Development of clinical decision tools for management of diarrhea of children in high and low resource settings

摘要 腹泻病是全世界儿童死亡的主要原因之一，大多数发生在 低收入国家。在高收入国家，儿科腹泻仍然是医疗保健的主要用途。 资源。治疗腹泻的基础是补液，尽管抗菌药物对腹泻有好处 一些例子。不幸的是，鉴于治疗通常是经验性的，主要基于临床怀疑 细菌原因、抗菌剂在全世界的腹泻疾病管理中被过度使用。在高收入人群中 在许多国家，诊断性检测往往被过度使用。因此，需要临床决策支持工具 用于许多环境中的抗菌和诊断管理。目前的临床预测工具大多基于 根据患者的内在属性，如临床检查和特定于该患者的症状史。我们有 初步数据表明，患者-外部数据的整合，包括气候和季节性 参数和人群水平的测试前概率(来自以前的患者和前几年的流行率)可以 改进临床预测模型的性能。我们也有初步数据表明， 一种评估腹泻病因以减少抗生素的电子临床决策支持工具(ECDST) 处方费。我们的总体目标是：1)通过整合 患者-外部数据源，以及2)探索eCDST的潜在可行性和用途，例如 智能手机应用程序或电子健康记录工具。在目标1中，我们将使用几个预期数据 儿童腹泻的临床研究以建立包含患者-外在因素的改进的临床预测模型 数据源。在目标2中，我们将确定eCDST的潜在可行性、实用性和经济影响 抗生素和诊断管理，通过深入检查临床和护理人员的视角 访谈和焦点小组。我们还将在美国环境下对eCDST进行经济评估。 这些目标的完成将导致使用大数据的优化临床预测模型，并为 为设计用于儿科管理的eCDST实施研究提供信息所需的基础工作 拉肚子。

Abstract Diarrheal diseases are the among the leading cause of death in children worldwide, most of which occur in low-income countries. In high-income countries, pediatric diarrhea remains a major utilization of healthcare resources. The cornerstone for management of diarrhea is rehydration, though antimicrobials are beneficial in some instances. Unfortunately, given that treatment is frequently empiric, based mostly on clinical suspicion for bacterial causes, antimicrobials are overused in management of diarrheal illness worldwide. In high-income countries, diagnostic testing is oftentimes overutilized. Thus, there is a need for clinical decision support tools for antimicrobial and diagnostic stewardship in many settings. Current clinical prediction tools are based mostly on patient-intrinsic properties such as the clinical exam and symptom history specific for that patient. We have preliminary data suggesting that the integration of patient-extrinsic data, including climate and seasonality parameters, and population-level pre-test probabilities (from prior patients and prior years’ prevalence), can improve the performance of a clinical prediction model. We also have preliminary data showing the potential for an electronic clinical decision-support tool (eCDST) that estimates diarrheal etiology to decrease antibiotic prescription rates. Our overarching goal is to: 1) improve diarrhea clinical prediction through integration of patient-extrinsic data sources, and 2) explore the potential feasibility and utility of an eCDST, such as a smartphone application or an electronic health record tool. In Aim 1, we will use data from several prospective clinical studies of pediatric diarrhea to build improved clinical prediction models that includes patient-extrinsic data sources. In Aim 2, we will determine the potential feasibility, utility, and economic impact of an eCDST for antibiotic and diagnostic stewardship by examining clinican and caregiver perspectives through in-depth interviews and focus groups. We will also perform an economic evaluation of eCDST in a US setting. Completion of the Aims will result in an optimized clinical prediction model using big data and lay the groundwork needed to inform the design of implementation studies of eCDSTs for management of pediatric diarrhea.