Abstract:
Hereditary hemochromatosis (HH) is not only the most common genetic condition in persons of northern European descent (de Graaff et al., 2015), it also remains significantly misdiagnosed. The primary reason for this failure to diagnose the disease is that it lacks significant clinically apparent features. This problem is compounded by the fact that early and correct diagnosis is key to mitigating the effects of HH on patients. Untreated, HH can result in iron overload, which in turn may cause significant liver, heart, and endocrine damage. Therefore, this study will fill a gap in the research and in practice by developing an improved model of HH diagnosis and screening utilizing patient claims records and diagnosed comorbidities. This study utilized likelihood ratios and LASSO regression to reduce the feature set to develop manageable models. The results indicated effective models can be developed utilizing patient claims data to detect and diagnose Hereditary Hemochromatosis.
