PBI analyses | Number of trials | Number of meta-analyses | PBI (95% CI*) |
---|---|---|---|

Primary analysis | 118 | 31 | 0.57 (0.50 to 0.63) |

Subgroup analyses | |||

Availability of a systematic review protocol | |||

With a protocol | 27 | 12 | 0.55 (0.39 to 0.72) |

Without a protocol | 91 | 19 | 0.57 (0.49 to 0.65) |

Subgroup difference†=−0.02 (95% CI −0.19 to 0.18; two-tailed p value of test of interaction 0.87) | |||

Availability of a core outcome measurement set for the clinical condition of the review | |||

Available core set (RA/OA systematic reviews) | 55 | 15 | 0.62 (0.53 to 0.71) |

No core set (depressive/anxiety disorder systematic reviews) | 63 | 16 | 0.51 (0.41 to 0.62) |

Subgroup difference†=0.11 (95% CI −0.04 to 0.24; two-tailed p value of test of interaction 0.15) |

*Percentile-based CIs for the PBI were constructed using bootstrap methods by resampling individual trials 2000 times.12

†The confidence limits and p value for the difference in PBI between subgroups was constructed using bootstrap methods from 2000 replicates.

OA, osteoarthritis; PBI, Potential Bias Index; RA, rheumatoid arthritis.