Skip to main content
. 2020 Jan 31;17(1):e1003019. doi: 10.1371/journal.pmed.1003019

Table 3. Application of the Bland–Altman limits of agreement to a plausible range of AD meta-analysis HRs.

Observed AD meta-analysis HR Potential IPD meta-analysis HR (after dividing by the limits of agreement: 0.90 and 1.05) Observed control group survival band Potential IPD meta-analysis absolute survival effects at a representative control group value
0.70 0.67 to 0.78 <10% 5% to 8%
10%–19% 8% to 13%
20%–49% 9% to 14%
50%–69% 7% to 11%
≥70% 5% to 7%
0.75 0.71 to 0.83 <10% 3% to 7%
10%–19% 6% to 10%
20%–49% 7% to 11%
50%–69% 5% to 9%
≥70% 3% to 6%
0.80 0.76 to 0.89 <10% 2% to 5%
10%–19% 4% to 8%
20%–49% 4% to 10%
50%–69% 4% to 8%
≥70% 2% to 5%
0.85 0.81 to 0.94 <10% 1% to 4%
10%–19% 2% to 6%
20%–49% 2% to 7%
50%–69% 2% to 6%
≥70% 1% to 4%
0.90 0.86 to 1.00 <10% 0% to 2%
10%–19% 0% to 4%
20%–49% 0% to 5%
50%–69% 0% to 4%
≥70% 0% to 3%
0.95 0.90 to 1.05 <10% −1% to 1%
10%–19% −2% to 3%
20%–49% −2% to 3%
50%–69 −2% to 3%
≥70% −1% to 2%

AD, aggregate data; HR, hazard ratio; IPD, individual participant data.