Identification, Extraction and Display of Clinical Data Patterns with Application to Anesthesia Workflows

该项目的目标是使用数据科学/大数据社区的尖端方法来提供 复杂临床数据模式的快速可解释可视化，使临床医生能够快速回答 精选的临床问题，他们一天要面对很多次。传统的电子健康档案展示 表格数字数据和叙述性临床文本的格式使得识别复杂的关系变得困难 以及几个变量之间的趋势，特别是如果这些关系随着时间的推移而改变的话。我们有 以前开发的计算方法来识别临床上重要的、时间变化的关系和 医学数据之间的模式，在这个项目中，我们寻求扩展这些方法并产生定制的 可视化所发现的模式，以支持涵盖以下范围的特定临床任务 了解患者、程序和人群。 具体地说，我们寻求支持涉及回答以下广泛临床问题的认知任务 问：1)患者的术前临床情况如何？2)常见的麻醉药是什么？ 这种外科手术的方法是什么？和3)每个患者的视力水平和复杂性是什么 明天要做手术的人有多少？我们从临床上挑选了这些特定的问题 麻醉领域，因为这一领域在机构之间有相当一致的做法，但我们打算 我们的解决方案可以很容易地扩展到临床专业的类似问题。 该项目包括开发基于Web的工具，临床医生可以使用这些工具在 他们的日常临床实践。我们计划采用迭代开发方法，从定性的用户研究开始 与这三个问题相关的工作流程和信息需求，以及随后的设计迭代 包括最终用户临床医生在每次迭代时反馈。此外，我们将在设计时考虑是否可能 采用障碍，而不是将其留到部署时间，我们希望能够降低这些 具有适当设计决策的障碍。 如果成功，该项目将促进临床护理的日常实践，提高其效率， 效率和质量。

The goal of this project is to use cutting-edge methods from the data science/Big Data community to provide rapidly interpretable visualizations of complex clinical data patterns that allow clinicians to quickly answer selected clinical questions that they face many times a day. The traditional Electronic Health Record display formats of tabular numeric data and narrative clinical text make it difficult to identify complex relationships and trends among several variables at once, particularly if those relationships change over time. We have previously developed computational methods to identify clinically important, time-changing relationships and patterns among medical data, and in this project we seek to extend those methods and produce tailored visualizations of the discovered patterns to support specific clinical tasks that cover the spectrum of understanding patients, procedures, and populations. Specifically, we seek to support the cognitive tasks involved in answering the following broad clinical questions: 1) What is the preoperative clinical status of this patient? 2) What are the common anesthetic approaches for this surgical procedure? And 3) What is the acuity level and complexity of each patient in the population of those who will be operated on tomorrow? We selected these specific questions from the clinical domain of anesthesia because that domain has fairly consistent practices between institutions, but we intend for our solutions to be easily extendable to analogous questions across clinical specialties. This project includes developing web-based tools that clinicians can use to answer these questions during their daily clinical practice. We plan an iterative development approach, starting with qualitative user studies of workflows and information needs relevant to the three questions, and followed by design iterations that include end-user clinician feedback at each iteration. Additionally, we will consider at design time possible barriers to adoption, rather than leaving this until deployment time, and we expect to be able to lower those barriers with appropriate design decisions. If successful, this project will facilitate the daily practice of clinical care, increasing its efficiency, effectiveness, and quality.