Clinical Data
A chart-abstraction form was designed in accordance with published recommendations.23 This common collection instrument was developed with a Research Electronic Data Capture form shared between all sites.24 Electronic and paper medical records were reviewed. Demographic data were extracted from existing eMERGE databases. The information extracted from medical records included medical history of hemochromatosis or iron-overload-related conditions, iron-related laboratory studies, imaging studies, and physical findings consistent with HH (Table S2).
In the absence of secondary iron-overload state, HH diagnosis was based on the physician annotation of HH in the medical records, including a recent clinical note, problem list, or ICD-9 (International Code of Diseases, Ninth Revision) codes 275.01, 275.02, or 275.03. We used a broad definition of liver disease and included individuals with liver cirrhosis, other chronic hepatic phenotypes (i.e., alcoholic liver disease, chronic viral hepatitis, fatty liver disease, or non-specific liver enzyme elevation), hepatocellular carcinoma, hepatomegaly, ascites, history of liver biopsy, and elevated liver function. A manual of procedures was written for the data abstractors’ training and reference, and common examples of situations encountered by abstractors were discussed before the final form was distributed. A pilot test form was designed to ensure the reliability and validity of the data-collection instruments. Monthly teleconferences with all sites took place during the form-development and abstraction phases of the project. These calls addressed questions about coding to ensure consistency, accounting for differences in medical-record structures (including different measuring units in the laboratories), and disparities in the availability of certain data.