Cost-effectiveness analysis (CEA) is increasingly utilized as a tool for decision makers to efficiently allocate scarce resources in health care. Cost-effectiveness analyses use data from a variety of sources to simulate what the cost, benefits and harms will be of choosing between different interventions for a given clinical scenario. In the last 20 years cost-effectiveness has become increasingly more quantitative, relying heavily on theory and methods from many fields including statistics, engineering, economics, psychology and of course, medicine. As models become more realistic, complex and relevant to health decisions, the methodological rigor and principles underling their findings are paramount. This dissertation presents work on three methodologic improvements to compute inputs to future cost-effectiveness studies. Each paper presents a problem frequently encountered when conducting a CEA, and proposes a novel solution. The first is a methodology for estimating the cost of pharmaceuticals when conducting a CEA from the health sector or societal payer's perspective. This method, which relies on publicly available data should improve transparency and consistency between studies when costing a difficult to estimate parameter. The next paper presents an approach to control for unobserved sources of variation in longitudinal estimates of cost of care of a population, adapting a statistical methodology popular in other disciplines, latent class growth mixture models, to a CEA context. Having unobserved sources of variation is often ignored by models informing CEA, and the acknowledgment that this is a problem, and in certain circumstance can be controlled for, benefits future CEAs. The final paper improves on a recent methodologic development in CEA, network meta-analysis of survival curves, and proposes how to best utilize this methodology in CEA. It provides a tutorial on the method offering recommendations for how to derive base case and uncertainty estimates from the complex statistical procedure. All three papers make a marginal improvement to an area in CEA; their adoption into standard practice can move the field forward and inform better value based decision making in health care.