faculty of Psychology, BNU

伦理委员会批准日期:

Date of approved by ethic committee:

2026-01-05 00:00:00

伦理委员会联系人:

熊炳森

Contact Name of the ethic committee:

Xiong Bingsen

伦理委员会联系地址:

中国北京市海淀区新街口外大街19号

Contact Address of the ethic committee:

19 Xinjiekouwai St, Haidian District, Beijing, China

伦理委员会联系人电话:

Contact phone of the ethic committee:

+86 10 5880 6836

伦理委员会联系人邮箱:

Contact email of the ethic committee:

psy_ethics@bnu.edu.cn

研究实施负责(组长)单位:

北京师范大学心理学部

Primary sponsor:

Faculty of Psychology, Beijing Normal University

研究实施负责(组长)单位地址:

中国北京市海淀区新街口外大街19号

Primary sponsor's address:

19 Xinjiekouwai St, Haidian District, Beijing, China

试验主办单位(项目批准或申办者):

Secondary sponsor:

国家:

中国

省(直辖市):

北京

市(区县):

Country:

China

Province:

Beijing

City:

单位(医院):

北京师范大学心理学部

具体地址:

中国北京市海淀区新街口外大街19号

Institution
hospital:

Faculty of Psychology, Beijing Normal University

Address:

19 Xinjiekouwai St, Haidian District, Beijing, China

经费或物资来源:

自筹

Source(s) of funding:

Self support

研究疾病:

情绪障碍  

Target disease:

Emotional disorders

研究疾病代码:

Target disease code:

研究类型:

干预性研究

Study type:

Interventional study

研究所处阶段:

探索性研究/预试验 

Study phase:

0

研究设计:

随机平行对照 

Study design:

Parallel 

研究目的:

本研究旨在探究生成式人工智能(Artificial Intelligence, AI)在基于情绪障碍跨诊断治疗的统一方案(Unified Protocol, UP)数字心理健康干预中的起效机制。重点考察传统咨询中的共同因素(如期望、真实关系、工作联盟)与AI特有因素(如可及性、技术可信度与去社会化等)如何在干预过程中影响咨询效果。并以此构建AI心理干预的纵向中介模型,为未来AI心理干预的设计提供理论依据。  

Objectives of Study:

This study aims to investigate the therapeutic mechanisms of Generative Artificial Intelligence (AI) in digital mental health interventions based on the Unified Protocol for Transdiagnostic Treatment of Emotional Disorders (UP). Specifically, it examines how common factors in traditional counseling (e.g., expectancy, real relationship, working alliance) and AI-specific factors (e.g., accessibility, technological credibility, and de-socialization) jointly influence therapeutic outcomes during the intervention process. By constructing a longitudinal mediation model of AI-based interventions, this study seeks to provide a theoretical basis for the design of future AI psychological interventions.

药物成份或治疗方案详述:

 

Description for medicine or protocol of treatment in detail:

 

纳入标准:

Inclusion criteria

排除标准:

1. 基线筛查显示 PHQ-9 第 9 题(自杀意念)得分 > 0,或在实验中发现存在自伤、自杀风险者; 2. 既往或现病史符合精神分裂症、双相情感障碍或精神病性症状诊断者; 3. 目前正在接受其他高强度心理咨询或服用精神类药物; 4. 存在酒精或药物滥用史。

Exclusion criteria:

1. A score of > 0 on item 9 of the PHQ-9 (suicidal ideation) at baseline, or identified risk of self-harm or suicide during the experiment; 2. A history or current diagnosis of schizophrenia, bipolar disorder, or psychotic symptoms; 3. Currently receiving other high-intensity psychological counseling or taking psychotropic medications; 4. A history of alcohol or substance abuse.

研究实施时间:

Study execute time:

From 2025-09-01 00:00:00 To 2027-06-01 00:00:00  

征募观察对象时间:

Recruiting time:

From 2026-04-13 00:00:00 To 2027-04-12 00:00:00

干预措施:

Interventions:

组别:

AI即时干预组

样本量:

375

Group:

AI+UP Intervention Group

Sample size:

干预措施:

微信小程序接入Claude Opus 4.6大模型,基于UP(跨诊断统一方案)取向,包含8节结构化课程(情绪觉察、认知重评等)及全天候自由对话模式;干预周期4周(28天)

干预措施代码:

Intervention:

WeChat mini-program integrated with Claude Opus 4.6 large model, based on UP (cross-diagnostic unified protocol) approach, including 8 structured modules (emotional awareness, cognitive reappraisal, etc.) and 24/7 free conversation mode; intervention duration of 4 weeks (28 days)

