Implementation of a polling protocol for predicting celiac disease in videocapsule analysis

Table 1 Number of votes cast by automata, celiac patient data

Mar	C1	N1	Vote	C2	N2	Vote	C3	N3	Vote	C4	N4	Vote
Celiac patients
IIIC	13	11	C	13	11	C	20	4	C	18	6	C
IIIC	13	11	C	16	6	C	17	7	C	16	8	C
IIIC	13	11	C	21	3	C	13	11	C	15	9	C
IIIA	17	7	C	16	8	C	19	5	C	14	10	C
IIIA	16	8	C	16	8	C	14	10	C	12	12	-
ND	16	8	C	17	7	C	16	9	C	12	12	-
IIIA	14	10	C	15	9	C	9	15	N	12	12	-
II	10	14	N	14	10	C	9	15	N	12	12	-
II	19	5	C	10	14	N	6	18	N	12	12	-
Mean ± SD	14.56 ± 2.70	9.44 ± 2.70	P = 0.022	15.33 ± 3.00	8.44 ± 3.13	P = 0.009	13.67 ± 4.85	10.44 ± 4.80	P = 0.345	13.67 ± 2.24	10.33 ± 2.24	P = 0.056
Control patients
0	9	15	N	11	13	N	5	19	N	8	16	N
0	9	15	N	8	16	N	6	18	N	8	16	N
0	10	14	N	12	13	N	9	15	N	9	15	N
0	11	13	N	10	14	N	4	20	N	11	13	N
0	16	8	C	10	14	N	9	15	N	6	18	N
0	16	8	C	11	13	N	10	14	N	9	15	N
0	10	14	N	10	14	N	10	14	N	15	9	C
Mean ± SD	11.57 ± 3.10	12.43 ± 3.10	P = 0.727	10.29 ± 1.25	13.86 ± 1.07	P = 0.006	7.57 ± 2.51	16.43 ± 2.51	P = 0.003	9.43 ± 2.88	14.57 ± 2.88	P = 0.056

Automata vote tallies for celiac (C) and non-celiac controls (N). The number of votes for each are shown for small intestinal levels 1, 2, 3, and 4. Vote = the overall vote for celiac, non-celiac, or tie (-) based on the tally for each. The vote tallies for celiac patients with villous atrophy and control patients lacking villous atrophy confirmed by intestinal biopsy. The Marsh score based on the intestinal biopsy is given (Mar) except in one patient who was a hemophiliac it was not determined (ND). Means and standard deviations are provided in the lower two rows, with significances based on the paired t-test shown.