Table 2. Regression model estimating associations between diagnostic interval and number of ‘competing demand’ and/or ‘alternative explanation’ conditions.

	Coeff. (95% CI)	P-value	Exponentiated coeff. (95% CI)	Diagnostic interval change in days (95% CI)a
Female gender	0.02 (−0.05, 0.09)	0.544	1.02 (0.95, 1.09)	2 (−3, 7)
Age group (years)b	0.07 (0.05, 0.09)	<0.001
40–59 (reference)	0.00	—	1.00
60–64	−0.13 (−0.27, 0.02)	0.083	0.88 (0.76, 1.02)	−9 (−18, 1)
65–69	0.05 (−0.09, 0.18)	0.505	1.05 (0.91, 1.20)	4 (−7, 15)
70–74	0.06 (−0.08, 0.19)	0.395	1.06 (0.93, 1.21)	5 (−6, 16)
75–79	0.10 (−0.03, 0.23)	0.132	1.11 (0.97, 1.26)	8 (−2, 20)
80–84	0.32 (0.19, 0.45)	<0.001	1.38 (1.21, 1.57)	29 (16, 44)
85+	0.34 (0.21, 0.48)	<0.001	1.41 (1.23, 1.62)	31 (17, 47)
No. of 'competing demand' conditionsb	0.09 (0.06, 0.12)	<0.001
None (reference)	0.00	—	1.00
One	0.12 (0.03, 0.21)	0.009	1.13 (1.03, 1.24)	10 (2, 18)
Two	0.20 (0.10, 0.30)	<0.001	1.22 (1.11, 1.34)	17 (8, 26)
Three	0.30 (0.18, 0.41)	<0.001	1.34 (1.20, 1.51)	26 (15, 39)
Four or more	0.35 (0.20, 0.49)	<0.001	1.41 (1.22, 1.63)	32 (17, 49)
No. of 'alternative explanation' conditionsb	0.09 (0.03, 0.14)	0.003
None (reference)	0.00	—	1.00
One	0.12 (0.04, 0.19)	0.003	1.12 (1.04, 1.21)	9 (3, 16)
Two or more	0.13 (−0.01, 0.27)	0.061	1.14 (0.99, 1.32)	11 (0, 24)
Inflammatory bowel disease	0.29 (0.17, 0.41)	<0.001	1.34 (1.18, 1.51)	26 (14, 39)
Non-colorectal cancer	0.11 (−0.05, 0.28)	0.173	1.12 (0.95, 1.32)	9 (−4, 24)

Abbreviation: CI=confidence interval.

^aCalculated using the geometric mean (as used by the log-tranformed regression model) of 76.71.

^bSensitivity analyses entered these covariates as ordinal variables to assess their global effects, which are reported above the effects of their separate levels.