Intervention code:

组别:

等待对照组

样本量:

125

Group:

Waitlist Control Group

Sample size:

干预措施:

等待期(无干预);等待期结束后(研究结束前)获得完整干预机会

干预措施代码:

Intervention:

Waitlist period (no intervention); full intervention opportunity provided after waitlist period (before study end)

Intervention code:

研究实施地点:

Countries of recruitment and research settings:

国家:

中国

省(直辖市):

北京 

市(区县):

 

Country:

China

Province:

Beijing

City:

单位(医院):

北京师范大学心理学部 

单位级别:

大学 

Institution
hospital:

Faculty of Psychology, Beijing Normal University

Level of the institution:

University

测量指标:

Outcomes:

指标中文名:

抑郁症状

指标类型:

主要指标

Outcome:

Depressive Symptoms

Type:

Primary indicator

测量时间点:

基线 (T0)、第 2 周 (T1)、第 4 周 (T2)、第 6 周 (T3)、第 8 周 (T4)

测量方法:

患者健康问卷 -9 项 (PHQ-9)

Measure time point of outcome:

Baseline (T0), Week 2 (T1), Week 4 (T2), Week 6 (T3), Week 8 (T4)

Measure method:

Patient Health Questionnaire-9 (PHQ-9)

指标中文名:

焦虑症状

指标类型:

主要指标

Outcome:

Anxiety Symptoms

Type:

Primary indicator

测量时间点:

基线 (T0)、第 2 周 (T1)、第 4 周 (T2)、第 6 周 (T3)、第 8 周 (T4)

测量方法:

广泛性焦虑量表 -7 项 (GAD-7)

Measure time point of outcome:

Baseline (T0), Week 2 (T1), Week 4 (T2), Week 6 (T3), Week 8 (T4)

Measure method:

Generalized Anxiety Disorder-7 (GAD-7)

指标中文名:

真实关系量表 - 来访者版

指标类型:

次要指标

Outcome:

Real Relationship Inventory - Client Version (RRI-C)

Type:

Secondary indicator

测量时间点:

第 2 周 (T1)、第 4 周 (T2)、第 6 周 (T3)、第 8 周 (T4)

测量方法:

Measure time point of outcome:

Week 2 (T1), Week 4 (T2), Week 6 (T3), Week 8 (T4)

Measure method:

指标中文名:

工作联盟量表 - 短版

指标类型:

次要指标

Outcome:

Working Alliance Inventory - Short Revised (WAI-SR)

Type:

Secondary indicator

测量时间点:

第 2 周 (T1)、第 4 周 (T2)、第 6 周 (T3)、第 8 周 (T4)

测量方法:

Measure time point of outcome:

Week 2 (T1), Week 4 (T2), Week 6 (T3), Week 8 (T4)

Measure method:

指标中文名:

可信度与预期

指标类型:

次要指标

Outcome:

Credibility and Expectancy Questionnaire (CEQ)

Type:

Secondary indicator

测量时间点:

基线 (T0)、第 2 周 (T1)、第 4 周 (T2)、第 6 周 (T3)、第 8 周 (T4)

测量方法:

Measure time point of outcome:

Baseline (T0), Week 2 (T1), Week 4 (T2), Week 6 (T3), Week 8 (T4)

Measure method:

指标中文名:

治疗参与度

指标类型:

次要指标

Outcome:

Treatment Engagement

Type:

Secondary indicator

测量时间点:

全程(每日)

测量方法:

后台提取对话轮数与总字数

Measure time point of outcome:

Throughout the study (daily)

Measure method:

Backend extraction of conversation turns and total word count

指标中文名:

AI 独特因素(可及性、去社会化、技术可信度、感知个性化、讨好性、非评判性、隐私保护、AI 偏见、感知安全风险)

指标类型:

次要指标

Outcome:

AI-specific Factors (Accessibility, De-socialization, Technological Trustworthiness, Perceived Personalization, Favorability, Non-judgmental Attitude, Privacy Protection, AI Bias, Perceived Safety Risk)

Type:

Secondary indicator

测量时间点:

基线 (T0)、第 2 周 (T1)、第 4 周 (T2)、第 6 周 (T3)、第 8 周 (T4)

测量方法:

自编问卷

Measure time point of outcome:

Baseline (T0), Week 2 (T1), Week 4 (T2), Week 6 (T3), Week 8 (T4)

Measure method:

Self-developed questionnaire

指标中文名:

安全性指标(包括:用户端高危风险(自杀/自伤意图)、AI 端生成风险(医疗越界、诱导伤害等)

指标类型:

主要指标

Outcome:

Safety Outcomes (including user-side high-risk events such as suicidal/self-harm intent, and AI-side generated risks such as medical boundary violations, inducement of harm, etc.)

