At a time when cancer research is increasingly complex and expensive to conduct, cost-effective study design is an imperative. Despite the well-known and long-standing problem of depending on unreliable estimates of the intraclass correlation (ICC) when designing studies, serious efforts to address this problem are lacking. The objective of the proposed research is to develop novel methods of predicting the ICC using mechanistic models of within-cluster dynamics and to incorporate these methods into sample size and power calculations. Cancer prevention studies motivate the work. The proposed research will have widespread practical applications for more cost-effective and scientifically defensible study designs.
Public Health Relevance When planning a cancer research study, good design is needed to ensure that the study will produce the data necessary to answer important research questions while minimizing cost. The proposed research will develop new methods of study design for a common data scenario, clustered binary data, thereby reducing costs and increasing the likelihood that research questions will be answered conclusively.
| Wu, Sheng; Crespi, Catherine M; Wong, Weng Kee (2012) Comparison of methods for estimating the intraclass correlation coefficient for binary responses in cancer prevention cluster randomized trials. Contemp Clin Trials 33:869-80 |
| Crespi, Catherine M; Maxwell, Annette E; Wu, Sheng (2011) Cluster randomized trials of cancer screening interventions: are appropriate statistical methods being used? Contemp Clin Trials 32:477-84 |
| Crespi, Catherine M; Wong, Weng Kee; Wu, Sheng (2011) A new dependence parameter approach to improve the design of cluster randomized trials with binary outcomes. Clin Trials 8:687-98 |
| Maxwell, Annette E; Bastani, Roshan; Danao, Leda L et al. (2010) Results of a community-based randomized trial to increase colorectal cancer screening among Filipino Americans. Am J Public Health 100:2228-34 |
