This dissertation applies techniques in comparative effectiveness research (CER) to understand the determinants of hospice stay duration, the cost-effectiveness of Human Papillomavirus (HPV) vaccination, and the optimal method for evaluating non-experimental and experimental data. The research on hospice stay duration uses data from the Dartmouth Atlas. It finds that for the logarithm of hospice days, nurse staffing numbers are a significant determinant of length-of-stay. Total Medicare reimbursements go down by $117 for each day that a patient stays in hospice. The HPV paper aggregates results across 7 published papers. It finds that male-female HPV vaccination is cost-effective at a $100,000/QALY cost-effectiveness threshold. Finally, the statistics paper compares the accuracy of linear regression, propensity score matching, and simple subtraction for approximating the results of a RCT. It finds all methods excepting simple subtraction and nearest neighbor propensity score matching produce results that are not significantly different from the RCT result.