Variation in Cancer Outcomes Amongst Children and Young Adults in Yorkshire

This study set out to improve upon the quality of research addressing variation in cancer outcomes amongst children and young adults (CYAs) through the novel application of multiple imputation (MI) to the population based Yorkshire cancer register. The study also sought to determine whether there were inequalities in disease severity according to age, ethnicity or deprivation for CYAs with cancer for the first time in the UK and to
quantify cardiovascular late effects amongst survivors of CYA cancer based on a unique data linkage approach to hospital admission data.

Key survival inequalities for CYAs with central nervous system (CNS) tumours (n=795), leukaemia (n=912) and germ cell tumours (GCTs) (n=846) between 1990 and 2009 were identified. Teenagers and young adults (TYA) compared to children across all three groups and those of non-white and non-Asian ethnicity diagnosed with CNS tumours had significantly poorer survival. Importantly, these survival inequalities could not be explained by differences in the stage at diagnosis according to age, ethnicity or deprivation. Survival rates from CNS tumours and leukaemia continued to improve over time. These improvements only became evident after using MI to adjust for missingness, which is often ignored by researchers. Despite no observed improvement for GCTs over time, the number of advanced stage tumours at diagnosis decreased significantly for this diagnostic group. For all cancers combined, the long term cardiovascular effects of cancer exist not only for children, but also for TYAs surviving their cancer.

Continued efforts should be made to ensure equal access to clinical trials and improved treatment protocols for TYAs. In addition, children as well as TYAs should be monitored for early signs of cardiovascular disease to maximise cardiovascular health. Finally, ignoring missing data can result in reduced study power and biased estimates, thus researchers should strive to use advanced techniques such as MI to account for missing data.