Using Google Forms to Create an Accurate, HIPAA Compliant Database for Rare and Poorly Coded Diseases in Uveitis

Purpose: The International Statistical Classification of Disease and Related Health Problems, tenth revision (ICD-10) coding in electronic health record (EHR) systems is unable to accurately characterize certain uveitis subtypes. We describe our experience with coding errors in our EHR system and our experience using Google Forms to create a Health Insurance Portability and Accountability Act (HIPAA) compliant database to accurately characterize rare and poorly coded diseases in uveitis. Methods: This pilot single-center retrospective, consecutive case series included patients diagnosed with uveitis to assess the frequency of inaccurate coding in our MDI (MDIntelleSys, Nextech Systems LLC) EHR using ICD-10 codes and the feasibility of using Google Forms to create an accurate HIPAA compliant database. Electronic medical records were reviewed to collect information on patient demographics and uveitis characteristics. Results: The charts of the first 100 consecutive patients evaluated by the uveitis unit between October 1, 2015 and October 27, 2015 were reviewed. Demographic information and uveitis characteristics were manually extracted to classify various uveitis subtypes under the correct diagnosis using an organizational framework created using Google Forms. This framework allowed for efficient selection of predetermined variables for each encounter such that data entry was efficient, accurate uveitis descriptors were used, and transcriptional errors were avoided. The most common anatomic location of inflammation was anterior uveitis (58.0%), followed by posterior uveitis (16.0%), panuveitis (15.0%), intermediate uveitis (10.0%), and primary retinal vasculitis (1.0%). Ninety four percent of diagnoses were non-infectious (94.0%). Forty six percent of the uveitides were undifferentiated (46.0%), followed by HLA B27- associated (12.0%), sarcoidosis (8.0%), and birdshot chorioretinopathy (6.0%). Of the 20 uveitic diagnoses identified with extraction using Google Forms, 7 (35.0%) did not have an associated diagnosis or ICD-10 code in MDI (22 patients) and 2 (10.0%) etiologies (acute retinal necrosis and birdshot chorioretinopathy) were coded under the same diagnosis and ICD-10 code (disseminated chorioretinal inflammation, H30.13, 8 patients). Conclusions: This pilot study successfully showed that Google Forms is useful in the creation of a HIPAA-compliant database which efficiently and accurately captures data on uveitis patients, which would otherwise be inaccurately coded using our current EHR and international disease classifications.