Aim To describe educational inequalities in cancer survival and to what extent these can be explained by comorbidity and health behaviours (smoking, physical activity and alcohol consumption).

Methods The GLOBE study sent postal questionnaires to individuals in The Netherlands in 1991 resulting in 18 973 respondents (response 70%). Questions were asked on education, health and health-related behaviours. Participants were linked for cancer diagnosis (1991–2008), comorbidity and survival (up to 2010) with the population-based Eindhoven Cancer Registry; 1127 tumours were included in the analyses.

Results 5-year crude survival was best in highly educated patients as compared with low educated patients for all cancers combined: 49% versus 32% in male subjects (log rank: p<0.0001), 65% versus 49% in female subjects (p=0.0001). Compared with highly educated, low educated prostate cancer patients had an increased risk of death (HR 2.9 (95% CI 1.7 to 5.1), adjusted for age, stage and year). No or inconsistent associations between educational level and risk of death were seen in multivariable analyses for breast, colon and non-small cell lung cancer. Although survival in prostate cancer patients was affected by comorbidities (HR₂__vs__{0_comorbidities}: 2.6 (1.5 to 4.4)), physical activity (HR_no/little__{vs__moderate}__{physical__activity}: 2.0 (1.2 to 3.4)) and smoking (HR_current__vs__{never_smokers}: 2.6 (1.0–6.8)), these did not contribute to educational inequalities in prostate cancer survival (HR_low__vs__high__education: 3.1 (1.6 to 5.8) with adjustment for comorbidity and lifestyle).

Conclusions Compared with low educated, highly educated prostate cancer patients had better survival. Although presence of comorbidities, physical activity levels and smoking status affected survival from prostate cancer, these did not contribute to educational inequalities in survival. The role of other factors for inequalities in cancer survival needs to be explored.