I had a fun conversation recently about a patient population that is near and dear to my heart. Near because, well, I am a member of this particular patient population, and dear because, even though I grew up in San Diego, I consider myself a Cincinnatian by heart. Long live the Queen City.
The conversation stemmed from discussions on the use of predictive analytics for the most oft mentioned subset of pretty much every patient population: diabetes patients. I’ve long joked that we healthcare technologists have a Diabetes Mention Punchcard and each ten mentions gets us a free snow cone. As part of this conversation, my data and analytics kin here at Perficient were using the power of predictive analytics to predict the behaviors of diabetes patients. As such, I will be guaranteed at least two snow cones by the end of this post.
One of the chief concerns within this discussion was readmissions of diabetes patients. Which diabetes patients can we predict will be readmitted? Why will these diabetes patients be readmitted? It’s all very fascinating work. Going beyond the buzzword of readmissions (we have a punchcard for that term as well) I had some other questions about the Cincinnati diabetes population that I contributed to the predictive analytics discussion:
- What diabetes patients are most likely to use the emergency department as their primary method of care?
- What diabetes patients are most likely to be diagnosed with costly myocardial infarction?
The reason I asked these two questions is because I know that there are healthcare organizations that have studied these two issues and found interesting correlations. In this case it is University of Southern California (by way of Los Angeles County Hospital) and Kaiser Permanente, respectively. The power of predictive analytics provides us with so much useful data that can have an incredible impact on rates or readmission and disease state for diabetes patients in Cincinnati and beyond.
But Here’s the Catch
What do you do with this data once you have it? Let’s say you know what diabetes patients are most likely to be readmitted. You know what diabetes patients will use the emergency department as their primary method of care. You know what diabetes patients are most likely to be diagnosed with myocardial infractions. Now what? How do you actually exact the change that will motivate and incentivize the necessary shifts in habitual behaviors? This, my friend, is precisely why predictive analytics and Connected Health are best friends forever. Let me demonstrate how Connected Health is actually used to solve these predictable outcomes:
- Text Messaging Programs that Change Diabetes Patient Habits: A Connected Health medium that is both low cost and highly effective at changing habits is text messaging. In fact, Los Angeles County Hospital used a text messaging program to make effective change in dissuading the use of the emergency department but also increased adherence to diabetes care protocols by study participants more generally. This, in turn, can also solve for the myocardial infarction issue as, Kaiser Permanente learned, by adding three simple pharmaceuticals to the diabetes care protocol with protocol reminders as part of the text message therapy.
- Conversation Therapy to Reduce Readmissions: Perficient had the great pleasure recently to be selected among three Adobe partners to demonstrate the power of the Adobe Mobile platform specifically within the healthcare industry. In this case we used data around reducing readmissions for the sickest of the sick post op heart care patients, and data from a study at Hahnemann University Hospital in Philadelphia, to design a conversational therapy application called Daily Dose that is used to not only set a schedule around discharge instructions but also provide the case manager/social worker with the ability to communicate with the patient post discharge. The most important corollary in the Hahnemann study was to get that patient in for their follow up appointments, and this is one of the core use cases the mobile application is designed around. This same mobile conversation therapy method can be used to help reduce readmissions for those diabetes patients we predict are most at risk.
So, there you have it. This is why predictive analytics and Connected Health are now the best of friends. This is also why I am going to spend the rest of my day enjoying my free snow cones.
