I’m off in an hour to tape some episodes of Healthcare Triage. Two of them are on NNT and NNH. I’m on the radio – literally right now (it’s a break) – so I hadn’t planned on blogging this morning. But then Austin sent me an email on a new study on mammograms. No one trolls me like he does.
Objective: To evaluate the effectiveness of contemporary mammography screening using individual information about screening history and breast cancer mortality from public screening programmes.
Design: Prospective cohort study of Norwegian women who were followed between 1986 and 2009. Within that period (1995-2005), a national mammography screening programme was gradually implemented, with biennial invitations sent to women aged 50-69 years.
Participants: All Norwegian women aged 50-79 between 1986 and 2009.
Main outcome measures: Multiple Poisson regression analysis was used to estimate breast cancer mortality rate ratios comparing women who were invited to screening (intention to screen) with women who were not invited, with a clear distinction between cases of breast cancer diagnosed before (without potential for screening effect) and after (with potential for screening effect) the first invitation for screening. We took competing causes of death into account by censoring women from further follow-up who died from other causes. Based on the observed mortality reduction combined with the all cause and breast cancer specific mortality in Norway in 2009, we used the CISNET (Cancer Intervention and Surveillance Modeling Network) Stanford simulation model to estimate how many women would need to be invited to biennial mammography screening in the age group 50-69 years to prevent one breast cancer death during their lifetime.
My interest here isn’t to talk about the results of this study.* But if it’s your interest, it’s a large cohort study of women in Norway between 1986 and 2009. It’s not an RCT (although it’s closer than most cohort studies). It’s also a study of people “invited to participate” in a screening program, not one of actually receiving a mammogram.
What I want to discuss, though, is the difference between relative and absolute risk. The conclusion of the manuscript is this:
Conclusion: Invitation to modern mammography screening may reduce deaths from breast cancer by about 28%.
That sounds amazing. Who wouldn’t want this? But that’s a relative risk reduction. It doesn’t tell you how much you reduced your absolute risk.
To the author’s credit, they reported an NNT for invitations to the program. You’d need to invite 368 women to participate in order for one breast cancer death to be prevented over a lifetime.** That means the absolute risk reduction is (100/368) = 0.27%.
Sticking to the NNT, that means that of the 368 women you invite, 367 will see no benefit. And they could see harms! Overdiagnosis, extra procedures, expense, potential sequelae, etc. And it’s possible that the one women saved might die of other causes. We’ve covered that before.
But whats more concerning is that I’m sure the “28% reduction” will be in many media stories that cover this paper. Very few will mention that the results show that the screening program will reduce absolute risk by 0.27%. The former number will blow past anyone’s concerns about harms. The latter would likely make them think hard about whether a program is worth it. Especially if we discussed the NNT along with NNH.
* Really, I don’t want to. We have data from many large RCTs, so the continued value of observational data to estimate causality is debatable.
** To be honest, I can’t figure out how they calculated the 368. It’s just reported. I don’t see an absolute risk reduction in the manuscript. I emailed the authors for more information. If I hear back, I’ll update this post. If one of you can figure out how they calculated the 368 from the data in the paper, let me know!