A system and method for generating and delivering patient-specific medication instructions are disclosed. Structured prescription data associated with a medication order is retrieved from an Electronic Health Record (EHR) data source in response to a request initiated through a user device. Personally identifiable information in the retrieved prescription data is masked to generate sanitized prescription data. Accessibility parameters associated with a patient profile, including preferred language, literacy level, visual impairment, dyslexia, and audio preference indicators, are retrieved. Medication instruction content is extracted from the sanitized prescription data and transformed into a normalized semantic representation comprising standardized medication instruction attributes. Patient-specific medication instructions are generated from the normalized semantic representation using artificial-intelligence-based constrained language generation engines. The generated instructions are rendered on the user device in various output formats, including audio output generated using a text-to-speech synthesis engine.
Full Text
What is claimed is: