The May meeting of Health TechNet was held on Friday, May 17 from noon to 2 pm at Nelson Mullins’ offices in Washington, D.C. The feature topic was the uses of data analytics for predicting and tracking patient conditions and diseases, and the challenges of implementing them in clinical settings. Two speakers presented information from their current research and work in this arena: Farouk Alemi, Ph.D., a Professor of Health Administration and Policy at George Mason University, and John Koenig, a principal data scientist at a healthcare analytics startup. We also had a thorough discussion of this topic from a number of our members who are involved in these issues.
Dr. Alemi described the use of predictive models for screening patients based on his research. This included screening for suicide, hospice use, diabetes, opioid abuse, substance use, anemia and other topics. He also talked about difficulties he faces in implementing these tools in clinical settings despite their effectiveness. Clinicians are frustrated by false alarms and stricken with alert fatigue. Patients are not well informed or involved. Insurance companies have not determined whether EHR-based automated screening is reimbursable. As a consequence, implementation of predictive models in clinics is lagging. Predictive medicine remains more of a promise than a reality despite accurate methods of screening patients.
John Koenig provided a review of data visualization techniques which can be used to understand patterns in healthcare data in clinical and non-clinical settings. He also answered questions about data visualization for healthcare data projects.