利用电子健康记录与机器学习识别假性痛风
本研究基于1991—2017年电子健康记录数据,纳入至少1个相关计费码或至少2次假性痛风/软骨钙化症自然语言处理提及的患者,抽取900例进行金标准病历复核,其中123例(13.7%)被判定为假性痛风(68例确诊、55例可能)。研究比较计费码、焦磷酸钙晶体、主题建模及联合算法的表现,联合算法将阳性预测值从计费码的22%提升至81%。
要点
- 计费码识别假性痛风敏感性65%、阳性预测值仅22%,单独使用不可靠。
- 联合自然语言处理、主题建模与滑液晶体结果可将阳性预测值提升至81%。
- 联合算法识别患者数比单用晶体结果多50%,并弥补了可能假性痛风的漏检。
核心结论
- 对无特异性计费码的发作性假性痛风,联合自然语言处理、机器学习与滑液实验室结果可显著提高阳性预测值。
研究设计与样本
- 研究构建1991—2017年电子健康记录数据仓,纳入至少1个相关计费码或至少2次自然语言处理提及假性痛风或软骨钙化症的患者。
- 从中选取900例进行金标准病历复核,分为确诊假性痛风(滑膜炎+滑液焦磷酸钙晶体)、可能假性痛风(滑膜炎+软骨钙化症)与无假性痛风。
- 采用主题建模方法识别确诊/可能假性痛风,并构建联合主题建模与人工复核焦磷酸钙晶体结果的算法。
- 比较计费码、焦磷酸钙晶体存在、主题建模与联合算法的性能及识别队列。
主要发现
- 900例中123例(13.7%)经病历复核为假性痛风,其中确诊68例、可能55例。
- 计费码识别假性痛风的敏感性为65%,阳性预测值为22%。
- 焦磷酸钙晶体存在的敏感性为29%,阳性预测值为92%。
- 不使用晶体结果时主题建模的敏感性为29%,阳性预测值为79%。
- 联合算法敏感性为42%、阳性预测值为81%,识别患者数比单用晶体结果多50%。
关键数值与推荐
- 计费码:敏感性65%,阳性预测值22%。
- 焦磷酸钙晶体存在:敏感性29%,阳性预测值92%。
- 主题建模(不用晶体结果):敏感性29%,阳性预测值79%。
- 联合算法:敏感性42%,阳性预测值81%。
- 假性痛风患病比例:900例中123例(13.7%),确诊68例、可能55例。
临床怎么用
- 在缺乏特异性计费码的假性痛风研究中,可考虑联合自然语言处理、主题建模与滑液晶体结果构建识别算法。
- 单用计费码阳性预测值低,不宜作为假性痛风病例识别的唯一依据。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVE: Identifying pseudogout in large data sets is difficult due to its episodic nature and a lack of billing codes specific to this acute subtype of calcium pyrophosphate (CPP) deposition disease. The objective of this study was to evaluate a novel machine learning approach for classifying pseudogout using electronic health record (EHR) data. METHODS: We created an EHR data mart of patients with ≥1 relevant billing code or ≥2 natural language processing (NLP) mentions of pseudogout or chondrocalcinosis, 1991-2017. We selected 900 subjects for gold standard chart review for definite pseudogout (synovitis + synovial fluid CPP crystals), probable pseudogout (synovitis + chondrocalcinosis), or not pseudogout. We applied a topic modeling approach to identify definite/probable pseudogout. A combined algorithm included topic modeling plus manually reviewed CPP crystal results. We compared algorithm performance and cohorts identified by billing codes, the presence of CPP crystals, topic modeling, and a combined algorithm. RESULTS: Among 900 subjects, 123 (13.7%) had pseudogout by chart review (68 definite, 55 probable). Billing codes had a sensitivity of 65% and a positive predictive value (PPV) of 22% for pseudogout. The presence of CPP crystals had a sensitivity of 29% and a PPV of 92%. Without using CPP crystal results, topic modeling had a sensitivity of 29% and a PPV of 79%. The combined algorithm yielded a sensitivity of 42% and a PPV of 81%. The combined algorithm identified 50% more patients than the presence of CPP crystals; the latter captured a portion of definite pseudogout and missed probable pseudogout. CONCLUSION: For pseudogout, an episodic disease with no specific billing code, combining NLP, machine learning methods, and synovial fluid laboratory results yielded an algorithm that significantly boosted the PPV compared to billing codes.
引用
- Tedeschi SK, Cai T, He Z, et al. Classifying Pseudogout Using Machine Learning Approaches With Electronic Health Record Data. Arthritis Care Res (Hoboken). 2021. PMID: 31910317. doi:10.1002/acr.24132
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。