Iris Kanera

6 149 Long-term effects of the Kanker Nazorg Wijzer disseminated at relatively low costs (Krebber et al., 2012; Ritterband & Tate, 2009). Based on the present results on moderate PA and previously published positive outcomes of the KNW on fatigue, depression, and QoL domains, the KNW seems very promising and suitable to meet a range of aftercare needs of a large group of early cancer survivors (Willems et al., 2016). Those cancer survivors who need more intensive treatment or additional support can be identified and referred to a subsequent program including more intensive assistance. Particularly, support is recommended to increase a healthy dietary pattern and to prevent a relapse of positive dietary changes. Moreover, based on our results, it should be taken into consideration that older cancer survivors might have different or additional support needs to increase moderate PA. Even though current results derived from a strong study design, some limitations should be acknowledged. With regard to generalizability, it should be noted that the participants of the KNW did not represent the overall cancer survivor population. A large proportion of participants were middle-aged, female survivors of breast cancer with relative low levels of physical and psychological complaints and comorbidities, and already relatively active at start.Thehighnumber of breast cancer survivorsmight bedue to the relativehighprevalence of breast cancer, the good overall prognosis, and the well-organized breast cancer care in the Netherlands, which was helpful during the recruitment of study participants. Moreover, the intervention was directed to survivors who were able to get support from a web-based program. Prior findings confirm that web-based interventions generally reach more women than men (Kohl et al., 2013). Consequently, the results of our web-based intervention might not be generalizable to the overall cancer survivor population. Furthermore, the present outcomes might be affected by selective dropout. However, dropout was very low, especially for eHealth interventions (Kohl et al., 2013). Moreover, analyses were corrected for the corresponding variables and ITT analyses revealed comparable results to complete cases analyses. Another limitation might be that the outcome variables were self-reported, allowing over- and underestimation to occur due to social desirability or recall bias (van Assema et al., 2002). Misreporting of dietary intake might also be due to misrepresentation of portion sizes and daily dietary variability (Desroches et al., 2013). As previous research suggests (Vassbakk-Brovold et al., 2016), overestimation could have occurred in moderate PA, although we used a validated measurement instrument and we accounted for over- reporting (Wendel-Vos et al., 2003). CONCLUSIONS Access to the web-based, fully automated KNW resulted in significant positive sustained changes of moderate PA in the long term among cancer survivors younger than 57 years of age. The increase of moderate PA was clinically relevant. Relatively older cancer survivors