Type:

Primary indicator

测量时间点:

测量方法:

Measure time point of outcome:

Measure method:

采集人体标本:

Collecting sample(s)
from participants:

标本中文名:

组织:

Sample Name:

None

Tissue:

人体标本去向

其它  

说明

Fate of sample:

0thers  

Note:

征募研究对象情况:

Recruiting status:

尚未开始

Not yet recruiting

年龄范围:

Participant age:

最小 Min age 18 years
最大 Max age 60 years

性别:

男女均可

Gender:

Both

随机方法(请说明由何人用什么方法产生随机序列):

本研究采用区组随机化(Block Randomization)方法产生分配序列,干预组与等待组的分配比例为 3:1。随机序列由一名研究人员使用 Microsoft Excel 生成。设置固定区组大小为 8(即每个区组内包含6个干预组名额与2个等待组名额),共生成60个区组以覆盖全部480名受试者。生成的随机分配表已通过数值固化保存,后续将严格按照受试者的入组顺序进行分配。

Randomization Procedure (please state who generates the random number sequence and by what method):

The allocation sequence was generated using block randomization with a 3:1 allocation ratio (intervention to waiting list control). The sequence was created by a researcher using Microsoft Excel. A fixed block size of 8 was utilized (comprising 6 intervention and 2 waiting list slots per block), resulting in 60 blocks to cover the total sample of 480 participants. The randomization list has been finalized and saved, and group assignment will be administered strictly according to the order of participant enrollment.

是否公开试验完成后的统计结果:

Calculated Results after the Study Completed public access:

公开/Public

盲法:

鉴于本研究采用等待组对照设计,且主试需承担受试者安全监控职责,干预实施阶段无法实现双盲。为最大限度减少实验者偏差,本研究在基线测评与数据分析阶段实施盲法控制: 1. 首先,在完成基线(T0)测评前,主试与受试者均处于盲态,确保基线数据不受期望效应影响; 2. 其次,数据分析阶段实行结果评估者盲法(Outcome Assessor Blinding),数据分析人员仅获知组别代码(如 Group A/B),而非干预组/等待组,直至分析全部完成后方可揭盲。

Blinding:

Given that this study adopts a waitlist control design and the experimenter is responsible for participant safety monitoring, double-blinding cannot be achieved during the intervention implementation phase. To minimize experimenter bias to the greatest extent possible, blinding controls are implemented at the baseline assessment and data analysis stages: 1. First, prior to the completion of baseline (T0) assessment, both the experimenter and the participant are blinded, ensuring that baseline data are not influenced by expectancy effects; 2. ,Second, outcome assessor blinding is implemented during the data analysis phase — the data analyst is provided only with group codes (e.g., Group A/B) rather than the intervention/waitlist group labels, and unblinding occurs only after all analyses have been completed.

试验完成后的统计结果(上传文件):

Calculated Results after
the Study Completed(upload file):

是否共享原始数据:

IPD sharing

是Yes

共享原始数据的方式(说明:请填入公开原始数据日期和方式,如采用网络平台,需填该网络平台名称和网址):

在2027年12月31日后共享原始数据,如有需要可联系项目负责人岸本鹏子。

The way of sharing IPD”(include metadata and protocol, If use web-based public database, please provide the url):

Share raw data after December 31, 2027, by contacting project lead Tomoko Kishimoto if required.

数据采集和管理(说明:数据采集和管理由两部分组成,一为病例记录表(Case Record Form, CRF),二为电子采集和管理系统(Electronic Data Capture, EDC),如ResMan即为一种基于互联网的EDC:

实验全程线上实施,问卷收集使用Credamo见数平台。在实验过程中,收集的信息不涉及受试者隐私。受试者与AI的对话记录和填写的问卷、量表仅供科研使用,将做匿名处理并严格保密。

Data collection and Management (A standard data collection and management system include a CRF and an electronic data capture:

The experiment was conducted online, and questionnaires were collected using the Credamo database platform. During the experiment, the information collected did not involve subject privacy. The records of the conversation between the subjects and the AI and the questionnaires and scales filled in are for scientific research use only, and will be anonymized and kept strictly confidential.

数据与安全监察委员会:

Data and Safety Monitoring Committee:

暂未确定/Not yet

注册人:

Name of Registration:

 2026-04-13 18:04:12