Working with Catholic Medical Center and Health Care for the Homeless of Manchester, I used a combination of structured and unstructured deidentified EMR data to perform a machine learning analysis that provided new insight into the characteristics and health care needs of HCH's clinic subpopulations.
With Dhara Kapoor, MSPT
After adjusting for counties’ mean race variation, proportion of sex, mean age, 2010 Health Factors Ranking (University of Wisconsin Population Health, 2010), the mean number of chronic conditions and the mean Elixhauser (Elixhauser et al., 1998) score per user, the counties with higher-than-expected Medicare costs were subjected to an exploratory factor analysis (EFA) uncovering 3 primary factors explaining 52% of excess Medicare spending. We also uncovered some important realities about Medicare underutilization in the United States.
At Elson S. Floyd College of Medicine, I had the opportunity to teach a brief course on R programming for data analysts in the Evaluation and Assessment units.
I wanted to give them a working knowledge that would hit all the R scripting basics and highlight important parametric and non-parametric tests, while providing some background on the history of computing, the internet, and the cloud.
Deepening our understanding of this important subject forces all of us to apply dedication, careful planning and empathy to our scientific and technical teamwork.
Rising to famous Data Science YouTuber Ken Jee's challenge in my own way, and enjoying every nerdy minute! 🤓💚