Editor’s response

CMAJ is grateful for the reminder that what matters to a patient is the absolute risk.1 We should have made this easier for readers to find, especially in the second article to which Shaw refers.

The total mortality on optimal medical therapy was easy to see in Wells and colleagues’ article;2 in Figure 2 it was 250/1013, or 24.5%. The absolute risk of cancer was less easy to find in Eisenberg and associates’ article.3 One estimate might be 12 020 cancers diagnosed in 82 861 patients, as reported in the abstract; these occurred over an average follow-up time of five years according to the results, which suggests about 2.9 cases per 100 person years.

What these summary absolute risks hide, however, is that they may not be appropriate for an individual patient. Risk may vary with characteristics such as age, sex, severity of illness, comorbidity and family history. There is a long tradition of debate about whether to present research findings in terms of relative measures, which tend to be more stable between patient populations, or absolute measures, which have more immediate interpretability for clinicians and patients.4^,5 We prefer to have both where possible, and we will make renewed efforts to remind authors to provide them.