Abstract
Purpose:
To assess the validity of International Classification of Diseases, 9th Edition (ICD9) codes used in the Pacific Ocular Inflammation Study, a population-based study using the electronic medical records of Kaiser Permanente Hawaii to investigate the epidemiology of ocular inflammatory disease in the Hawaiian islands.
Methods:
The electronic record system of Kaiser Permanente Hawaii was searched for any visit from January 1, 2006 to December 31, 2007 that referenced an ICD9 code that might be associated with a diagnosis of uveitis. The inclusive list of diagnosis codes used for this study was adapted from two prior uveitis incidence and prevalence studies using administrative data. A subset of these ICD9 codes included all diagnoses codes pertinent to herpes zoster ophthalmicus (HZO). Subsequently, all charts from this electronic query were individually reviewed by a uveitis specialist (NRA) to confirm a diagnosis of uveitis, as well as HZO.
Results:
Of the 873 patients identified as possibly having uveitis by ICD9 codes, 224 cases were confirmed as uveitis patient after medical record review. In our study, the most accurate ICD9 codes in identifying uveitis cases were herpes simplex iridocyclitis, 54.44 (75%), histoplasmosis retinitis, 115.92 (100 %), panuveitis, 360.12 (92%), disseminated chorioretinitis, 363.13 (100%), pars planitis, 363.21 (100%), Harada disease, 363.22 (100%), recurrent iridocyclitis, 364.02 (96%), and granulomatous uveitis, 364.1 (81%). Sixty-nine patients had an accurate uveitis diagnosis using ICD9 code 364.04, secondary iridocyclitis noninfectious. However, using this ICD9 code also contributed 122 patients with an inaccurate diagnosis of uveitis after chart review and only had an accuracy of 36%. Using ICD9 codes specific to HZO yielded 152 patients through electronic search. After chart review, 138 had a confirmed diagnosis of HZO. Overall, HZO codes had an accuracy of 91%.<br />
Conclusions:
With the increased use of data from electronic medical records for research, it is important to validate whether ICD9 diagnoses are accurate. These results suggest that in Kaiser Permanente Hawaii, using ICD9 codes alone to capture uveitis diagnoses is not always accurate. Chart review, as was done in this study, can help further elucidate accurate diagnoses. However, electronic search for ICD9 codes alone can be an accurate method for identifying cases of HZO.