We used descriptive statistics to describe the study population.
Two-tailed, paired t-tests or chi-squared tests were used to
investigate improvements in patients’ health behavior and
physical quality of life over time (difference between T0 and
T1). Changes in patients’ physical quality of life and health
behaviors were compared among DMPs with different chronic
conditions using analysis of variance or chi-squared tests. We
employed a multilevel random-effects model to investigate the
predictive role of (changes in) health behavior on patients’ physical
quality of life while controlling for patients’ physical quality of life
J.M. Cramm 138 et al. / Patient Education and Counseling 95 (2014) 137–142
at T0, age, gender, educational level, and marital status. SPSS
version 20 (IBM) was used for these statistical analyses. In
addition, we tested the final full model on imputed data (10
imputed datasets based on approximately 14,470 observations
using the Monte Carlo Markov Chain (MCMC) method) using SAS.
Results were considered statistically significant if two-sided p
values were 0.05.