Interest in health-related quality of life (HRQOL) is soaring because Americans are living longer and want to stay healthy and active for as long as possible, and also because many interventions do not have a large effects on mortality and are intended to improve HRQOL. Commonly used HRQOL measures, such as the Short Form 36-item Health Survey (SF-36), do not include death so that subjects who die will have undefined HRQOL, called truncation by death. Existing methods for dealing with truncation by death have some limitations. The major goal of this project is to develop a principal stratification framework that can correctly address the issue of truncation by death in causal inferences of interventions on HRQOL.
The study had three main objectives: 1.) To develop estimation of causal effects of interventions on HRQOL in randomized trials with cross-sectional data in the presence of truncation by death; 2.) To develop estimation of causal effects of interventions on HRQOL in randomized trials in the presence of truncation by death and with additional missing-data; 3.)To develop estimation of causal effects of interventions on HRQOL in randomized trials with longitudinal data in the presence of truncation by death.
We showed that under certain conditions the parameters of interest are identifiable even under different types of completely non-ignorable missing data: that is, the missing mechanism depends on the outcome. To improve identifiability, we used an approach which requires a covariate associated with principal strata. Finally, we focused on estimating the complier average causal effects (CACE) in a longitudinal clinical trial with truncation by death.
We have found that the appropriate statistical methods for analysis of health related quality of life (HRQOL) due to death are lacking due to some analytic problems associated with missing HRQOL due to deaths. We have developed several new statistical methods that can handle this special feature better than the existing methods. We believe that if health service researchers can adopt these new methods in their analysis of HRQOL, the validity of their conclusions would be greatly improved.
The proposed methods will make the following contributions to the VA. First, the methods may enable the VA investigators to make better estimates of causal effects of a new intervention or program on HRQOL outcomes. As a result, VA policy makers can make better decisions regarding resource allocation and measuring effectiveness of new interventions or programs. Second, these validated statistical models can be utilized by other VA researchers as useful tools in their studies on HRQOL outcomes. Findings from this study will be disseminated through annual meetings and manuscripts.
External Links for this Project
None at this time.