Artificial Intelligence (AI) is a crucial component of China's new-generation information technology development strategy. With its extensive applications in the healthcare field, the integration of AI into general practice for auxiliary diagnosis and chronic disease management plays a significant role in enhancing primary healthcare services and improving the professional skills and clinical decision-making capabilities of general practitioners.

Through a preliminary investigation into general practitioners' cognition of AI technology in the medical field and their needs for constructing auxiliary diagnosis and treatment systems across various provinces and cities, this study aims to provide references for establishing more clinically applicable AI-assisted diagnosis and treatment systems.

From March to April 2024, we selected general practitioners (GPs) from across China's provinces and municipalities as study participants through national GP WeChat groups. Using a self-designed questionnaire, we collected data on: GPs' basic demographic information; their attitudes toward and understanding of AI-assisted diagnosis and treatment systems; and their needs regarding AI applications in daily clinical practice (including clinical consultations, auxiliary diagnosis and treatment, referral and follow-up, and application formats) . Data collection was conducted via the"Questionnaire Star" platform, and descriptive analysis methods were employed to present and characterize the results.

This study enrolled 382 general practitioners (GPs) from 27 provinces across China, comprising 52.36% (200/382) from general hospitals and 47.64% (182/382) from primary healthcare institutions. The survey revealed that 57.69% (220/382) of respondents believed physicians and AI should work synergistically, while 77.75% (297/382) of practitioners expressed willingness to adopt or continue using AI-assisted diagnostic technologies in clinical practice; The survey identified key advantages of AI adoption, including improved diagnostic efficiency [89.01% (340/382) ] , reduced clinical workload [84.29% (322/382) ] , and decreased misdiagnosis rates 80.10% (306/382) . However, significant concerns were noted regarding: over-reliance on AI diagnosis [75.13% (287/382) , ethical implications [64.14% (245/382) ] , algorithmic diagnostic bias [63.09% (241/382) ] , data security vulnerabilities [57.33% (219/382) ] , and liability ambiguity in medical incidents [53.66% (205/382) ] ; In the clinical consultation needs assessment, general practitioners most frequently requested AI-enabled alerts for"life-threatening conditions" [61.25% (234/382) ] and prompts for"urgently actionable clinical signs" [62.04% (237/382) ] . Significant demand was also observed for"associated symptom inquiry prompts" [58.90% (225/382) ] and"systematic clinical reasoning guidance" [62.57% (239/382) ] ; In the auxiliary diagnosis and treatment domain, rapid and precise imaging evaluation [67.80% (259/382) ] and prescription review [67.01% (256/382) ] emerged as the most anticipated AI functions. Significant attention was also given to personalized medication recommendations [63.88% (244/382) ] and insurance coverage alerts [66.49% (254/382) ] ; Regarding referral and follow-up management, general practitioners showed the strongest demand for three key AI functions: automated reminders for follow-up visit scheduling [72.25% (276/382) ] , remote monitoring capabilities through smart devices [71.46% (273/382) ] and personalized delivery of patient education content [70.42% (269/382) ] ; In terms of system application formats, general practitioners showed the strongest preference for"reviewing patients' historical medical records" [73.04% (279/382) ] and"voice-to-text transcription" [71.73% (274/382) ] ; In terms of priorities for incorporating undifferentiated diseases into AI-assisted diagnostic and treatment systems, 37.43% (143/382) of general practitioners recommended prioritizing the inclusion of fever in AI-assisted diagnosis and treatment systems.

General practitioners demonstrated overall high willingness to adopt AI technology for clinical diagnosis and treatment. The AI-assisted clinical decision support system showed significant demand across multiple clinical workflows, including medical history collection, physical examination, diagnostic testing, treatment planning, referral criteria determination, follow-up management, and system application formats. Moving forward, the development of AI-assisted diagnosis and treatment systems tailored for Chinese general practitioners—through overcoming technological limitations, establishing clear legal frameworks and guidelines for AI, and enhancing GPs' AI literacy—will significantly improve primary care physicians' first-contact diagnostic capabilities and facilitate the implementation of tiered healthcare delivery systems.

人工智能（AI）是国家新一代信息技术发展战略的重要部分，在医疗健康领域应用广泛，将AI应用于全科医疗辅助诊断和慢性病管理，对提升基层医疗水平、提高全科医师专业技能和临床决策能力有作用。

调查各省（区、市）全科医师对AI技术在医疗领域的认知及对构建辅助诊疗系统的需求，为建立更贴合临床应用的辅助诊疗系统提供参考。

2024年3—4月通过国内全科医师微信群选取全国各省（区、市）全科医师为调查对象，采用自行设计的调查问卷，包括全科医师的基本信息、对AI辅助诊疗系统的应用态度及认知、AI辅助诊疗应用于日常诊疗（临床接诊、辅助诊治、转诊随访、应用形式）的需求，利用"问卷星"平台进行数据收集，采用描述性分析方法对结果进行呈现和描述。

本研究共纳入382名全科医师，覆盖27个省（区、市），其中综合医院占52.36%（200/382），基层医疗卫生机构占47.64%（182/382）。调查显示，57.69%（220/382）认为医生与AI应相辅相成；77.75%（297/382）的医师愿意尝试或继续使用AI辅助诊疗技术；AI应用优势以提升诊疗效率［89.01%（340/382）］、减轻临床工作负担［84.29%（322/382）］、减少误诊［80.10%（306/382）］为主，但仍存在过度依赖AI诊疗、涉及伦理问题、AI算法易出现诊断偏差、数据安全性欠佳、医疗事故无法划分医疗责任等弊端，分别占75.13%（287/382）、64.14%（287/382）、63.09%（241/382）、57.33%（219/382）、53.66%（205/382）；在临床接诊需求调查中，全科医师最需要AI提供与主诉相关的"致命性疾病"警示［61.25%（234/382）］和"紧急需要处理的阳性体征"提示［62.04%（237/382）］，同时对"伴随症状"询问提示［58.90%（225/382）］和"系统性思维导图"参照提示［62.57%（239/382）］也有较高需求；辅助诊治环节中，快速精准影像结果评估［67.80%（259/382）］和"处方审核"诊疗辅助提示［67.01%（256/382）］成为最受期待的功能，个体化"推荐用药"辅助提示［63.88%（244/382）］和"医保提醒"提示［66.49%（254/382）］也备受关注；在转诊随访管理方面，全科医师对个体化推送患者"健康教育"内容［70.42%（269/382）］、及时提示患者按照随访时间预约复诊［72.25%（276/382）］和利用智能设备远程监护功能［71.46%（273/382）］需求最高；系统应用形式上，全科医师偏好设置"回顾患者历史就诊记录"功能［73.04%（259/382）］和设置"语音转文字录入"功能［71.73%（259/382）］；最希望未分化疾病病种率先纳入AI辅助诊疗系统方面，37.43%（143/382）的全科医师建议优先将发热纳入AI辅助诊疗系统。

全科医师对AI技术用于临床诊疗应用意愿总体较高。AI辅助诊疗系统对病史采集、体格检查、实验室检查、诊疗计划、转诊指征、随访管理、应用形式均有一定需求。未来通过改善信息技术的局限性、明确AI法律框架和指导原则、加强全科医师AI素养等方式，构建适用于中国全科医师的AI辅助诊疗系统，从而提高全科医师的疾病首诊能力，推动分级诊疗制度落地。