Developing, Validating, and Implementing a CKD Predictive Model (DELVECKD)

描述（由申请人提供）：Khaled Abdel-Kader 博士已完成肾病学前期培训和医学教育硕士学位，并接受过成人学习、医疗错误和认知理论的正式培训以及临床研究和生物统计学的入门课程。他的主要研究兴趣是描述和解决初级保健机构中慢性肾脏病 (CKD) 护理的缺陷。他获得了个人博士后资助以支持他在该领域的工作。他的职业目标是成为 CKD 流行病学专家和独立临床研究者，研究基于电子病历 (EMR) 的干预措施，以改善 CKD 护理和结果。在这项职业发展奖 (CDA) 中，他专注于改善初级保健医生 (PCP) 的 CKD 筛查。该奖项为他提供了流行病学、研究设计、医学信息学、决策分析、卫生服务研究和生物统计学方面的指导、正式课程和实践经验。他组建了一批高素质的导师，他们将指导他并帮助他发展成为一名独立的临床研究者。支持性研究环境已经培养了众多成功的独立临床研究者，对这些个人导师起到了补充作用。此外，独特的机构资源，包括发达的电子病历、电子病历研究基础设施和庞大的患者基础，使当地环境成为候选人及其研究的理想环境。 Abdel-Kader 博士将利用这些研究经验、课程、指导、机构资源和承诺，继续发展成为一名独立资助的临床研究员。申请的一个重要组成部分是候选人的研究计划。他将分两个阶段进行他的研究项目。在第一阶段，他将利用当地完善的 EMR 数据库和匹兹堡大学的大型门诊患者库（过去 2 年超过 450,000 名独特患者）开发决策树预测模型，以在不使用血清化学的情况下识别 CKD 高风险患者。他将把决策树模型的性能与最近开发的著名 CKD 风险逻辑回归模型进行比较。在确定具有最佳性能的模型后，候选人将对 PCP 进行随机对照试验，检查在 EMR 中实施 CKD 预测模型作为临床警报与常规护理的效果。临床警报将提醒 PCP 对尚未筛查 CKD 高风险患者进行筛查。这种新颖的方法将 CKD 风险因素的机器建模与 EMR 临床决策支持系统 (CDSS) 结合起来，为 PCP 提供实时指导，从而改善对隐匿性 CKD 患者的护理。该项目将为申请人提供数据挖掘、医学信息学、决策分析、卫生服务研究以及临床试验设计和实施方面的宝贵经验。这些经历对于他成为一名独立的临床研究员至关重要。除了这些直接经验之外，申请人还将受益于其熟练的导师和顾问团队提供的教学和指导。该提案的主要导师 Mark Unruh 博士是一位资金雄厚的独立临床研究者，拥有流行病学、临床试验和 CKD 方面的专业知识。匹兹堡大学公共卫生研究生院卫生政策与管理系主任马克·罗伯茨 (Mark Roberts) 博士在预测模型、决策树分析和 CKD 方面拥有良好的独立资助记录和良好的研究兴趣。 Shyam Visweswaran 博士是一名生物医学信息学研究员，带来了生物医学数据挖掘、预测建模和 CDSS 方面的专业知识。 Charity Moore 博士是一位技术精湛的卫生服务统计学家，在临床试验方面拥有丰富的经验。她将把她在研究设计、实施、分析和解释方面的专业知识带入该项目。 Gary Fischer 博士是普通内科门诊主任，在将 EMR 和 CDSS 与医生工作流程集成方面拥有丰富的经验。 Douglas Landsittel 博士是一位统计学家，对使用决策树对疾病结果进行分类感兴趣，在构建和验证预测模型方面拥有丰富的经验。这个跨学科团队将独特的合格研究人员与成功完成拟议研究和候选人培训所需的多样化经验和专业知识结合起来。为了补充这些实践经验和指导活动，候选人将通过匹兹堡大学公共卫生研究生院、生物医学信息学系和临床研究教育研究所（该大学临床和转化科学研究所的一部分）完成正式课程。这些课程将包括研究设计和临床试验实施、应用医学信息学和决策分析以及卫生服务研究和生物统计学方法方面的正式培训。此外，医疗中心还举办许多研讨会、讲习班和领导力课程，候选人将参加这些课程以建立合作关系并提高其技能。总之，候选人对提高 CKD 护理服务质量的研究兴趣，加上良好的医学教育背景和临床研究的早期培训，使他成为使用此 CDA 来研究 EMR 干预措施的理想候选人，这些干预措施可以广泛改善 PCP 的 CKD 筛查。申请人经验丰富的多学科指导团队、强大的机构资源和支持，以及他将在该奖项下完成的正式培训，将确保他继续发展成为一名成功的独立临床研究者，研究方法以改善对 CKD 患者的 PCP 护理服务。

DESCRIPTION (provided by applicant): Dr. Khaled Abdel-Kader has completed prior training in nephrology and a master's in medical education with formal training in adult learning, medical errors, and cognitive theory as well as introductory coursework in clinical research and biostatistics. His primary research interest is characterizing and addressing chronic kidney disease (CKD) care deficiencies in the primary care setting. He has received individual post-doctoral funding to support his work in this area. His career goal is to become an expert in CKD epidemiology and an independent clinical investigator studying electronic medical record (EMR)-based interventions to improve CKD care and outcomes. In this career development award (CDA), he focuses on improving primary care physician (PCP) screening for CKD. This award provides him with mentorship, formal coursework, and hands-on experience in epidemiology, research design, medical informatics, decision analysis, health services research, and biostatistics. He has assembled a group of highly skilled mentors who will guide him and help him develop into an independent clinical investigator. A supportive research environment that has already cultivated the development of numerous successful independent clinical investigators complements these individual mentors. In addition, unique institutional resources including a well-developed EMR, EMR research infrastructure, and large patient base make the local environment an ideal setting for the candidate and his research. Dr. Abdel- Kader will use these research experiences, coursework, mentorship, and institutional resources and commitment to continue his progression to becoming an independently funded clinical researcher. An important element of the application is the candidate's research proposal. He will conduct his research project in 2 phases. In the first phase, he will leverage the local, well-developed EMR database and the University of Pittsburgh's large ambulatory patient base (>450,000 unique patients in the prior 2 years) to develop a