Abstract: Due to the high growth rate in claiming disability benefits, Social Security Administration (SSA) faces a real overload challenge. Disability determination process has turned out to be time-consuming, complicated and expensive. By unlocking patients details, we can gain valuable information that could lead to improvement in the quality of healthcare, reducing time and healthcare cost. This study presents an approach to ease the process of disability determination. Our approach uses natural language processing and biomedical text mining to deal with data stored in patients Electronic Healthcare Records (EHRs). Such data may encode significant information about the patients case. The developed system extracts relevant medical entities and builds relations between symptoms and other clinical signature modifiers. The proposed system uses extracted information as evaluation features. Such features decide whether an applicant should gain disability benefits. Evaluations show that the proposed system accurately extracts symptoms and other laboratory marks with high F-measures (93.5-95.6%). The proposed automated system deduces right assessments to approve or reject the applicants for disability benefits.
Eslam Amer and Mohammed Abel Elfatah, 2017. Adaptive Model for Disability Determination Decision Process Based on Natural Language Processing. Research Journal of Applied Sciences, 12: 384-391.