MIS GrApH AI
The emerging uses of artificial intelligence are transforming many industries, but the healthcare sector lags behind this progress partly because the production databases holding patient data are too poorly organized to run artificial intelligence algorithms. To address this problem, we have organized an open, sharing community of physicians, data scientists, data engineers, software engineers, students, and others inspired by a common vision: to make healthcare data AI-friendly from the ground up.
We are working use-case by use-case with real patient data to build a foundation of knowledge, a toolkit, and a corps of colleagues to act as a springboard for real-world implementations. In addition to use cases, we work to address global issues such as human-machine interfaces and knowledge graphing.
If you are interested in joining the working group, please contact Tim McLerran at tmclerran@gmail.com or through [LinkedIn](https://www.linkedin.com/in/tim-mclerran/).
Repositories
-
Public Repository
Public Repository Description
-
Credentialed Access Repository
In order to access this repository, you must submit proof that you have completed the credentialling process outlind by PhysioNet
-
Awesome List
Currated list of useful material as far as AI in EHR is concerned
Blog / Updates
Use Cases
-
Food intolerance prediction
As a nutritionist caring for a patient with a poorly defined food intolerance
-
New Clinical Trials
What new clinical trials (or treatment or medical technology) can I recommend my patient to enroll or consider?
-
Interactive Visual Learning Graph
WIP
Design and develop clinical health assisted intelligence via an interactive visual knowledge graph
-
Impact Analysis
Impact analysis for treatment decisions based on current patient state and current treatments
-
Recommend Plan
Recommend plan items for a specific problem
-
Display Problem Courses
Display the typical courses of a given problem over time in patients similar to the one I'm treating
-
Visualize History
Visualize the history of a patient's problem
-
Automated Discharge Summaries
Automated drafting of discharge summaries
-
Create Patient Summary
Create a running summary of the patient's problems
-
Unexpected Effects
Predict unexpected effects of interventions based on patient characteristics
-
What to do next
What should I do next to work up this patient's problem?
-
Likely Problems
Recommend which problems the patient is likely to have
-
Minimize Stay
Safely minimize length of hospital stay
-
Predict Staffing
Predict staffing needs
-
Bed Assigning
Assign scarce hospital beds to those who need them the most
-
Predict Transfer
Predict which patients will be transferred to a lower level of care soon
-
Optimize Scanner Utilization
Pro-actively optimize utilization of the CT scanner
-
Collect Billing Info
Automate collection of information necessary for billing
-
Doctor Assignment
Assign each patient to the doctor who can provide the best care for that individual patient
-
Medication Side Effects
Find out which patient problems likely come from medication side-effects or interactions