decision tree predictive model to identify patients at high risk for CKD without the use of serum chemistries. He will compare the performance of the decision tree model to a prominent, recently developed logistic regression model of CKD risk. After identifying the model with the best performance, the candidate will conduct a randomized controlled trial of PCPs examining the effect of implementing the CKD predictive model in the EMR as a clinical alert versus usual care. The clinical alert will remind PCPs to screen high-risk patients for CKD if they have not already done so. This novel approach pairs machine modeling of CKD risk factors with an EMR clinical decision support system (CDSS) to provide real-time guidance to PCPs to improve the care delivered to patients with occult CKD. This project will provide the applicant with valuable experience in data mining, medical informatics, decision analysis, health services research, and clinical trial design and implementation. These experiences will be integral to his development into an independent clinical researcher. In addition to these direct experiences, the applicant will also benefit from the teaching and guidance provided by his team of proficient mentors and consultants. Dr. Mark Unruh, primary mentor for the proposal, is a well-funded, independent clinical investigator who brings expertise in epidemiology, clinical trials, and CKD. Dr. Mark Roberts, Chair of Health Policy and Management at the University of Pittsburgh's Graduate School of Public Health, brings a strong record of independent funding and well-established research interests in predictive modeling, decision tree analysis, and CKD. Dr. Shyam Visweswaran, an investigator in biomedical informatics, brings expertise in biomedical data mining, predictive modeling, and CDSS. Dr. Charity Moore is a highly skilled health services statistician with extensive experience in clinical trials. She will bring her expertise in research design, implementation, analysis, and interpretation to the project. Dr. Gary Fischer, director of the general internal medicine ambulatory clinic, has substantial experience in the integration of the EMR and CDSS with physician workflow. Dr. Douglas Landsittel, a statistician with an interest in the classification of disease outcomes using decision trees, has extensive experience in building and validating predictive models. This interdisciplinary team combines uniquely qualified investigators with the diversity of experience and expertise necessary for the successful completion of the proposed research and the candidate's training. To complement these hands-on experiences and mentorship activities, the candidate will undertake formal coursework through the University of Pittsburgh's Graduate School of Public Health, Department of Biomedical Informatics, and the Institute for Clinical Research Education (part of the university's Clinical and Translational Science Institute). These courses will include formal training in research design and clinical trial implementation, applied medical informatics and decision analysis, and methods in health services research and biostatistics. In addition, the medical center has numerous seminars, workshops, and leadership courses that the candidate will participate in to form collaborative relationships and enhance his skills. In summary, the candidate's research interest in improving the quality of CKD care delivery coupled with a sound background in medical education and early training in clinical research make him an ideal candidate to use this CDA to investigate EMR interventions that can broadly improve CKD screening by PCPs. The applicant's experienced, multidisciplinary mentorship team, strong institutional resources and support, and the formal training he will complete under this award will ensure that he continues to develop into a successful independent clinical investigator studying methods to improve PCP care delivery to CKD patients.