This research paper by Valgimigli and colleagues (BMJ 2021;373:n1332; doi:10.1136/bmj.n1332) used data from publications that have since been corrected. Haemoglobin values were not reported correctly in the GLASSY trial, so data on haemoglobin in table 1 and supplementary table E of the paper have been corrected. The PRECISE-DAPT score was not correctly reported in the GLASSY trial because the score was calculated using haemoglobin, so data on the PRECISE-DAPT score in table 1, figure 2, supplementary table E, and supplementary figures C, D, F, and G have been corrected. Four procedural variables were not correctly reported. “Number of vessels treated at index PCI,” “Number of lesions treated at index PCI,” and “Bifurcation lesion treated with at least two stents” have been corrected in the GLASSY trial. “Number of vessels treated at index PCI” has been corrected in the STOPDAPT-2 trial. “Bifurcation lesion treated with at least two stents” and “TIMI pre-PCI 0-1” have been corrected in the TWILIGHT trial. Accordingly, data in supplementary tables F and G have been corrected. In addition, data on the “Complex PCI” subgroup have been corrected in figure 2 and supplementary figures C, D, F, and G. These corrections were necessary because the “Complex PCI” subgroup was defined based on procedural variables (see above).