Shah Shailja C, Gupta Rohan, Bustamante Ranier, Lamm Mark, Yassin Hanin, Earles Ashley, Hung Adriana, Halvorson Alese, Greevy Robert, Gupta Samir, Demb Joshua, Liu Lin, Roumie Christianne L
Gastroenterology Section, Jennifer Moreno VA San Diego Healthcare System, San Diego, California.
Division of Gastroenterology, University of California San Diego, San Diego, California.
Gastro Hep Adv. 2023 Sep 13;3(1):78-83. doi: 10.1016/j.gastha.2023.09.005. eCollection 2024.
There are limited contemporary population-based data on epidemiology and outcomes in the United States. Our primary aim was to create a validated cohort of veterans with testing or treatment using Veterans Health Administration data.
Using Veterans Health Administration structured and unstructured data, we developed and validated 4 algorithms for infection (3 algorithms) and treatment status (1 algorithm). During the development phase, we iteratively modified each algorithm based on a manual review of random sets of electronic health records (reference standard). The validation goal was to achieve a one-sided 95% confidence lower bound (LB) for positive predictive value (PPV) and/or negative predictive value (NPV) >90%. We applied the Bonferroni correction when both PPV and NPV were relevant.
For infection, we achieved 99.0% PPV (LB = 94.6%) and 100% NPV (LB = 96.4%) for discriminating positive vs negative status using structured (ie, laboratory tests) and 95% PPV (LB = 90.3%) and 97.9% NPV (LB = 93.9%) using unstructured (ie, histopathology reports) data. Diagnostic codes achieved 98% PPV (LB = 93.0%) for diagnosis. The treatment algorithm was composed of multiple antimicrobial combinations and overall achieved ≥98% PPV (LB = 93.0%) for treatment, except for amoxicillin/levofloxacin (PPV<60%). Application of these algorithms yielded nearly 1.2 million veterans with testing and/or treatment between 1999 and 2018.
We assembled a validated national cohort of veterans who were tested or treated for infection. This cohort can be used for evaluating epidemiology and treatment patterns, as well as complications of chronic infection.
在美国，关于流行病学和结局的当代基于人群的数据有限。我们的主要目的是利用退伍军人健康管理局的数据创建一个经过验证的接受检测或治疗的退伍军人队列。
利用退伍军人健康管理局的结构化和非结构化数据，我们开发并验证了4种算法，分别用于检测感染情况（3种算法）和治疗状态（1种算法）。在开发阶段，我们根据对随机抽取的电子健康记录集（参考标准）进行人工审核，对每种算法进行迭代修改。验证目标是使阳性预测值（PPV）和/或阴性预测值（NPV）的单侧95%置信下限（LB）>90%。当PPV和NPV都相关时，我们应用了Bonferroni校正。
对于检测感染情况，使用结构化数据（即实验室检测）区分阳性与阴性状态时，我们实现了99.0%的PPV（LB = 94.6%）和100%的NPV（LB = 96.4%）；使用非结构化数据（即组织病理学报告）时，实现了95%的PPV（LB = 90.3%）和97.9%的NPV（LB = 93.9%）。诊断代码对诊断的PPV达到98%（LB = 93.0%）。治疗算法由多种抗菌药物组合构成，总体上对治疗的PPV≥98%（LB = 93.0%），阿莫西林/左氧氟沙星除外（PPV<60%）。应用这些算法在1999年至2018年间产生了近120万接受检测和/或治疗的退伍军人。
我们组建了一个经过验证的全国退伍军人队列，这些退伍军人接受了感染检测或治疗。该队列可用于评估流行病学和治疗模式，以及慢性感染的并发症。
