Fig. 2: CRs of CD and UC across epidemiologic stages 1, 2 and 3.

a, The CR-I of CD (left) and UC (right) across the three epidemiologic stages. Number of observations: n = 263 (CD stage 1), n = 1,011 (CD stage 2), n = 796 (CD stage 3), n = 277 (UC stage 1), n = 847 (UC stage 2), n = 760 (UC stage 3). b, The CR-P of CD (left) and UC (right) across the three epidemiologic stages. Number of observations: n = 86 (CD stage 1), n = 247 (CD stage 2), n = 443 (CD stage 3), n = 118 (UC stage 1), n = 238 (UC stage 2), n = 435 (UC stage 3). Data are categorized by type (incidence/prevalence), disease type (CD/UC) and epidemiologic stage (stage 1, stage 2, stage 3), as determined by the random-forest classifier. For the box plots, the centre line shows the median, the lower hinge shows the 25th percentile (that is, first quartile) and the upper hinge shows the 75th percentile (that is, third quartile). The 25th and 75th percentiles are labelled and correspond to the CRs. Statistical analysis was performed using negative binomial regression with post hoc comparisons of estimated marginal means with Tukey adjustment for multiple comparisons, showing significant differences between all stages for the incidence and prevalence of CD and UC (P < 0.001 for all comparisons).