A Longitudinal Model and Graphic for Beneft-risk Analysis, With Case Study
A novel method for simultaneously visualizing beneft and risk over time is presented. The underlying model represents a subject’s beneft-risk state at a given time as one of fve discrete clinical states, one being premature study withdrawal. The new graphic uses colors to represent each subject’s changing state over the course of the clinical trial. The user can quickly grasp how a treatment affects subjects in aggregate, then further examine how individuals are affected. It is possible to tell whether the benefcial and harmful outcomes are correlated. The method is particularly appropriate for treatments that provide only symptomatic relief. An approved drug for chronic pain is presented as a worked example.
KeywordsBeneft-risk Graphics Longitudinal Safety Missing data
Unable to display preview. Download preview PDF.
- 2.National Research Council. The Prevention and Treatment of Missing Data in Clinical Trials. Panel on Handling Missing Data in Clinical Trials. Committee on National Statistics, Division of Behavioral and Social Sciences and Education. Washington, DC: National Academies Press; 2010.Google Scholar
- 7.Szarfman A, Talarico L, Levine JG. Analysis and risk assessment of hematological data from clinical trials. In IG Sipes, CA Mcqueen, AJ Gandolf, eds. Comprehensive Toxicology 4. Amsterdam: Elsevier Science; 1997.Google Scholar
- 8.Tufte ER. The Visual Display of Quantitative Information. Cheshire, CT: Graphics Press; 1983.Google Scholar
- 9.Ceesay P, Entsuah R. Beneft risk assessment incorporating time component. Paper presented at Joint Statistical Meetings, August 4, 2010, Vancouver.Google Scholar
- 10.Norton JD. A longitudinal model for medical beneft-risk analysis, with case study. Paper presented at International Chinese Statistical Association 2010 Applied Statistics Symposium, June 21, 2010, Indianapolis.Google Scholar
- 12.Mantel N. Chi-square tests with one degree of freedom; extensions of the Mantel-Haenszel procedure. J Am Stat Assoc. 1963;58:690–700.Google Scholar