Howson Stephanie N, McShea Michael J, Ramachandran Raghav, Burkom Howard S, Chang Hsien-Yen, Weiner Jonathan P, Kharrazi Hadi
Applied Physics Laboratory, Johns Hopkins University, Baltimore, MD, United States.
Center for Population Health Information Technology, Johns Hopkins School of Public Health, Baltimore, MD, United States.
JMIR Med Inform. 2022 Mar 24;10(3):e33212. doi: 10.2196/33212.
A small proportion of high-need patients persistently use the bulk of health care services and incur disproportionate costs. Population health management (PHM) programs often refer to these patients as persistent high utilizers (PHUs). Accurate PHU prediction enables PHM programs to better align scarce health care resources with high-need PHUs while generally improving outcomes. While prior research in PHU prediction has shown promise, traditional regression methods used in these studies have yielded limited accuracy.
We are seeking to improve PHU predictions with an ensemble approach in a retrospective observational study design using insurance claim records.
We defined a PHU as a patient with health care costs in the top 20% of all patients for 4 consecutive 6-month periods. We used 2013 claims data to predict PHU status in next 24 months. Our study population included 165,595 patients in the Johns Hopkins Health Care plan, with 8359 (5.1%) patients identified as PHUs in 2014 and 2015. We assessed the performance of several standalone machine learning methods and then an ensemble approach combining multiple models.
The candidate ensemble with complement naïve Bayes and random forest layers produced increased sensitivity and positive predictive value (PPV; 49.0% and 50.3%, respectively) compared to logistic regression (46.8% and 46.1%, respectively).
Our results suggest that ensemble machine learning can improve prediction of care management needs. Improved PPV implies reduced incorrect referral of low-risk patients. With the improved sensitivity/PPV balance of this approach, resources may be directed more efficiently to patients needing them most.
一小部分高需求患者持续占用大部分医疗服务并产生不成比例的费用。人群健康管理（PHM）项目通常将这些患者称为持续高利用者（PHU）。准确的PHU预测能够使PHM项目更好地将稀缺的医疗资源与高需求的PHU相匹配，同时总体上改善治疗效果。虽然之前关于PHU预测的研究已显示出前景，但这些研究中使用的传统回归方法准确性有限。
我们试图在一项使用保险理赔记录的回顾性观察性研究设计中，通过集成方法改进PHU预测。
我们将PHU定义为在连续4个6个月期间医疗费用处于所有患者前20%的患者。我们使用2013年的理赔数据来预测未来24个月的PHU状态。我们的研究人群包括约翰霍普金斯医疗保健计划中的165,595名患者，其中8359名（5.1%）患者在2014年和2015年被确定为PHU。我们评估了几种独立机器学习方法的性能，然后评估了一种结合多个模型的集成方法。
与逻辑回归（分别为46.8%和46.1%）相比，具有互补朴素贝叶斯和随机森林层的候选集成模型产生了更高的敏感性和阳性预测值（PPV；分别为49.0%和50.3%）。
我们的结果表明，集成机器学习可以改善对护理管理需求的预测。PPV的提高意味着低风险患者的错误转诊减少。随着这种方法的敏感性/PPV平衡得到改善，资源可以更有效地导向最需要它们的患者。
