Table 3 Sensitivity analyses to examine factors contributing to heterogeneity in randomised controlled trials of antidepressants versus placebo in irritable bowel syndrome (IBS)
FactorNumber of trialsNumber of patientsRelative risk of IBS symptoms persisting (95% CI)p Value for the differenceI2 (p value)
Setting
Secondary care73690.57 (0.44 to 0.75)0.0839.2% (0.13)
Tertiary care64200.76 (0.64 to 0.91)0% (0.56)
Criteria used to define IBS
Rome I or II74740.65 (0.50 to 0.84)0.8641.7% (0.10)
Clinical diagnosis63150.67 (0.54 to 0.83)11.7% (0.34)
Method of randomisation
Stated75010.68 (0.54 to 0.85)0.6635.8% (0.14)
Not stated62880.63 (0.49 to 0.82)20.4% (0.28)
Concealment of allocation
Stated33130.78 (0.63 to 0.96)0.125.1% (0.37)
Not stated104760.62 (0.50 to 0.75)24.1% (0.22)
Blinding
Double127540.70 (0.61 to 0.80)0.036.9% (0.38)
Not stated1350.30 (0.13 to 0.59)N/A
Score on Jadad scale
⩾4106510.70 (0.59 to 0.83)0.2919% (0.26)
<431380.51 (0.29 to 0.90)56.6% (0.10)
  • CI, confidence interval.