Table 1.
Mean difference scores and standard errors for bowel habit data from all subjects in Trials A and B, and subjects with low initial stool frequency (LISF) from Trials A and B (A & B-LISF). Data shown as ‘Mean’ are the average of difference scores for weeks 3 and 5, with corresponding SEM.
| Outcome | Data from Trial | Number of subjects | Mean | SEM | p-value -interaction LISF versus HISF‡ | P-value |
|---|---|---|---|---|---|---|
| Stool frequency | A & B | 28 | 0.08 | 0.05 | 0.025* | 0.98 |
| A & B-LISF only | 16 | 0.91 | 0.06 | 0.032*§ | ||
| Stool consistency | A & B | 29 | 0.28 | 0.10 | 0.91 | 0.008* |
| Flatulence | A & B | 30 | −0.14 | 0.09 | 0.39 | 0.125 |
| Rumbling | A & B | 30 | −0.02 | 0.05 | 0.43 | 0.67 |
| Bloating | A & B | 29 | 0.03 | 0.04 | 0.125 | 0.39 |
| Cramping | A & B | 30 | −0.02 | 0.03 | 0.088 | 0.67 |
*denotes significant difference <0.008 (for secondary outcomes with Bonferroni correction).
† p-values for ‘Trials A & B’ indicate if the difference scores for all subjects together are different from 0.
‡ p-value interaction LISF versus High Initial Stool Frequency (HISF) show if the subjects in both trials with LISF had a significantly different outcome than those in Trial B subjects with HISF. § p-value for stool frequency, where this interaction is significant for subjects in Trial A & B-LISF, shows the difference between inulin treatment and placebo was significant for those subjects in both trials who had low initial frequency.