Curtis (Curt) B. Storlie, Ph.D., has methodological expertise in Bayesian hierarchical modeling, longitudinal and spatial modeling, missing data, nonparametric regression, machine learning, and artificial intelligence (AI) approaches. Most of his journal publications discuss complex statistical modeling challenges resulting from collaborative research, which is now focused on improving clinical practice with data science.
Dr. Storlie's research at Mayo Clinic has been extraordinarily interdisciplinary in nature. The goal of this research is to understand the critical aspects of a specific problem and what is needed statistically to help achieve the objective. Inevitably, the scientific questions and data then motivate his methodological research as well.
- Palliative care. Identify patients who have uncontrolled symptom burdens using demographics, prior history, vital signs, lab results and notes. Conduct an intervention to arrange a palliative care consult with a specialist.
- Drug diversion detection. Identify bad actors (medication theft) among providers on the basis of medication transaction, charting, and waste data.
- Intensive care unit (ICU) bleeding and other adverse event monitoring and prediction. Use vital signs, lab results and notes of ICU patients to detect suspected bleeding, shock and other adverse events earlier.
- Surgical site infection. In real-time, identify surgical patients with the highest risk of developing or having infection during the course of their encounter using demographics, history, procedure, vital signs, lab results and notes. Use transfer learning to assist in training models for lower volume or event-rate surgical specialties and institutions.
Significance to patient care
Dr. Storlie's goal is to take data science and AI research to the next level at Mayo Clinic by not only applying state-of-the-art approaches to build models but also integrating these models into clinical practice in a way that improves the processes and care provided to patients.