Fig. 5
From: A Bayesian natural cubic B-spline varying coefficient method for non-ignorable dropout
a) BNSV, CLM and Frailty Model Posterior Mean Estimated Changes in CD4 + Count per Year by Dropout Time for Untreated Subjects in the WIHS, Assuming a Baseline CD4 + Count of 478.5 b) BNSV, CLM and Frailty Model Posterior Mean Estiamted Changes in ln(Odds of Suppression) per Year for Subjects that Initiated HAART in the WIHS, Assuming a Baseline CD4 + count of 267, Baseline log10(viral load)=4.2, and Random Effects = 0