Sensitivity of Claims-Based Algorithms to Ascertain Smoking Status More Than Doubled with Meaningful Use
Sensitivity of Claims-Based Algorithms to Ascertain Smoking Status More Than Doubled with Meaningful Use
复制标题
DOI：
10.1016/j.jval.2017.09.002
复制
发表时间：
2018-03-01
期刊：
影响因子：
4.5
通讯作者：
Shih, Ya-Chen Tina
中科院分区：
文献类型：
--
作者：
Huo, Jinhai;Yang, Ming;Shih, Ya-Chen Tina
背景资料：“有意义地使用经认证的电子健康记录”政策要求符合条件的专业人员在2011年至2012年期间记录超过50%的13岁或以上个人的吸烟状况。目的：探讨记录吸烟行为的编码是否随时间推移而增加，并评估吸烟相关诊断和程序代码在识别既往和当前吸烟者方面的准确性。方法：我们在Truven Health Analytics数据库中对2009年至2014年的5，423，880名入组者进行了观察性研究。测量吸烟编码的时间趋势、敏感性、特异性、阳性预测值和阴性预测值。结果：研究结束时吸烟行为编码率明显提高。在6年期间，索赔数据中记录为当前吸烟者的患者比例增加了2.3倍，记录为既往吸烟者的患者比例增加了4倍。各国际疾病分类第九版临床修改编码的敏感性一般小于10%。烟草使用障碍诊断码（305.1X）是识别吸烟者最敏感的诊断码（9.3%）。这些编码与现行手术术语编码的特异性均在98%以上。结论：自有意义使用政策开始以来，对当前和以前吸烟行为的编码有了很大的改进。然而，使用诊断和程序代码来识别吸烟行为的管理数据仍然是不可靠的。这表明，需要提高吸烟行为医学编码的质量，以提高吸烟相关结果研究的索赔数据的能力。
Background: The "meaningful use of certified electronic health record" policy requires eligible professionals to record smoking status for more than 50% of all individuals aged 13 years or older in 2011 to 2012. Objectives: To explore whether the coding to document smoking behavior has increased over time and to assess the accuracy of smoking-related diagnosis and procedure codes in identifying previous and current smokers. Methods: We conducted an observational study with 5,423,880 enrollees from the year 2009 to 2014 in the Truven Health Analytics database. Temporal trends of smoking coding, sensitivity, specificity, positive predictive value, and negative predictive value were measured. Results: The rate of coding of smoking behavior improved significantly by the end of the study period. The proportion of patients in the claims data recorded as current smokers increased 2.3-fold and the proportion of patients recorded as previous smokers increased 4-fold during the 6-year period. The sensitivity of each International Classification of Diseases, Ninth Reuision, Clinical Modification code was generally less than 10%. The diagnosis code of tobacco use disorder (305.1X) was the most sensitive code (9.3%) for identifying smokers. The specificities of these codes and the Current Procedural Terminology codes were all more than 98%. Conclusions: A large improvement in the coding of current and previous smoking behavior has occurred since the inception of the meaningful use policy. Nevertheless, the use of diagnosis and procedure codes to identify smoking behavior in administrative data is still unreliable. This suggests that quality improvements toward medical coding on smoking behavior are needed to enhance the capability of claims data for smoking-related outcomes research.