Healthcare professionals …. please help fuel the research. I know you have a lot to share about these new developments (artificial intelligence) and hence will request your comments. Thanks in advance.
A few years back, I served as a board member of a newly formed accountable care entity. This multi-hospital, multi-county accountable care entity spent significant time and effort to develop a risk stratification and care coordination model. The goal was that the model will not only provide efficient and effective care but will also be based upon evidence based medicine.
If you are from the healthcare field, I know what was the first word that just went through your mind: pipedream (or perhaps more creative forms of it “wishful thinking”, “yeah sure”, “bull”, or simply a “ha”). And if that is what you thought, you weren’t too far off.
Getting participating systems to agree on what is the right way to measure risk, classify patients into a risk classification system, and identify the appropriate approach to coordinate care was not easy. In fact, at times it felt as if it was not even possible. Despite the claims of adherence to evidence based medicine, medicine is still practiced with a wide degree of variability across systems, and sometimes even within a given department of the same hospital.
Lomiata, a predictive analytics company, recently launched an artificial intelligence based risk stratification and care coordination system. The company claims that it offers a deep learning platform that takes in data from traditional sources such as electronic health records, claims, and labs and helps in risk stratification. The system has been trained using 175 million patient-record years. The company claims that it is the only solution that provides “evidentiary clinical rationale with each prediction to enable stronger alignment amongst all clinical stakeholders”. Based upon my experience in working with accountable care entities and hospital systems, this will certainly be of great value. (Note: I am not invested in or affiliated with the company in any manner). You can read their story here.
The good part: Finally a system that provides more than just risk stratification based upon preset probabilities – and actually backs it up with evidentiary rationale. While this may still invoke a lot of discussions among the physicians, but at least those discussions will be focused and relevant. Another good part is that the system has the ability to learn and will become better with experience. Linking care coordination with risk stratification is also a powerful feature. Given the constant development in technology and care coordination models, it will be good to have a more evolutionary system than a stationary frozen in time system.
The challenging part: How is the learning acquired by the system with 175 million patient-record years different from the traditional models for risk stratification? In other words, if the system is using tons of computing power and data to essentially classify a patient in the same category that an eight-question excel based questionnaire can do, then why make things more complicated. Second, what are the quality controls of the system? Third, how is the learning shared across multiple implementations? Fourth, how are changes in care coordination models incorporated in the system? Fifth, what are the sources of the evidence based medicine?
What other issues, opportunities, or problems you see with the usage of artificial intelligence for risk stratification and care coordination? Please communicate in the comments section and help fuel the research.