Cristian Baicus is correct in pointing out the inconsistency in terms in our article.1 We intended to refer to relative risks in all instances. The type of analysis (univariate or multivariate) would not determine the type of value generated.
Madhu K. Natarajan Rizwan Afzal Division of Cardiology Population Health Research Institute McMaster University Hamilton, Ont.
Reference
- 1.↵