The proposed research is designed to contribute to an understanding of the impact and implications of the regression to the mean (RTM) effect in health research and to provide some statistical solutions to the RTM effects when assessing the effectiveness of the intervention on marker value(s) obtained at or prior to screening. RTM means that when repeated measurements of a variable are taken for a non-randomly selected subgroup, the average of the subsequent measurements will tend to move toward the population mean. The bias introduced by RTM can be encountered in practically every type of health study. Therefore, the results of this research will provide a valuable contribution to a broad spectrum of health research. This application will focus on analyzing the change in event probability resulting from the implementation of an intervention among a selected sub-population when the event of interest post-selection is a binary outcome. The marker used in selection might be the same as the outcome variable or it might be a continuous variable believed to reflect the likelihood of the outcome. The methods will be applied to real data sets from NIDDK interstitial cystitis observational study and the Pennsylvania Health Care Cost Containment Council (HC4) Reports. The proposed research is innovative and significant because current available methods for adjustment for RTM have focused only on data that are normally distributed, which are no longer adequate to cope with the complexities of health services outcomes and clinical studies in the real world.
| Lin, Hung-Mo; Lyles, Robert H; Williamson, John M et al. (2005) Estimation of the intervention effect in a non-randomized study with pre- and post-mismeasured binary responses. Stat Med 24:419-35 |
| Lyles, Robert H; Williamson, John M; Lin, Hung-Mo et al. (2005) Extending McNemar's test: estimation and inference when paired binary outcome data are misclassified. Biometrics 61:287-94 |
| Lyles, Robert H; Lin, Hung-Mo; Williamson, John M (2004) Design and analytic considerations for single-armed studies with misclassification of a repeated binary outcome. J Biopharm Stat 14:229-47 |
| Lin, Hung-Mo; Lyles, Robert H; Williamson, John M (2002) Bias in a placebo-controlled study due to mismeasurement of disease status and the regression effect. Control Clin Trials 23:497-501 |
