Encoding and Processing Patient Allergy Information in EHRs
The research team developed and evaluated a natural language processing allergy module that was used to study different types of allergies in an electronic health record.
The research team developed and evaluated a natural language processing allergy module that was used to study different types of allergies in an electronic health record.
This research combined the artificial intelligence technology technique Dynamic Logic with natural language processing to create a model to predict risk of death over the next 12 months and found it was better than benchmark statistical and machine learning algorithms.
Researchers created a drug allergy module that detects inconsistencies in allergy information within the electronic health record and uses a dynamic picklist that puts answers in order of how important they are based on the allergen input.
This research iteratively designed and developed a standards-based, interoperable, and publicly available clinical decision support resource to aid primary care practices in instituting routine fall risk assessment and prevention care plans.
This research will use digital health tools leveraging patient-reported outcomes and data from electronic health records to engage individuals with multiple chronic conditions to improve understanding of individualized risk of adverse events during care transitions.
This study will develop, evaluate, and disseminate a multicomponent intervention including a mobile health application and digital navigator training to support safe care transitions for patients with multiple chronic conditions.
This research will refine a current health information exchange platform to improve data exchange for inter-hospital transfers, evaluate its impact, and create a dissemination toolkit so that others may adopt this model.
This research will create clinical decision support artifacts for three patient-centered outcomes research guidelines around advanced diagnostic imaging using standards to allow them to be shareable, interoperable, and scalable; and implement them in different workflows and settings measuring their impact.