Skip to content

Internet Explorer is no longer supported by this website.

For optimal browsing we recommend using Chrome, Firefox or Safari.
Heart icon (animated) heart icon (static)
Champions of Science

Latin America and Caribbean Storytelling Challenge: Meet Rafael Teixeira Sousa

Diabetes is a rapidly expanding health problem in Brazil and throughout the world. Without regular monitoring, the chronic disease often comes with high cost and high mortality. To help reverse this trend, Teixeira Sousa used artificial intelligence to develop an app that can help identify diabetic patients most at-risk for developing serious complications.
Today's Top Reads Close

Rafael Teixeira Sousa is a Ph.D. candidate in Computer Science at the Universidade Federal de Goiás in Brazil.

Using his research in artificial intelligence, Teixeira Sousa created an app for health insurers that can evaluate the risk of individual diabetic patients developing serious complications of the disease. In a matter of minutes, the tool can assess millions of subjects and accurately identify those with higher risk. Health insurers can then create more effective monitoring programs for high-risk patients. Ultimately, Teixeira Sousa hopes his app will help decrease the mortality and hospitalization of diabetic patients.

Thanks to his noteworthy contributions championing scientific advancements, Teixeira Sousa was recognized as one of five winners of the Champions of Science® Storytelling Challenge: Latin America and Caribbean Edition.

Teixeira Sousa's Full Submission
I am a doctoral student in Computer Science in Brazil and have been working in a Deep Learning application, an Artificial Intelligence, in risk evaluation in diabetic patients in order to reduce diabetes complications through personal monitoring. I especially believe that Artificial Intelligence will be able to help professionals with their cognitive abilities. This way, I hope my project may turn diabetes monitoring into something more accurate and effective.

The project started as a challenge for a University subject´s project. People from health insurance branch in Brazil came to us with the concern of monitoring diabetic patients. As I had to develop a final project for the subject, I embraced the problem and invested about three months developing the first results.

The inspiration to develop this work comes from a large and fast growth in diabetes in Brazilian population and throughout the world. In 2014 it had a prevalence of 8.9%, totaling 422 million people. As this is a chronic disease with a slow and subtle progression, a follow-up is necessary to avoid the evolution of the disease. For health insurance companies, the lack of monitoring means high cost and high mortality. Therefore, a movement in the Brazilian health industry has emerged in order to employ prevention programs for chronic patients. However, evaluating people to determine their needs is an epic task, as the number of diabetics is huge, making the manual selection of high-risk diabetics for a more intense monitoring impossible.

Thinking on monitoring programs, we started to develop a tool to evaluate and identify the patients who are advancing to a possibly severe complication, as for example infarction, cerebrovascular accident, amputations and renal failure. With a more customized and oriented follow-up for each person, we will be able to identify each need and to treat the progression of the disease individually. Today this is difficult because most of the monitoring programs are focused on the population and cannot evaluate the subject. We believe that an individual treatment could save lives and bring health quality to diabetics. We will not reach this only with specific drugs and procedures, but also by understanding and prioritizing health, and not the disease.

Currently, the tool has an initial version that in a few minutes can evaluate millions of subjects in health insurances and identify those of higher risk with good precision. The tool has been used in the screening of subjects to elaborate a more effective and oriented monitoring. In the future, we wish to report positive effects in the decrease of mortality and hospitalization of diabetic patients.

We hope that, in the next years, we may take the application to the whole country and then worldwide, for a better control and treatment of diabetes. Lastly, I believe that telling the story of the project may show people how academia and industry connect in the search of solutions. And also, how artificial intelligence applications will improve healthcare services and the lives of everyone in the future.

About the Latin America and Caribbean Storytelling Challenge
Through the Champions of Science® Storytelling Challenge: Latin America and Caribbean Edition, Johnson & Johnson invited innovators working in the region to share their stories to help engage the public, encourage advocacy for scientific innovation and inspire youth in the region to pursue STEM careers that will help change the trajectory of health for humanity.

After receiving nearly 100 submissions between January and March 2019, an independent committee of scientists, policymakers and science journalists reviewed the applications and determined the winners.

This site uses cookies as described in our Cookie Policy. Please click the "Accept" button or continue to use our site if you agree to our use of cookies.
Close cookie banner icon
Back to top
You are now leaving The site you’re being redirected to is a branded pharmaceutical website. Please click below to continue to that site.