Improving data capture in clinical research using a chatbot

项目摘要。 直接从研究参与者那里收集完整和准确的结果数据变得越来越多 重要.临床研究人员需要一种成本效益高的方法来捕获高质量的患者报告 结果。通常，直接从参与者处获取的数据是通过自填问卷或 通过人工面试，每个人都有自己的优点和缺点。有效的新数据 捕获技术，可以通过人工访谈收集患者报告的结果， 自我管理调查的费用将为临床研究建立巨大的能力。Dokbot，LLC和 南卡罗来纳州医科大学（MUSC）合作开发了Dokbot，一个简单、可扩展的 使用基于文本的对话从临床研究参与者收集数据的聊天机器人， 他们的移动的设备上的浏览器。聊天机器人是一种创新和有效的方式来捕获临床数据 research.不幸的是，当前的聊天机器人技术不足以支持临床应用中的数据捕获。 research. Dokbot可用于增强临床研究中的数据采集。然而， 需要适应、改进和完善来扩展和优化Dokbot，使其能够理想地支持 临床研究为了实现这一目标，我们首先需要了解临床研究中的机会和障碍。 研究利益相关者使用Dokbot（目标1），然后适应和迭代完善的功能原型， 用于临床研究的Dokbot（目标2）。通过展示Dokbot作为一种简单、低成本 在临床研究环境中收集数据的方法，我们将有一个明确的路径来开发技术， 专业知识和证据，对改善临床研究数据收集产生重大影响。与 通过STTR奖的支持，Dokbot可以成为一个有效的工具，帮助临床研究人员提高 研究参与者数据的质量和效率

PROJECT SUMMARY . Collecting complete and accurate outcome data directly from research participants is becoming increasingly important. Clinical researchers needs a cost-effective approach to capture high-quality patient-reported outcomes. Typically, data captured directly from participants is through self-administered questionnaires or through a human interviewer, each with their own advantages and disadvantages. An effective new data capture technology that can collect patient-reported outcomes with the engagement of human interviews at the cost of self-administered surveys would build tremendous capacity for clinical research. Dokbot, LLC and the Medical University of South Carolina (MUSC) have partnered to develop Dokbot, a simple, scalable chatbot that uses text-based conversations to collect data from clinical research participants using the browser on their mobile devices. Chatbots are an innovative and effective way to capture data for clinical research. Unfortunately, current chatbot technologies do not adequately support data capture in clinical research. Dokbot can be adapted to enhance data capture in clinical research. However, significant adaptation, improvement, and refinement is needed to extend and optimize Dokbot for it to ideally support clinical research. To achieve this, we first need to understand opportunities and barriers among clinical research stakeholders using Dokbot (Aim 1) and then adapt and iteratively refine a functional prototype of Dokbot for clinical research (Aim 2). By demonstrating the feasibility of Dokbot as a simple, low-cost approach for collecting data in clinical research settings, we will have a clear path to develop the technology, expertise, and evidence to make a significant impact on improving clinical data collection for research. With support through the STTR award, Dokbot could become an effective tool to help clinical researchers improve the quality and efficiency of data from research participants