Patients with diagnosed diabetes mellitus can be accurately identified in an Indian Health Service patient registration database
Patients with diagnosed diabetes mellitus can be accurately identified in an Indian Health Service patient registration database
复制标题
DOI：
10.1093/phr/116.1.45
复制
发表时间：
2001-01-01
影响因子：
3.3
通讯作者：
Peterson, D
中科院分区：
文献类型：
--
作者：
Wilson, C;Susan, L;Peterson, D
客观的。印第安人卫生服务中心 (IHS) 维护的计算机化患者登记数据库是有关美洲印第安人和阿拉斯加原住民糖尿病流行情况的潜在重要数据来源。本研究的目的是确定该数据源的准确性，并确定最佳搜索标准，以识别 IHS 患者注册数据库中诊断为糖尿病的患者。方法。作者将一系列计算机搜索的结果与来自大型 IHS 机构的 465 个人工审核图表的“黄金标准”样本进行了比较。结果。在 15 岁及以上的患者中，识别诊断为糖尿病的患者的最佳标准是至少存在一个由 250.00 至 250.93 ICD-9 代码识别的就诊目的叙述。单一糖尿病计算机编码的存在可识别诊断为糖尿病的患者，其敏感性为 92%（95% 置信区间 [CI] 81, 97），特异性为 99%（95% Cl 98, 99），计算出的阳性预测值为 94%（95% Cl 85, 99）。在对数据库中至少记录了一个 250.00 至 250.93 ICD-9 代码的 462 名患者进行的单独图表审查中，435 名患者被诊断为糖尿病，观察到的阳性预测值为 94%。由于糖尿病的患病率随患者年龄的不同而变化，因此识别糖尿病患者的能力的阳性预测值也随年龄的变化而变化。结论。对 IHS 患者数据库进行计算机化搜索可以识别诊断为糖尿病的患者，其准确性与其他医疗保健系统数据库报告的准确性相似。
Objective. The computerized patient registration databases maintained by the Indian Health Service (IHS) represent a potentially important source of data about the epidemic of diabetes among American Indian and Alaskan Native people. The purpose of this study is to determine the accuracy of this data source, and to identify the optimal search criteria to identify patients with a diagnosis of diabetes in an IHS patient registration database.Methods. The authors compared the results of a series of computerized searches to a "gold standard" sample of 465 manually reviewed charts from a large IHS facility.Results. Among patients ages 15 years and older, the best criterion for identifying patients diagnosed with diabetes was the presence of at least one purpose of visit narrative identified by a 250.00 to 250.93 ICD-9 code. The presence of a single computerized code for diabetes identified patients with diagnosed diabetes with a sensitivity of 92% (95% confidence interval [CI] 81, 97), a specificity of 99% (95% Cl 98, 99), and a calculated positive predictive value of 94% (95% Cl 85, 99). In a separate chart review of 462 charts of patients who had at least one 250.00 to 250.93 ICD-9 code recorded in the database, 435 had a diagnosis of diabetes for an observed positive predictive value of 94%. Because the prevalence of diabetes varies by age of the patient, the positive predictive value of the ability to identify patients with diabetes also varies by age.Conclusion. A computerized search of an IHS patient database can identify patients with a diagnosis of diabetes with an accuracy that is similar to the reported accuracy from other health care system databases.