Sramana Mitra: These models that are helping with the recommendations, do you have heuristics? How are the models architected?
Jas Grewal: The models are actually based on past research. One of the things about CareSkore is that we are a pretty academic company. We started following patients. We have historical data from Stanford and Yale. We also work directly with Medicare. In our data bank, we have data from 15 different hospitals on around 39 million patients.
I think there are 90,000 physicians that took care of those patients. We based our model based on that training data and understanding that if you are a 45-year-old male, African-American, live in a rich neighborhood, and have chronic heart failure, we would compare your profile to the millions of profiles that we have to say what are the chances of a bad event happening or a risk onsetting for you.
The interesting thing about CareSkore is that we not only go with data in the electronic health records, but we also bring in a lot of socio-economic demographic and behavioral data from external sources. 60% of a patient profile is from clinical and lab results. 40% of it is what your socio-economic and demographic status is. That has a huge impact on your healthcare. We are one of those companies that are going after 43 different data sources. We pull that information in and make the predictions.
Sramana Mitra: Fantastic. How many healthcare systems are using this system at this point?
Jas Grewal: Right now, we are working in 18 organizations and they have multiple sites. We’re deployed in over 50 sites.
Sramana Mitra: Across the 50 sites, what is your estimate of the number of patients who are being serviced?
Jas Grewal: They’re in the millions.
Sramana Mitra: You have a large data volume that you’ve already been active and deployed on.
Jas Grewal: We have two parts to the solution. The first is, how do you predict and understand the patient better? How do you understand the risk and the gaps for a patient? The second part is, once you’ve understood what the patient needs, how do you engage them? How do you communicate with them? On the engagement side, we are starting to do more and more. Right now, we are predicting and looking at well over 1.8 million patients.
Sramana Mitra: Are there trends or conclusions that you could draw from this body of patients that you are engaged with?
Jas Grewal: Absolutely. There are massive conclusions that we are looking at. The data is very insightful. We have been practicing a lot of evidence-based medicine in this country, but this is a time where hospitals and even insurance companies need to think about evidence-based management of patients.
One of the interesting things that we’re seeing is the huge impact of weather. Local snow has a huge impact on no-shows and cancellation. We can predict those volumes out on when patients are going to cancel. The second thing that we’re seeing is the onset of flu in a particular area has a huge impact on the Emergency Department utilization. We are bringing that third-party information.