Diabetes and Obesity Control
My role and business objective
14.6% of adults in Texas, over the age of 18, have diabetes. That’s almost 3 million people. Of those three million, the largest concentration is in south Texas, along the border of Mexico. Our clients, the University of Texas at Brownsville, Valley Baptist and Su Clinica were located there. Through our engagement with UT, we were able to team up with care providers of this population who were interested in exploring how Watson could help improve the quality of care from both the provider and the patient’s perspective.
My role was to meet with providers and administrators from all three facilities and identify user needs. I then worked with my team to apply Design Thinking to the problems and pain points that our users were faced with and find opportunities for Watson.
The key users, how they were identified, and the challenges I faced
Through the contract with these facilities we were given access to several nurses and primary care physicians. The administrators who were helping manage the engagement set up interviews with the providers and the patients as well. We learned that many of the physicians worked full time at larger hospitals and actually volunteered their time at Su Clinica, the clinic that serviced much of the Brownsville, TX lower income population. We learned that many patients did not speak English and some were illiterate. In most cases, it wasn’t the patient himself who was managing his own care. It was a relative, usually a son or daughter who helped manage it for them. This was beneficial because the children of the senior diabetic patients were often bilingual and that opened up the lines of communication. We also found that the parents often had no means of transportation and had to rely on the child for that, or use public transit. Appointments were often missed or never scheduled in the first place.
We needed to find a way that the providers, who were stretched for time and resources, could better serve the patients. And a way for the patients to take charge of their own care when a family member wasn’t available to do it for them.
How I practiced Design Thinking
As a team, we first held a workshop to look at the end to end experience for both providers and patients. We did this using the Design Thinking framework. We were also able to pull from our previous work with MD Anderson. We knew we wanted to use some of the capabilities we leveraged in oncology. For example, we would use Natural Language Processing to allow the physician to quickly search and find information in clinic notes that he would normally have to sift through manually. We enabled an alert system that would notify the provider in the event that a patient’s blood labs dipped to a dangerous level. This was achieved by arming the patient with a digital glucose meter that fed results through the system back to the physician. The patients had a blood pressure cuff that they used for the same purpose. Patients were able to take these devices home so they could monitor their diabetes and be in close contact with the physician without having to rely on transportation to the clinic.
The actions I took
Creating the physician facing app was less of a challenge than working on the app that the patients would use. Luckily, we were able to leverage many of the design patterns we employed in the Oncology Expert Advisor application.
We decided to create a mobile app for the patients. Surprisingly, our research showed that most of them did have Android smartphones, although they often had to rely on their children and grandchildren for instructions on how to use them. Many of the senior patients had diminishing eyesight due to having diabetes. With that limitation, as well as needing to have it translated to Spanish, I knew the app needed an interface that was easy for them to cognitively digest.
What artifacts I created and the key principles I applied
Working in an agile environment with the development team, we planned our sprints and gathered feedback from the physicians in Texas as often as possible. I delivered the user flows and graphic assets to the developers and we used an actual developed app to gather feedback instead of a prototype. The developers were able to implement user feedback into the code on the fly.
For the patient facing app, I was able to use large icons, alerts, and graphs to ensure it would be legible for patients who may be illiterate or have poor eyesight.
