Emergency Medicine, Disease Surveillance, and Informatics

紧急医疗系统（EMS）机构作为医疗紧急情况的第一响应者，具有独特的地位，可以为监测和预警系统提供有价值的数据。在大多数州，所有进入EMS的患者都通过院前护理到急诊室，使用院前护理报告进行跟踪。聚合酶链反应用于收集重要的患者信息，卫生保健管理人员将这些信息用作通过宏观分析确定趋势的资源。目前，PCR数据主要以纸质形式存在，将这些数据输入可以处理和挖掘信息的数据库的过程可能需要长达数年的时间，因为没有从医院提取这些数据并将其传输到中心位置进行整理和分析的自动化手段。本研究的目的是通过文档图像理解和手写体分析来弥补这一不足。该研究团队致力于从具有松散约束布局结构、大量医学词汇、大量缩写和非标准写作风格的医疗表格中提取和识别手写文本数据的解决方案。如果成功的话，这项工作将推动目前仅限于小于一千个单词的小型“原始”词汇应用的手写识别技术的发展。从扫描表格中对PCR数据进行机器识别的任务，需要对数以万计的词汇表进行识别，其中还必须允许非标准缩写和医学术语。在这种情况下，研究的挑战将是使用表单上可以准确有效地识别的字段的线索，例如复选框数据，来约束叙事字段中单词的词汇量。尽管围绕防范可能的生物恐怖袭击的准备问题进行了宣传，但还没有提出许多短期到中期的实用技术解决方案来解决这个问题。拟议的项目将大大提高收集聚合酶链反应数据的速度，以填充国家紧急医疗服务数据库，从而为紧急医疗服务提供者和卫生保健管理人员提供丰富的数据，可用于趋势分析，以便及早发现疾病爆发，并为紧急服务分配公共资源。

Emergency Medical Systems (EMS) agencies, as first responders in medical emergency scenarios, are uniquely positioned to provide valuable data for surveillance and early warning systems. In most states, all patients who enter the EMS are tracked through their pre-hospital care to the emergency room using the Pre-Hospital Care Report. The PCR is used to gather vital patient information that is used by health care administrators as a resource to identify trends through macro analysis. Currently, the PCR data exists mostly on paper forms and the process of keying this data into a database that can be processed and mined for information can take up to several years as there is no automated means to extract this data from hospitals and transmit them to a central location for collating and analysis. The goal of this research is to attach this deficiency by using document image understanding and handwriting analysis. Intellectual Merit The research team work toward solutions to extraction and recognition of handwritten text data from medical forms with loosely constrained layout structure, large medical vocabulary, and abundance of abbreviations and non-standard writing styles. If successful, this work will advance the state-of-the-art in handwriting recognition limited to applications with small 'pristine' vocabularies of less than a thousand words. The proposed task of machine recognition of PCR data from scanned forms requires recognition with vocabularies of tens of thousands of words which must also allow for non-standard abbreviations and medical jargon. The research challenge in this case will be to use cues from fields on the form that can be recognized accurately and efficiently, such as check box data, to constrain the size of the vocabulary of the words in the narrative fields. Broader Impact Despite the publicity surrounding the issue of preparedness to guard against possible bioterrorist attacks, there have not been many short to middle-term practical technological solutions proposed to address the problem. The proposed project will greatly improve the speed of collecting PCR data to populate a national EMS database, thereby providing emergency medical service providers and health care administrators a wealth of data that can be used in trend analysis for the early detection of outbreak of diseases as well as allocation of public resources for emergency services.