Natural history, risk prediction and cost of cirrhosis in insured Americans.

摘要 在美国，肝硬变是导致死亡的主要原因，在150-940万美国人中确诊， 每年导致超过40,000人死亡。每年有5-7%的肝硬变患者会有生命危险 失代偿事件，如腹水、肝性脑病(HE)、胃肠道出血(GIB)或 发展为肝细胞癌(HCC)。这些事件往往会导致住院、残疾甚至死亡。 挑战是准确地预测那些可能发生失代偿事件的患者 并且很可能会死亡。对肝硬变患者进行准确的风险分层将有助于及早识别 以及指南建议护理和新兴疗法(例如他汀类药物)的优先顺序。几个 预测模型是存在的，但没有一个模型能充分回答这个问题。此外，没有纵向成本 在美国已经对肝硬变患者进行了护理分析和成本预测。关爱 对肝硬变患者的治疗是复杂的，通常涉及昂贵的反复住院和手术。2015年， 据报道，仅医院费用就高达163亿美元。我们提议的研究将分析一个大型国家 行政健康支付者，提供有关诊断、程序、实验室测试、药物、 2011年至2018年期间的住院和门诊护理，以及参保美国人的标准化费用。这样的一个 包括成本数据在内的大量基于人群的肝硬变队列提供了一个独特的、前所未有的机会 研究疾病进展，开发高度准确的预测模型，并研究成本。 目的1.在一大群被保险人的纵向队列中描述随时间推移的肝硬变的自然病史 在美国患有肝硬变的美国人 目标1.1：数据准备和变量转换 目标1.2：由肝硬变利益相关者小组裁决潜在的风险相关协变量 目标1.3：描述肝硬变的自然病程 目的2.预测肝硬变患者失代偿、住院和死亡的风险 大型纵向管理数据集(联合健康集团) 目标2.1：建立失代偿风险模型 目标2.2：建立住院风险模型 目标2.3：将UHG数据集与国家死亡指数合并，并对死亡风险进行建模 目标3：预测与肝硬变患者各方面护理相关的费用 目标3.1：确定按肝硬变状态/表型分层的标准化成本 目标3.2：建立随时间推移的护理成本模型

SUMMARY Cirrhosis is a leading cause of mortality in the United States (US), diagnosed in 1.5-9.4 million Americans and results in over 40,000 deaths each year. Every year 5-7% patients with cirrhosis experience life-threatening decompensating events, such as ascites, hepatic encephalopathy (HE), gastrointestinal bleeding (GIB), or develop hepatocellular carcinoma (HCC). These events often result in hospitalization, disability, or even death. The challenge is to accurately predict those patients who are likely to develop decompensating events and are likely to die. Accurate risk stratification of persons with cirrhosis will allow for early identification and prioritization for guideline recommended care and emerging therapies (e.g., statins). Several predictive models exist but none of them adequately answers this question. Furthermore, no longitudinal cost of care analyses and cost prediction has been performed in the US for persons with cirrhosis. The care of patients with cirrhosis is complex, often involving costly recurrent hospitalizations and procedures. In 2015, the hospital costs alone were reported to be $16.3 billion. Our proposed research will analyze a large national administrative health payer with detailed information on diagnoses, procedures, laboratory tests, medications, in- and outpatient care, as well as standardized costs for insured Americans between 2011 and 2018. Such a large population-based cirrhosis cohort, which includes cost data, offers a unique and unprecedented opportunity to study disease progression, develop highly accurate prediction models, and study costs. Aim 1. To describe the natural history of cirrhosis over time in a large longitudinal cohort of insured Americans with liver cirrhosis in the United States Aim 1.1: Data preparation and variable transformation Aim 1.2: Adjudicate potentially risk-relevant covariates by a cirrhosis stakeholder panel Aim 1.3: Describe the natural history of cirrhosis Aim 2. To predict the risk of decompensation, hospitalization and death in patients with cirrhosis using a large longitudinal administrative dataset (UNITED Health Group) Aim 2.1: Model the risk of decompensation Aim 2.2: Model the risk of hospitalization Aim 2.3: Merge the UHG dataset with the National Death Index and model the risk of death Aim 3: To predict costs associated with all aspects of care in patients with liver cirrhosis Aim 3.1: Ascertain standardized cost stratified by state/phenotypes of liver cirrhosis Aim 3.2: Model the cost of care over time