SBIR Phase I: Improving Fertility Patient-Provider Communication with a Personalized Natural Language Processing Platform

小型企业创新研究(SBIR)第一阶段项目的更广泛影响/商业潜力是开发自然语言处理软件，该软件将简化生育诊所患者及其提供者之间的沟通。生育诊所是专门的医疗机构，提供临床诊断和治疗，以帮助怀孕。他们的患者在怀孕过程中经历了强烈的焦虑和巨额的自付费用，在生育诊所，多达三分之二的患者将停止治疗方案，情绪压力被认为是主要原因。患者满意度比仅仅实现预期结果更复杂，这与护理团队如何有效地驾驭巨大的信息差距和弥合沟通差距相关。该项目推进了促进这种交流的技术。这个小型企业创新研究(SBIR)第一阶段项目将解决与医疗保健中的自然语言处理相关的主要技术障碍，特别是生育护理。这些挑战包括，会话代理必须理解与生育相关的问题所特有的词汇，自然语言平台必须克服在连续的多人对话中保持上下文感知的困难，以及在更广泛的对话上下文中理解模棱两可的问题，以及算法必须以针对特定患者的临床准确响应进行回复。为了实现这些目标，将根据生育利益相关者(患者、提供者和诊所工作人员)的反馈开发特定于生育的语料库或词典，并将开发一种机器学习算法，分析患者信息的意图，并从开发的语料库中提取除外部来源外的个性化回应。最后，将进行可用性研究，以改进平台的用户界面、语气和信息内容。拟议研究的预期结果是一个提高患者与护理团队沟通质量和效率的平台。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact /commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to develop natural language processing software that will streamline communication between fertility clinic patients and their providers. Fertility clinics are specialized medical facilities that offer clinical diagnoses and treatments to assist with the conception of a child. Their patients undergo intense anxiety and large out-of-pocket expenses throughout their journey to conceive, and up to two-thirds of patients at fertility clinics will discontinue their treatment regimen with emotional stress being cited as the primary reason. Patient satisfaction, which is more complex than just achieving desired outcomes, is correlated with how effectively care teams can navigate the large information disparity and bridge the communication gap. This project advances technologies to facilitate this communication. This Small Business Innovation Research (SBIR) Phase I project will address major technical hurdles associated with natural language processing in healthcare, specifically fertility care. These challenges include that a conversational agent must understand the vocabulary particular to fertility-related questions, the natural language platform must overcome the difficulty in maintaining contextual-awareness throughout a continuous multi-person dialogue as well as understand ambiguous questions within a broader conversational context, and that the algorithm must reply with clinically accurate responses personalized to a specific patient. To meet these objectives, a fertility-specific corpus, or dictionary, will be developed from fertility stakeholders (patients, providers, and clinic staff) feedback, and a machine-learning algorithm will be developed that analyzes patient messages for intent and provides personalized responses drawing from the developed corpus in addition to external sources. Finally, a usability study will be performed to refine the platform’s user-interface, tone, and informational content. The anticipated result of the proposed research is a platform that improves both the quality and efficiency of patient-care team communication.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.