In the US, 37% of adults lack the health literacy needed to navigate the healthcare system. 1 Moreover, there exists a negative correlation between health literacy and age. 1 In other words, older adults, who are the largest demographic group interacting with the healthcare system, are often the least qualified. With low levels of health literacy resulting in worse health outcomes, 2 we find it necessary to reduce the gap between the health literacy of patients and the health literacy demands of the US healthcare system.

High literacy demands of the US healthcare system are due in part to the providers’ complicated usage of language. EHR systems normalize providers’ language through drop-down menus, with the provision to use free text when needed. The nature of the text occurring in the free text tends to be complex from a layperson’s standpoint for multiple reasons. One reason is abbreviations such as the term “RA”, which can stand for “Rheumatoid Arthritis”. Another reason is professional terms originating from Latin such as “prn”(pro re nata), which means “as required”. Also, professional terms such as “prophylaxis”, which means “prevention”, introduce complexity to the providers’ language. Not only do abbreviations increase the complexity of the used language, they also introduce ambiguity. For example, the abbreviation “RA” could mean “Rheumatoid Arthritis” or “Refractory Anemia” among other possibilities depending on the context it appears in. Borrowing from the field of Machine Translation (MT), we propose an automated system, which, to the best of our knowledge, is the first of its kind. By means of a unified framework, our system is designed to tackle both issues (complexity and ambiguity). It translates the multiple types of complicated terms to their simpler counterparts, so that patients with low health literacy can better understand their health information.