Iris Kanera

7 167 General Discussion Statistical analyses An important strength of the effect studies in Chapters 5 and 6 were the statistical analyses. To increase internal validity, advanced effectiveness analyses were conducted (multilevel linear regression analyses) that controlled for the baseline value of the outcome variable, potential covariates (including participation in traditional cancer aftercare), baseline differences, predictors of dropout, time-effects, and hospital variance. By applying multilevel linear regression analysis, missing data could be handled adequately (Twisk, 2006). Nevertheless, and despite the low dropout rates, multiple imputations of missing data were conducted in order to reduce non-response bias and to increase power (Enders, Mistler, & Keller, 2016; Montori & Guyatt, 2001). On the one hand, we addressed possible multiple testing problems (elevated chance of Type 1 error) by applying the false discovery rate correcting procedure (Benjamini & Hochberg, 1995). On the other hand, this might have been too conservative. SUGGESTIONS FOR FUTURE RESEARCH In the context of the entire KNW project aimed at increasing QoL of cancer survivors, data on lifestyle change and coping with cancer residues were collected. In order to achieve the objectives of the present thesis, the lifestyle-related data were analyzed focusing on lifestyle intervention development, identifying the main intervention effects, and the overall use (of the modules) and appreciation of the KNW. In continuation of these studies, further research questions can be answered. It would be valuable to investigate the extent to which the KNW intervention was able to effectively influence the targeted social cognitive variables (i.e., attitude, self-efficacy, social support, intention) and whether the changes in these variables were relevant to the intervention outcomes after six and 12months. Besides assessing these variables as potential mediators, awareness of the healthiness of one’s behavior should also be considered as a mediator, since the MRA appeared to be a possible key intervention element that aimed to target awareness. Assessing awareness should therefore be added to the mediator measurements. Sustainable diet behavior change seems to be difficult in cancer survivors. It might be worthwhile to investigate the determinants of behavior continuation. So far, research on determinants of maintenance is evolving (Rothman, 2000; Rothman, Sheeran, & Wood, 2009). Therefore, behavioral maintenance should be the main focus of future studies with regard to lifestyle behavior change and continuation, which requires long-term access to behavior change programs and longer term follow-up measurements. Within the overall KNW research project, the main studies were divided between the themes lifestyle (present dissertation) and psychosocial aspects of cancer aftercare (Willems,