“…Finally, even though studies were conducted among patients who had agreed to participate, patients who refused to participate did not appear to be significantly different from those participants in terms of general health status and health awareness. 32,33 All four studies reporting mortality demonstrated a bias towards the null caused by early interventions to the control group prior to data collection (for example, health check to the control group prior to mortality data collection) 17,[26][27][28] or the merging of study groups (for example, in the OXCHECK study, the initial health check was conducted in 3 consecutive years in the intervention group). 29 …”