International Center for Health Outcomes and Innovation Research
Clinical Research
Methodological Research

Randomized clinical trials

Analysis of non-experimental data

Cost-effectiveness analysis

Policy Analysis
Patient Safety


Randomized Clinical Research

Much of the methodology for clinical trials has been developed for pharmacological treatments. Our research has highlighted some of the unique technical, methodological, and ethical challenges of applying randomized clinical trials to surgical procedures. These challenges are especially acute in life-threatening conditions, as is evident in two NIH-supported clinical trials that we are involved in: the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH trial), and that of lung volume reduction surgery (LVRS) for end-stage emphysema.

We have undertaken research on a number of methodological questions that arose in the REMATCH trial and have more general applications to trials where cost and quality are key outcomes. In particular, we have developed a method for predicting dates when interim analysis milestones would be reached, to better facilitate the process of scheduling DSMB meetings. The method is now in use in REMATCH, and a manuscript is under review. We are also developing statistical methods for jointly modeling costs, quality of life and survival in clinical trials such as REMATCH. This work will improve current methods that assume overly simple data structures and do not account for censoring of data.

Analysis of Non-experimental Data

Center faculty have a strong interest in the heterogeneous outcomes of medical treatment and the utilization of non-experimental, i.e., observational data, for purposes of medical decision-making. Recent work focused on how to utilize instrumental variable methods (an econometrics technique) to estimate treatment effects in non-experimental medical outcome studies. Current work looks at the heterogeneity of response in the outcomes observed in several empirical data sets to establish a predictive model for estimating the effects of changes in practice patterns.

back to top

Cost-effectiveness Research

CE research is becoming an important part of modern medicine. Practical problems in its application, however, are often significant. A key area of methodological research that applies to clinical trials is capturing and expressing the uncertainty in the cost-effectiveness ratio. Current solutions have serious limitations, and we suggest that one approach to the current problems is to use the Bayesian concept of probability intervals. We published an explicit application of this method in the context of pneumococcal vaccination of the elderly.

COST EFFECTIVENESS AND ITS IMPACT ON DECISION-MAKING -- Goal is to review the methodological strengths and weaknesses of current cost-effectiveness research, and examine its impact on clinical and health policy decisions. This research is funded by the Organization for Economic Cooperation and Development (OECD).

The Center is currently working to incorporate the effect of uncertainty surrounding new technologies into cost-effectiveness analysis of new health care interventions. This work is based on experience in other sectors of the economy. A great deal of work, for instance, has been done in finance on how to value a variety of financial instruments, such as options, in the face of uncertainty. This work has now been extended to corporate finance in which decisions about real investments (e.g., building a factory) are evaluated in light of the uncertainty about future developments in prices, regulations and so on.

Another direction for the Centerís health economic research is to investigate the underlying assumption of risk-neutrality that is commonly made in cost-effectiveness analyses. Risk-neutrality implies that decision-making concerning health would be no different in the presence or absence of a measurable health risk. Our research examines the evidence against the assumption that society, in general, is risk-neutral and explores the impact of failing to consider risk aversion in current studies of cost-effectiveness. This work specifies a framework for conducting cost-effectiveness analysis under conditions where society is believed to be risk averse and demonstrates the important interdependency of the mean and variance of the effectiveness of a medical intervention in determining the degree to which an intervention enhances the overall welfare of society.

back to top



Mission & Objectives |Institutional Capabilities | Research Projects |Faculty | Contact Us