Cancer incidence and stage at diagnosis among people with recent-onset psychotic disorders: A retrospective cohort study using health administrative data from Ontario, Canada.

OBJECTIVE: Prior evidence on the relative risk of cancer among people with psychotic disorders is equivocal. The objective of this study was to compare incidence and stage at diagnosis of cancer for people with psychotic disorders relative to the general population. METHOD: We constructed a retrospective cohort of people with a first diagnosis of non-affective psychotic disorder and a comparison group from the general population using linked health administrative databases in Ontario, Canada. The cohort was followed for incident diagnoses of cancer over a 25-year period. We used Poisson and logistic regression models to compare cancer incidence and stage at diagnosis between people with psychotic disorders and the comparison group, adjusting for confounding factors. RESULTS: People with psychotic disorders had an 8.6% higher incidence (IRR = 1.09, 95%CI = 1.05,1.12) of cancer overall relative to the comparison group, with effect modification by sex and substantial variation across cancer sites. People with psychotic disorders also had 23% greater odds (OR = 1.23, 95%CI = 1.13,1.34) of being diagnosed with more advanced stage cancer relative to the comparison group. CONCLUSIONS: We found evidence of elevated cancer incidence in people with non-affective psychotic disorders relative to the general population. The higher odds of more advanced stage cancer diagnoses in people with psychotic disorders represents an opportunity to improve patient participation in recommended cancer screening, as well as timely access to services for cancer diagnosis and treatment. Future research should examine confounding effects of lifestyle factors and antipsychotic medications on the risk of developing cancer among people with psychotic disorders.