Breiner Ari, Young Jacqueline, Green Diane, Katzberg Hans D, Barnett Carolina, Bril Vera, Tu Karen
Department of Medicine, Division of Neurology, University of Toronto, Toronto, Ont., Canada.
Neuroepidemiology. 2015;44(2):108-13. doi: 10.1159/000375463. Epub 2015 Mar 7.
Incidence and prevalence estimates for myasthenia gravis (MG) have varied widely, and the ability of administrative health data (AHD) records to accurately identify cases of MG is yet to be ascertained. The goal of the current study was to validate an algorithm to identify patients with MG in Ontario, Canada using AHD - thereby enabling future disease surveillance.
A reference standard population was established using automated key word searching within EMRALD (Electronic Medical Record Administrative data Linked Database) and chart review of potential cases. AHD algorithms were generated and tested against the reference standard. The data was used to calculate MG prevalence rates.
There were 123,997 eligible adult patients, and 49 patients had definite MG (forming the reference standard). An algorithm requiring: (1 hospital discharge abstract with MG listed as a reason for hospitalization or a comorbid condition), or (5 outpatient MG visits and 1 relevant diagnostic test, within 1 year), or (3 pyridostigmine prescriptions, within 1 year) identified MG with sensitivity = 81.6%, specificity = 100%, positive predictive value = 80.0% and negative predictive value = 100%. The population prevalence within our cohort was 0.04%.
This novel validation method demonstrates the feasibility of using administrative health data to identify patients with myasthenia gravis among the Ontario population.
重症肌无力（MG）的发病率和患病率估计差异很大，行政卫生数据（AHD）记录准确识别MG病例的能力尚待确定。本研究的目的是验证一种算法，该算法使用AHD在加拿大安大略省识别MG患者，从而实现未来的疾病监测。
通过在EMRALD（电子病历行政数据链接数据库）中使用自动关键词搜索以及对潜在病例进行病历审查，建立了一个参考标准人群。生成了AHD算法并与参考标准进行了测试。这些数据用于计算MG患病率。
有123,997名符合条件的成年患者，49名患者患有确诊的MG（构成参考标准）。一种算法要求：（1）医院出院摘要中将MG列为住院原因或合并症），或（1年内5次门诊MG就诊和1次相关诊断测试），或（1年内3次吡啶斯的明处方），该算法识别MG的灵敏度为81.6%，特异性为100%，阳性预测值为80.0%，阴性预测值为100%。我们队列中的人群患病率为0.04%。
这种新的验证方法证明了使用行政卫生数据在安大略省人群中识别重症肌无力患者的可行性。
