About Me

I’m currently the Senior Research Analyst at the Rutgers University Center for Health Outcomes, Policy, and Economics. My interests include statistical programming, data analysis, and health economics and outcomes research. Current studies include modeling the costs and utilization of platelets from the hospital transfusion service perspective; assessing costs and outcomes of several behavioral interventions for persons with dementia and their caregivers; and analyzing costs and outcomes of community-based health interventions, including programs for vaccine education and glaucoma screening.

*

Fall 2019

• 30:725:341: Economic Modeling of Pharmaceuticals and Other Health Interventions
  • Graduate course at the Rutgers University Ernest Mario School of Pharmacy.
  • Covers concepts in HEOR modeling, including budget impact analysis, decision analytic modeling, and univariate and probabilistic sensitivity analyses.

Spring 2019

• MAT 8444: Time Series & Forecasting
  • Final project: Disposable Income, Debt, and Savings: Q1 1980 through Q3 2018
  • Objective of this project was to analyze US household real per-capita disposable income, debt services payments, and personal savings over time, to understand trends in personal disposable income, and to determine whether savings can accurately be predicted from debt.

Fall 2019

• CSC 8515: Machine Learning
  • Final project: Understanding Correlates of Obesity: Supervised and Unsupervised Machine Learning Approaches
  • Objective of this project was to apply supervised and unsupervised machine learning approaches in R and Python to understand demographic and behavioral characteristics associated with obesity using the 2015-2016 National Health and Nutrition Examination Survey (NHANES) dataset.
  • Supervised approach employed a random forest using n-by-k-fold crossvalidation to classify cases into weight category by Body Mass Index (BMI).
  • Unsupervised approach applied agglomerative and divisive hierarchical clustering methods to unlabeled data, plotting the corresponding dendrograms along with BMI category labels, assessing cluster homogeneity, and comparing performance for the two clustering methods.

• kmp at prioli dot net