The heatmap shows the estimates for pairwise comorbidity. These estimates were computed using cox proportional hazards model with adjustment for i) country, age-cohort and sex (Model A) ii) adjusted for country, age-cohort, sex, type and number of mental disorders with onset preceding the prior disorder (Model B)
We found that associations are particularly strong for the following pairs (yellow in the heatmap):
There were moderate evidence of associations for the following pairs (purple in the heatmap):