Osborne M L, Vollmer W M, Johnson R E, Buist A S
Department of Medicine, Oregon Health Sciences University, Portland 97201-3098, USA.
J Clin Epidemiol. 1995 Nov;48(11):1393-7. doi: 10.1016/0895-4356(95)00065-8.
We used medication-dispensing information for 4 years (1/1/87 through 12/31/90) to examine the utilization of anti-asthma medications among 175,562 members of a large health maintenance organization. A total of 297,863 anti-asthma medications was dispensed during the study period, over one-half of which (55%) were beta-agonists, followed by aminophylline preparations (23%) and inhaled corticosteroids (13%). Next, we compared the predictive value of three algorithms for identifying individuals with asthma: (1) two or more beta-agonist dispensings, (2) both a beta-agonist and an inhaled corticosteroid dispensing, and (3) five or more total anti-asthma dispensings. We performed chart reviews for 40 subjects aged 5-45 years in each of these three groups and made a clinical judgment, based on all available information in the chart, as to whether each patient had asthma. Two levels of certainty were used: "any asthma" and "definite asthma." All 120 charts reviewed presented a clinical picture consistent with asthma. However, patients identified by the algorithm that included both a beta-agonist and an inhaled corticosteroid were more likely to meet our criteria for "definite" asthma and more likely to have moderate to severe asthma. These results demonstrate the feasibility of using an automated outpatient pharmacy database to identify patients with asthma.
我们利用了4年（1987年1月1日至1990年12月31日）的药物配给信息，来研究一家大型健康维护组织的175,562名成员中抗哮喘药物的使用情况。在研究期间，共配给了297,863剂抗哮喘药物，其中超过一半（55%）是β受体激动剂，其次是氨茶碱制剂（23%）和吸入性糖皮质激素（13%）。接下来，我们比较了三种用于识别哮喘患者的算法的预测价值：（1）两次或更多次β受体激动剂配给；（2）一次β受体激动剂配给和一次吸入性糖皮质激素配给；（3）总共五次或更多次抗哮喘药物配给。我们对这三组中每组40名年龄在5至45岁之间的受试者进行了病历审查，并根据病历中的所有可用信息，对每位患者是否患有哮喘做出临床判断。使用了两个确定程度的标准：“任何哮喘”和“确诊哮喘”。所有审查的120份病历都呈现出与哮喘相符的临床症状。然而，通过包括一次β受体激动剂配给和一次吸入性糖皮质激素配给的算法识别出的患者，更有可能符合我们“确诊”哮喘的标准，也更有可能患有中度至重度哮喘。这些结果证明了使用自动化门诊药房数据库识别哮喘患者的可行性。
