By now, most practices and hospitals use EHRs if for no other reason than to meet mandated requirements to receive reimbursements under the Affordable Care Act. But to manage chronic diseases that are costing us billions, such as heart disease and diabetes to name just two, we can do much more than just meet basic mandates. For starters, we can couple EHRs with artificial intelligence capable of producing predictive analytics in order to produce better health outcomes.
We are still subpar when it comes to utilizing technological advances to manage chronic disease. Artificial intelligence and predictive analytics can combine to give us ways to predict heart arrhythmias and blood sugar levels that can help doctors help patients to engage and participate in managing their illnesses. It’s a balance of giving the right information at the right time to the right person.
Traditional Methods are Limited
We have so much data today that traditional data analysis doesn’t cut it anymore. It’s not enough to try and stay within a safe heart rate. Certainly, we can capture the heart rate during a visit and enter it into the patient’s record— but what is that really doing to help prevent a heart attack? Many other factors make a difference in an individual patient’s case than whether that person is prone to a heart attack or not.
Processing the unstructured data floating around is one key to moving beyond our current state of crisis management. To take a proactive approach we need predictive analytics. Using AI and predictive algorithms, we can marry the data to improve clinical decisions.
Artificial Intelligence and Healthcare
AI combines unstructured data, crunches the numbers together and produces information that can then produce predictive analytics that can help patients. In an online article regarding AI and healthcare, authors Anne Bruce and Dolly Hinshaw explain how AI supports providers:
“AI not only compiles information but also translates it in usable intelligence. For example, it might crunch millions of EMRs, journal articles and cancer registry entries to predict the optimal treatment for a patient with a rare form of breast cancer. Recommendations would be personalized based on the person’s genome, health history and response to past treatments. (Such programs are currently in the works at the cancer centers of Cedars-Sinai Medical Center and Memorial Sloan Kettering.)” [1]
Bruce and Hinshaw provide a few data sources from which AI can pull unstructured data and turn it into meaningful information:
- Textbooks
- Public databases (e.g., cancer registries, voluntary reporting programs)
- Electronic medical records (EMRs)
- Journal articles
- Diagnostic images
- Prescriptions
- Results of clinical trials
- Insurance records
- Genomic profiles
- Provider notes
- Wearable devices and activity trackers (e.g., Fitbit) [1]
Artificial Intelligence Can Save Money
A 2013 study at Indiana University by Casey Bennett and Kris Hauser used an AI framework combining Markov Decision Processes and Dynamic Decision Networks. The team used 500 randomly selected patients and “compared actual doctor performance and patient outcomes against sequential decision-making models, all using real patient data. They found great disparity in the cost-per-unit of outcome change when the AI model’s cost of $189 was compared to the treatment-as-usual cost of $497.” [2]
According to Bennett and Hauser, the AI framework was able to “simulate numerous alternative treatment paths out into the future; maintain beliefs about patient health status over time even when measurements are unavailable or uncertain; and continually plan/re-plan as new information becomes available. In other words, it can ‘think like a doctor.’” [2]
The ability to take information in EHRs such as vital signs and combine it with unstructured data such as provider notes and diagnostic images and create results that can, in effect, predict and possibly offset life-threatening episodes is here today. Biotricity continues to innovate within the Internet of Things space to bring these realities to providers and patients.
