Sepsis phenotypes at risk for infections caused by multidrug resistant Gram-negative bacilli: elucidating the impact of sepsis definition and patient case mix on prediction performance

脓毒症表型面临由多重耐药革兰氏阴性杆菌引起的感染风险：阐明脓毒症定义和患者病例组合对预测性能的影响

• 批准号: 10256063
• 负责人: Maria Cristina Vazquez Guillamet
• 金额: $ 16.55万
• 依托单位: WASHINGTON UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-10 至 2024-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10256063
• 关键词: American; Antibiotic Therapy; Antibiotics; Antimicrobial Resistance; Bacillus; Big Data; Case Mixes; Cause of Death; Characteristics; Clinical; Clinical Data; Clinical Informatics; Clinical Medicine; Clinical Research; Communicable Diseases; Communication; Critical Care; Critical Illness; Data; Development; Development Plans; Empiricism; Environment; Etiology; Excess Mortality; Frequencies; Functional disorder; Funding; Goals; Health Care Costs; Healthcare; Healthcare Systems; Hospitals; Infection; Informatics; Institution; Leadership; Learning; Life; Linear Regressions; Link; Mathematics; Medical; Medicine; Mentors; Mentorship; Methodology; Methods; Modeling; Morbidity - disease rate; Multi-Drug Resistance; National Institute of General Medical Sciences; Organ; Outcome; Patients; Performance; Phenotype; Physicians; Policy Maker; Population; Prevalence; Probability; Reproducibility; Research; Research Personnel; Resistance; Risk; Risk Estimate; Role; Sepsis; Standardization; Surveys; Syndrome; Testing; Theoretical model; Time; Training; Uncertainty; United States; Universities; Washington; Work; antimicrobial; attributable mortality; base; big data management; career; career development; clinical care; clinical database; clinical decision support; clinical epidemiology; cohort; combat; experience; hands on research; high risk; improved; individual patient; infection risk; innovation; interest; machine learning method; mortality; mortality risk; multi-drug resistant pathogen; multilevel analysis; novel; personalized risk prediction; predictive modeling; predictive tools; primary outcome; response; responsible research conduct; risk prediction; risk prediction model; risk stratification; septic patients; skills; supervised learning; support tools; tool; user-friendly

PROJECT ABSTRACT Sepsis is a devastating syndrome that represents a leading cause of death, morbidity, and healthcare costs. Its impact is amplified by rising rates of antimicrobial resistance. Improving sepsis outcomes primarily results from prescribing timely antibiotics based on the estimated risk of multidrug resistance (MDR). Previous models grossly overestimated the MDR risk and exacerbated the escalating rates of antimicrobial resistance and excess mortality. The overall goal of this proposed K08 research is to identify common sepsis phenotypes that will enable better prescribing practices and standardized comparisons across hospitals, which will help practicing clinicians, researchers, healthcare institutions, and policy makers. These themes correlate with NIGMS's interest in finding innovative methods and leveraging big data to improve sepsis outcomes. Our three specific aims reflect these goals: (1) establish resistance thresholds for MDR Gram- negative bacilli (GNB) that cause sepsis, (2) assess the impact of sepsis definition on the performance of risk prediction models for MDR GNB, and (3) identify sepsis phenotypes at high risk for MDR GNB in a well-balanced cohort and assess the impact of case mix on risk prediction model performance. We will mathematically derive resistance thresholds that link population resistance rates to individual patient risk of death in sepsis caused by MDR GNB, assess factors that impact prediction performance, and incorporate rich clinical data from 15 hospitals in our healthcare system to identify stable common sepsis phenotypes. Dr. Vazquez Guillamet has training in Infectious Diseases and Critical Care Medicine and experience in antimicrobial resistance in critically ill patients. This proposal will build on her clinical work and previous research experience in finding innovative methods to solve challenging problems at the intersection of infectious diseases and critical care medicine. Dr. Vazquez Guillamet has six career objectives: (1) pursue advanced training in clinical epidemiology; (2) acquire skills in advanced linear regression and multilevel modeling; (3) learn supervised machine learning methods; (4) acquire skills in big data management in healthcare and methods to handle missing data; (5) improve scientific communication, grantsmanship, and leadership, and (6) participate in training in the responsible conduct of research. She will achieve these goals through didactic coursework, hands-on research experience, and active mentoring from experts in Infectious Diseases, Critical Care Medicine, and applied clinical informatics. She will continue to develop innovative methods to mitigate the antimicrobial resistance crisis, especially in critically ill patients, and become an analytics translator at the intersection of clinical medicine and clinical applied informatics. The fertile research environment at Washington University in St. Louis, the experienced mentorship team, and a well-crafted career development plan will enable Dr. Vazquez Guillamet to achieve her long-term goal of becoming an independently funded clinician-investigator utilizing big data to develop applications for risk prediction, surveillance, and outcome comparisons in antimicrobial resistance and sepsis.

项目摘要 败血症是一种毁灭性的综合征，是导致死亡、发病率和医疗费用的主要原因。它的 抗菌素耐药率的上升放大了影响。改善脓毒症预后的主要结果是 根据多药耐药(MDR)的估计风险，及时开出抗生素。以前的型号 严重高估了多药耐药风险，加剧了抗菌素耐药性的上升和过剩 死亡率。这项拟议的K08研究的总体目标是确定常见的脓毒症表型， 将使更好的处方实践和跨医院的标准化比较成为可能，这将有助于 执业临床医生、研究人员、医疗机构和政策制定者。这些主题相互关联 随着NIGMS对寻找创新方法和利用大数据改善脓毒症的兴趣 结果。我们的三个具体目标反映了这些目标：(1)建立耐多药Gram的耐药阈值- 引起败血症的阴性杆菌(GNB)，(2)评估败血症定义对风险表现的影响 MDR GNB的预测模型，以及(3)在平衡良好的情况下确定MDR GNB的高危脓毒症表型 对病例组合对风险预测模型性能的影响进行分类和评估。我们将从数学上推导出 将人群耐药率与由以下原因引起的脓毒症患者死亡风险联系起来的耐药阈值 MDR GNB，评估影响预测性能的因素，并纳入来自15家医院的丰富临床数据 在我们的医疗系统中，以确定稳定的常见败血症表型。Vazquez Guillamet博士接受了 传染病和危重病医学与危重病患者的抗菌药耐药经验。 这项建议将建立在她的临床工作和之前的研究经验的基础上，寻找创新的方法来 解决传染病和重症监护医学交叉领域的挑战性问题。巴斯克斯博士 Guillamet有六个职业目标：(1)进行临床流行病学方面的高级培训；(2)获得 高级线性回归和多水平建模；(3)学习有监督机器学习方法；(4)获取 医疗保健大数据管理技能和缺失数据处理方法；(5)提高科学性 沟通、勇气和领导力，以及(6)参加有关负责任的行为的培训 研究。她将通过讲授课程、实践研究经验和积极的行动来实现这些目标 来自传染病、重症监护医学和应用临床信息学专家的指导。她会的 继续开发创新方法以缓解抗菌素耐药性危机，特别是在危重病人 患者，成为临床医学和临床应用交汇点的分析翻译者 信息学。圣路易斯华盛顿大学肥沃的研究环境，经验丰富的导师 团队，一个精心设计的职业发展计划将使巴斯克斯·吉拉梅博士能够实现她的长期 目标是成为独立出资的临床医生和研究人员，利用大数据开发风险应用程序 抗菌素耐药性和败血症的预测、监测和结果比较。