Validation of an Artificial Intelligence (AI)-Powered Clinical Decision Support System for Anxiety and Depression in Primary Care Settings
试验速览
- 阶段
- 不适用
- 状态
- 招募中
- 发起方
- 入组人数
- 100
- 试验地点
- 2
- 主要终点
- Successful Production and Delivery of the Companion Brief
研究概览
简要总结
This prospective, single-arm, observational pilot study evaluates a clinical decision support system developed by Singula Health for use in primary care. Adult patients complete a self-report assessment before a routine primary care visit. The system uses the responses to generate a clinician-facing Companion Brief that summarizes relevant information for the treating clinician. Each Companion Brief is reviewed by the Principal Investigator before it is provided to the clinician. The study will evaluate whether the system can reliably generate and deliver the Companion Briefs, whether clinicians find the Briefs useful and acceptable, and whether patients can successfully complete and accept the pre-visit assessment.
详细描述
The Singula clinical decision support system is designed to organize information collected from a patient before a routine primary care visit into a structured Companion Brief for the treating clinician. The pre-visit assessment includes validated screening measures and structured questions addressing behavioral health symptoms and other factors relevant to the patient's clinical context. After a patient completes the assessment, an automated software pipeline processes the responses and generates a draft Companion Brief. The Principal Investigator reviews each Artificial Intelligence (AI)-generated Brief for clinical accuracy, appropriate safety language, and potential errors before it is released to the treating clinician. The treating clinician receives the reviewed Brief before the patient's routine primary care visit and retains full independent authority over clinical assessment and decision-making. The Companion Brief is informational and does not direct treatment. This Phase 0.1 pilot is intended primarily to characterize the feasibility and operational performance of this workflow in real-world primary care. The study will examine successful generation and delivery of Companion Briefs, pipeline performance, clinician feedback regarding the utility and acceptability of the Briefs, patient completion of the pre-visit assessment, and errors or other operational issues identified during the pilot. Findings from this study are intended to inform subsequent technical, operational, and clinical development of the system.
研究设计
- 研究类型
- Observational
- 观察模型
- Cohort
- 时间视角
- Prospective
入排标准
- 年龄范围
- 18 Years 至 —（Adult, Older Adult）
- 性别
- All
- 接受健康志愿者
- 是
入选标准
- •Adult, aged 18 years or older at the time of consent.
- •Scheduled for a routine, non-urgent primary care visit at a participating practice.
- •Able to read and understand English at a level sufficient to complete the electronic informed consent and Survey 0.1 independently.
- •Has personal-device access to complete the survey (smartphone, tablet, or computer with internet).
- •Capable of providing informed consent.
排除标准
- •Non-English-speaking, defined as inability to read English at a level sufficient to complete electronic informed consent and the survey independently.
- •Visually impaired to a degree that prevents independent reading of an electronic consent form on a personal device, without specialized assistive technology beyond standard mobile-device accessibility features.
- •Unable to read for any reason (literacy, cognitive impairment, or other).
- •Presenting with an acute psychiatric emergency or in acute crisis at the time of the scheduled visit, as identified by clinic staff during routine check-in.
- •Patient is a known minor at the time of scheduling (under 18 years).
- •Patient or clinician declines participation.
研究组 & 干预措施
Primary Care Patient Cohort
Adult primary care patients evaluated during routine clinical care using the Singula Artificial Intelligence (AI)-powered clinical decision support system for assessment of anxiety and depression. Participants are included in a single observational cohort; no study-assigned intervention is administered.
结局指标
主要结局
Successful Production and Delivery of the Companion Brief
时间窗: Up to 7 days
Proportion of enrolled patient cases for which an Artificial Intelligence (AI)-generated Companion Brief is successfully produced, reviewed by the Principal Investigator, and delivered to the treating clinician prior to the patient's scheduled visit.
Clinician-Rated Overall Clinical Utility of the Companion Brief
时间窗: Periodic feedback during active enrollment, approximately 30 to 90 days
Clinician-rated overall clinical utility of the Companion Brief, captured on the periodic batched feedback survey using a 5-point Likert scale. Target: ≥70% of feedback batches indicating overall utility of "useful" or "very useful" across the cases covered in each batch.
次要结局
- Patient Survey Dropout and Skip Patterns(Up to 7 days)
- Survey 0.1 Completion Rate(Up to 7 days)
- Pipeline End-to-End Latency(Up to 72 hours)
- Principal Investigator Corrections to Large Language Model (LLM)-Generated Reports(Up to 7 days)
- Number of Safety-Flag Triggers Generated by the Clinical Decision Support System(Up to 8 days)
- Number of Safety-Flag Triggers by Documented Disposition(Up to 8 days)
- Clinician-Rated Most Useful Companion Brief Section(Periodic feedback during active enrollment, approximately 30 to 90 days)
- Number of Feedback Batches With Clinician-Reported Influence on Clinical Decision-Making, Measured by the Singula Clinician Feedback Survey(Periodic feedback during active enrollment, approximately 30 to 90 days)
- Clinician-Reported Willingness to Continue Using the Companion Brief, Measured by the Singula Clinician Feedback Survey(Periodic feedback during active enrollment, approximately 30 to 90 days)