This blog is a contribution from our client. Mannheim University Medical Centera leading university hospital in Europe. Learn how your team is leveraging DataRobot to accelerate clinical research with AI.
As doctors and researchers, we constantly work to improve quality of life. To do this, we need a holistic, data-driven approach to help us understand the full impact of a particular treatment. How does a certain treatment affect the patient’s symptoms? What makes your symptoms better or worse?
At the University Medical Center Mannheim, we are investigating data to answer these questions. But it’s difficult for our small teams to balance research with the critically important treatment of our patients. We don’t learn data science in medical school, but it is an increasingly essential piece of the healthcare puzzle.
Fortunately, easy to use artificial intelligence platforms like DataRobot help us close the gaps between our medical expertise and the data science we need to provide more thoughtful care to our patients.
The benefits of AI in healthcare: findings we can trust
What excites me most about AI in healthcare is the potential to discover new explanations for diseases or efficacies of innovative therapies that we are too blind to see using classical statistical methods. Our goal is to uncover new influences on disease progression, predict disease flares, and empower patients to better manage treatment compliance.
DataRobot gives us exciting new ways to gain insights from our data and grow our team without data scientists.
As clinicians, we can compare and validate models to find those with the highest degree of accuracy. In the Clinical Cooperation Unit: Healthy Skin and Joints, we have leveraged AI to evaluate data from a smartphone application, including images and other clinical datasets of anonymized patient data.
Compliance is also critical – from privacy measures around patient information to GDPR regulations that protect and secure sensitive data. When we publish our findings, the most important thing is their reproducibility. This is why documentation and explainability behind models are so critical. DataRobot makes these normally labor-intensive processes seamless and automatic.
With DataRobot, we trust our findings, knowing that they have been fully trained, retrained, validated and revalidated. We have a lot of statistics to show the level of accuracy, which we also need to publish the results. Because of that, I sleep better at night and our center can make an even greater contribution to the medical research community.
Discovering links in disease progression
Another example of the use of AI in healthcare: we have applied AI to several use cases in our collaboration between dermatology and rheumatology.
For a recently published studyWe use DataRobot to analyze data from clinical research with patients with chronic eczema or psoriasis. The analysis focused on itch, pain, quality of life, and use of a smartphone tracking app to track their symptoms. We seek to discover new influences on disease progression, try to predict disease flares or promote patient adherence to treatment.
Through our analysis, we learned that nearly 30% of patients see improved quality of life at six months, while another 30% showed a decrease in quality of life or had consistently poor quality of life. This knowledge and others will influence treatment decisions. This data is transformative because we can better understand our patients and know which patients benefit from certain therapies. It tells us when and how to change patients’ treatment course if necessary.
We are now helping other clinics in the medical center discover insights from their data. With the Department of Internal Medicine, we analyze blood lipids with the goal of predicting heart disease or heart attacks. In just a few weeks/months, we’ve been able to create some pretty accurate models and we hope to publish our findings in the near future.
Using AI to accelerate medical research
All of these findings may not have been discovered without DataRobot. Instead, we have been able to accelerate research from hours to seconds, even as we continue to care for patients and focus on improving their quality of life.
AI helps our daily work and, most importantly, helps patients.
When we first partnered with DataRobot, I told others that this new technology would change the face of the Earth and that they had to learn about it. I still say this today. AI offers enormous benefits to healthcare professionals and I am delighted to see the impact of the work of the University Medical Center Mannheim.
About the author
Victor Olsavszky is a physician and postdoctoral researcher at the Dermatology Clinic of the University Medical Center and Faculty of Medicine in Mannheim, Germany. His clinical expertise focuses on dermatological surgery and plaque psoriasis, while his research areas are vascular biology and AutoML analysis of medical data sets using DataRobot.
