Alan G Palestine1, Pauline T Merrill2,3, Sophia M Saleem4, Douglas A Jabs4,5,6, Jennifer E Thorne6,7. 1. Department of Ophthalmology, University of Colorado School of Medicine, Aurora. 2. Illinois Retina Associates, SC, Chicago. 3. Department of Ophthalmology, Rush University, Chicago, Illinois. 4. Department of Ophthalmology, Icahn School of Medicine at Mount Sinai, New York, New York. 5. Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York. 6. Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland. 7. Department of Ophthalmology, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Abstract
Importance: Electronic health record (EHR) systems based on International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) coding of disease entities are increasingly being used to generate large data sets for analysis. However, the reproducibility of ICD-10 coding in uveitis has not been assessed across EHR platforms, and imprecision in coding may lead to improper conclusions in big-data analyses. Objective: To compare ICD-10 coding of uveitis using 2 EHR systems. Design, Setting, and Participants: This study compares ICD-10 codes for 27 uveitic diseases generated by the Epic and MDIntelleSys EHR systems to the ICD-10 descriptions associated with the codes. No patient data were assessed in this study. Main Outcomes and Measures: The number of diseases for which ICD-10 coding differed between the 2 systems. Results: Thirteen of 27 uveitic diseases were coded differently by the 2 EHR systems. Coding imprecision was notable in that the Epic system returned 16 ICD-10 codes and the MDIntelleSys returned 12 ICD-10 codes to describe 13 diseases; 4 diseases had multiple codes returned, and 6 codes were used to describe more than 1 disease. For example, MDIntelleSys uses ICD-10 code H30.13 for both birdshot choroiditis and acute retinal necrosis, while Epic uses H30.9 for both birdshot choroiditis and multiple evanescent white dot syndrome; MDIntelleSys uses this code for multifocal choroiditis. Furthermore, the ICD-10 descriptions for certain codes lack specificity, allowing variable interpretation by the coder. Conclusions and Relevance: This study suggests there is substantial disparity in the ICD-10 codes that are generated for specific uveitides by the 2 EHR systems studied. This result implies that analysis of large databases generated from the pooling of EHR data could produce results with substantial bias because of misclassification resulting from conflicting and imprecise coding of uveitides. Therefore, research into outcomes, costs, health care utilization, and epidemiology in uveitis might be improved if a more uniform coding system to describe ocular inflammatory disease is implemented.
Importance: Electronic health record (EHR) systems based on International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) coding of disease entities are increasingly being used to generate large data sets for analysis. However, the reproducibility of ICD-10 coding in uveitis has not been assessed across EHR platforms, and imprecision in coding may lead to improper conclusions in big-data analyses. Objective: To compare ICD-10 coding of uveitis using 2 EHR systems. Design, Setting, and Participants: This study compares ICD-10 codes for 27 uveitic diseases generated by the Epic and MDIntelleSys EHR systems to the ICD-10 descriptions associated with the codes. No patient data were assessed in this study. Main Outcomes and Measures: The number of diseases for which ICD-10 coding differed between the 2 systems. Results: Thirteen of 27 uveitic diseases were coded differently by the 2 EHR systems. Coding imprecision was notable in that the Epic system returned 16 ICD-10 codes and the MDIntelleSys returned 12 ICD-10 codes to describe 13 diseases; 4 diseases had multiple codes returned, and 6 codes were used to describe more than 1 disease. For example, MDIntelleSys uses ICD-10 code H30.13 for both birdshot choroiditis and acute retinal necrosis, while Epic uses H30.9 for both birdshot choroiditis and multiple evanescent white dot syndrome; MDIntelleSys uses this code for multifocal choroiditis. Furthermore, the ICD-10 descriptions for certain codes lack specificity, allowing variable interpretation by the coder. Conclusions and Relevance: This study suggests there is substantial disparity in the ICD-10 codes that are generated for specific uveitides by the 2 EHR systems studied. This result implies that analysis of large databases generated from the pooling of EHR data could produce results with substantial bias because of misclassification resulting from conflicting and imprecise coding of uveitides. Therefore, research into outcomes, costs, health care utilization, and epidemiology in uveitis might be improved if a more uniform coding system to describe ocular inflammatory disease is implemented.
Authors: Douglas A Jabs; Andrew Dick; John T Doucette; Amod Gupta; Susan Lightman; Peter McCluskey; Annabelle A Okada; Alan G Palestine; James T Rosenbaum; Sophia M Saleem; Jennifer Thorne; Brett Trusko Journal: Am J Ophthalmol Date: 2017-11-06 Impact factor: 5.258
Authors: B Trusko; J Thorne; D Jabs; R Belfort; A Dick; S Gangaputra; R Nussenblatt; A Okada; J Rosenbaum Journal: Methods Inf Med Date: 2013-02-08 Impact factor: 2.176
Authors: Durga S Borkar; Lucia Sobrin; Rebecca A Hubbard; John H Kempen; Brian L VanderBeek Journal: Ophthalmic Epidemiol Date: 2018-12-06 Impact factor: 1.648
Authors: Timothy T Xu; Cole E Bothun; Tina M Hendricks; Sasha A Mansukhani; Erick D Bothun; Launia J White; Brian G Mohney Journal: Ophthalmic Epidemiol Date: 2021-11-25
Authors: Lucia Sobrin; Yinxi Yu; Samuel Han; Gayatri Susarla; John H Kempen; Rebecca A Hubbard; Brian L VanderBeek Journal: Graefes Arch Clin Exp Ophthalmol Date: 2021-05-29 Impact factor: 3.535
Authors: Mohammed Al Ani; George Garas; James Hollingshead; Drostan Cheetham; Thanos Athanasiou; Vanash Patel Journal: Med Princ Pract Date: 2022-05-18 Impact factor: 2.132