Table 3.

Multivariate predictors of ACN using the model fitted to the derivation (Phase I) cohort. The predictive score derived ranged from −4 to 24 points (Final model: C-Statistic = 0.865; Brier Score = 0.10; Hosmer-Lemeshow p-value = 0.86)

Risk Factor	OR	Q 2.5	Q 97.5	Points
Age 40–50 years	2.26	0.52	10.05	3
Age 50–60 years	7.58	2.02	28.56	8
Age > 60 years	11.88	3.27	43.13	10
Colonoscopy (in the previous 5 years) (Yes)	0.36	0.2	0.65	−4
Smoking (Yes)	1.49	1.01	2.27	2
IF MAXFIT [> 4–11]
NSAMPLES> 4	1.72	1.04	2.83	2x†
IF MAXFIT > 11
NSAMPLES> 4	2.89	2.45	3.44	4x†

MAXFIT: maximum f-Hb value; NSAMPLES > 4: Number of samples with FIT > 4 μg Hb/g faeces. Discrete variable (0 to 3). Points: Points assigned to each Risk Factor

†Note on 2x/4x: factor to multiply NSAMPLES

How do we interpret the model? Ex: Suppose a 55-year-old man has not smoked or has done a prior colonoscopy and with values of FIT 4, 6 and 10 μg Hb/g faeces. This individual has a MAXFIT value of 10, and 2 samples with FIT > 4 (2 positive samples); therefore its score will be 8 points by age + 2 points per FIT (corresponding to the OR = 1.72) multiplied by 2 positive samples, giving 4 points. Total: 12 points. But if their FIT values were 4, 8, and 12 μg Hb/g faeces, the MAXFIT value of 12 and two positive samples, their score would be 8 points by age + 4 points per FIT (corresponding to the OR = 2.89) multiplied by 2 positive samples, giving 8 points. Total: 16 points