Improving the Analysis of KCCQ Endpoints in Heart Failure Clinical Trials

Figures

• 

  Fig. 1. Baseline-dependency of treatment effects in conventional mean change from baseline analysis of KCCQ scores. Grey areas indicate unobservable changes caused by the boundedness of the score, lines indicate treatment group mean changes

• 

  Fig. 2. Examples of possible distributions for beta distributed KCCQ score as characterized by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha $$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta $$\end{document} parameters (multiplied by 100 from original beta variable)

• 

  Fig. 3. KCCQ distributions and change from baseline in simulated randomized trials. Distributions for two N = 4000 simulated trials with a large effect on KCCQ and either a low or high baseline KCCQ distribution, respectively. Grey lines indicate bounds of the scores, colored lines represent mean regression lines per treatment group

• 

  Fig. 4. KCCQ Overall Summary Score distributions and individual changes from baseline in PRIORITIZE-HF. Grey lines indicate bounds of the scores, colored lines represent mean regression lines per treatment group

• 

  Fig. 5. Mean treatment effects for KCCQ Overall Sumary Score change from baseline in PRIORITIZE-HF. Treatment effects represent differences in mean change from baseline at 12 weeks between treatment groups. Shaded areas represent 95% confidence intervals. Mean effects > 0 favors SZC treatment