One of the biggest hurdles in patient care faced by hospitals and medical practitioners around the world is to grasp signals provided by charts, test results, and other metrics for patients and to manage all that to improve patient care. While it might not be difficult for small number of patients, the data starts piling up exponentially as more and more patients are treated.
It can be difficult to integrate and monitor all of these data for multiple patients while making real-time treatment decisions, especially when data is documented inconsistently across hospitals. That’s where studies by researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) come in as scientists have explored ways for computers to help doctors make better medical decisions.
Researchers have developed a machine-learning approach called “ICU Intervene” that consumes huge amounts of intensive-care-unit (ICU) data, from vitals and labs to notes and demographics, and comes up with kinds of treatments that could best cater to different symptoms.
For the system to be able to grasp all this data and churn out meaningful to-the-point analytical gists, scientists used “deep learning” to make real-time predictions, learning from past ICU cases to make suggestions for critical care, while also explaining the reasoning behind these decisions.
On the same lines, another team developed an approach called “EHR Model Transfer” that can facilitate the application of predictive models on an electronic health record (EHR) system, despite being trained on data from a different EHR system. Specifically, using this approach the team showed that predictive models for mortality and prolonged length of stay can be trained on one EHR system and used to make predictions in another.
Both models were trained using data from the critical care database MIMIC, which includes de-identified data from roughly 40,000 critical care patients and was developed by the MIT Lab for Computational Physiology.
Read more about it here.