目的：建立和验证一个涉及多级临床场景的白内障协作通用的人工智能(artificial intelligence,AI)管理平台，探索基于AI的医疗转诊模式，以提高协作效率和资源覆盖率。方法：训练和验证的数据集来自中国AI医学联盟，涵盖多级医疗机构和采集模式。使用三步策略对数据集进行标记: 1)识别采集模式；2)白内障诊断包括正常晶体眼、白内障眼或白内障术后眼；3)从病因和严重程度检测需转诊的白内障患者。此外，将白内障AI系统与真实世界中的居家自我监测、初级医疗保健机构和专科医院等多级转诊模式相结合。结果：通用AI平台和多级协作模式在三步任务中表现出可靠的诊断性能: 1)识别采集模式的受试者操作特征(receiver operating characteristic curve,ROC)曲线下面积(area under the curve,AUC)为99.28%~99.71%)；2)白内障诊断对正常晶体眼、白内障或术后眼，在散瞳-裂隙灯模式下的AUC分别为99.82%、99.96%和99.93%，其他采集模式的AUC均 > 99%；3)需转诊白内障的检测(在所有测试中AUC >91%)。在真实世界的三级转诊模式中，该系统建议30.3%的人转诊，与传统模式相比，眼科医生与人群服务比率大幅提高了10.2倍。结论：通用AI平台和多级协作模式显示了准确的白内障诊断性能和有效的白内障转诊服务。建议AI的医疗转诊模式扩展应用到其他常见疾病和资源密集型情景当中。

Objective: To establish and validate a universal artificial intelligence (AI) platform for collaborative management of cataracts involving multilevel clinical scenarios and explored an AI-based medical referral pattern to improve collaborative efficiency and resource coverage. Methods: The training and validation datasets were derived from the Chinese Medical Alliance for Artificial Intelligence, covering multilevel healthcare facilities and capture modes. The datasets were labelled using a three step strategy: (1)capture mode recognition; (2) cataract diagnosis as a normal lens, cataract or a postoperative eye and (3) detection of referable cataracts with respect to aetiology and severity. Moreover, we integrated the cataract AI agent with a real-world multilevel referral pattern involving self-monitoring at home, primary healthcare and specialised hospital services. Results: The universal AI platform and multilevel collaborative pattern showed robust diagnostic performance in three-step tasks: (1) capture mode recognition (area under the curve (AUC) 99.28%–99.71%), (2) cataract diagnosis (normal lens, cataract or postoperative eye with AUCs of 99.82%, 99.96% and 99.93% for mydriatic-slit lamp mode and AUCs >99% for other capture modes) and (3)detection of referable cataracts (AUCs >91% in all tests). In the real-world tertiary referral pattern, the agent suggested 30.3%  of people be ’referred’, substantially increasing the ophthalmologist-to-population service ratio by 10.2-fold compared with the traditional pattern. Conclusions: The universal AI platform and multilevel collaborative pattern showed robust diagnostic performance and effective service for cataracts. The context of our AI-based medical referral pattern will be extended to other common disease conditions and resource-intensive situations.

目的：探索智能语音随访系统在医疗场景中的新型应用服务模式并分析其在新冠肺炎疫情期间的应用效果，以此评估该系统应用于互联网医院开展医疗咨询服务的实际效能。方法：本研究应用智能语音随访系统针对先天性白内障患儿术后的常见问题进行回访。首先，针对随访目的，设计出完善的结构化随访内容与步骤。其次，部署智能外呼系统自动拨打用户电话，并通过语音识别技术对用户的每次应答进行识别，根据用户的应答自动跳转到下一个随访步骤，在完成一系列问答后根据用户的回答给出恰当的建议，实现电话随访的自动化与智能化。收集2020年2月24日至2月28日期间，智能语音随访系统随访的电话内容、呼叫时间、患儿资料等数据，采用描述性统计分析。结果：2020年2月24日至2月28日期间，中山大学中山眼科中心应用智能语音随访系统电话共随访1154例，其中收到有效回访数据561例，平均有效回访率48.6%。有效回访人群中，有204位(36.4%)家属认为疫情期间复诊时间延长，对宝宝眼睛的恢复有影响，309位(55.1%)家属认为对宝宝眼睛的恢复没有影响。360位(64.2%)先天性白内障患儿眼睛恢复情况良好，没有出现不良反应，169位(30.1%)患儿出现不良反应和体征，包括瞳孔区有白点，眼睛发红和有眼屎流眼泪等。统计患儿不同行为显示，有417位(74.3%)患儿佩戴眼镜，135位(24.1%)患儿没有佩戴眼镜，另有9位(1.6%)患儿佩戴眼镜情况不清楚，经常揉眼的患儿更容易出现眼睛发红(20.4%)、眼睛有眼屎或流眼泪(17.0%)和瞳孔区有白点(6.8%)等不良反应。结论：智能语音随访系统在临床随访中显示出巨大的应用潜力，可作为一种新型的智能医疗服务模式。

Objective: This study was designed to explore its potential value for new medical service model based on the intelligent voice follow-up system and analyze its application effect during the outbreak of COVID-19. The actual effectiveness of this intelligent voice follow-up system applied in the Internet hospital to carry out medical consultation service was discussed. Methods: In this study, an intelligent voice follow up system was developed for postoperative follow-up of children with congenital cataract. First, a well-designed and structured questionnaire contents were developed for postoperative follow-up. Secondly, the intelligent voice follow-up system was deployed. The system would automatically jump to the next follow-up step according to the user’s response, and give appropriate suggestions. Finally, the data of telephone recording, call time, children’s attributes were collected and statistically analyzed. Results: From February 24 to March 15, 2020, 561 families of children with congenital cataract from Zhongshan Ophthalmic Center were recruited by using the intelligent voice follow-up system. The system completed a total of 1 154 calls, of which 561 cases received follow-up data, reaching an average effective call rate of 48.6%. Among 561 cases, 204 (36.4%) thought that the extended time of follow-up visit would affect the recovery of children, while 309 (55.1%) thought that it exerted no effect on the recovery. 360 children (64.2%) achieved good ocular recovery without complications, whereas 169 cases (30.1%) developed ocular symptoms. These include white spots in the pupil area, redness and eye secretions. Statistics of different behavior of children showed that there were 417 (74.3%) children wearing glasses, 135 (24.1%) children did not wear glasses, another 9 (1.6%) children wearing glasses were not clear, often rubbing the eyes of children were more likely to appear redness (20.4%), eye secretions (17.0%) and white spots in the pupil area (6.8%) and other adverse reactions. Conclusion: The intelligent voice follow-up system shows great application potential in clinical follow-up, which can be employed as a new service mode of intelligent medical treatment.