Pan Yi, Rose Charles E, Haber Michael, Ma Yan, Carrasco Josep L, Stewart Brock, Keitel Wendy A, Keyserling Harry, Jacobson Robert M, Poland Gregory, McNeil Michael M
Immunization Safety Office, Division of Healthcare Quality Promotion, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, GA, USA.
Int J Biostat. 2013 Jul 27;9(1):/j/ijb.2013.9.issue-1/ijb-2012-0001/ijb-2012-0001.xml. doi: 10.1515/ijb-2012-0001.
Cohen's kappa coefficient, which was introduced in 1960, serves as the most widely employed coefficient to assess inter-observer agreement for categorical outcomes. However, the original kappa can only be applied to cross-sectional binary measurements and, therefore, cannot be applied in the practical situation when the observers evaluate the same subjects at repeated time intervals. This study summarizes six methods of assessing agreement of repeated binary outcomes under different assumptions and discusses under which condition we should use the most appropriate method in practice. These approaches are illustrated using data from the CDC anthrax vaccine adsorbed (AVA) human clinical trial comparing the agreement for two solicited adverse events after AVA between the 1-3 day in-clinic medical record and the patient's diary on the same day. We hope this article can inspire researchers to choose the most appropriate method to assess agreement for their own study with longitudinal binary data.
科恩kappa系数于1960年被引入，是评估分类结果观察者间一致性最广泛使用的系数。然而，原始的kappa仅适用于横断面二元测量，因此，当观察者在重复的时间间隔对同一受试者进行评估时，它无法应用于实际情况。本研究总结了在不同假设下评估重复二元结果一致性的六种方法，并讨论了在实际应用中应在何种条件下使用最合适的方法。使用来自美国疾病控制与预防中心吸附炭疽疫苗（AVA）人体临床试验的数据对这些方法进行了说明，该试验比较了AVA接种后1 - 3天内临床病历与同一天患者日记中两种预期不良事件的一致性。我们希望本文能够启发研究人员为自己的纵向二元数据研究选择最合适的方法来评估一致性。
