News

“Citizens’ jury” disagrees over whether screening leaflet should put reassurance before accuracy

BMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e8047 (Published 26 November 2012) Cite this as: BMJ 2012;345:e8047
  1. Nigel Hawkes
  1. 1London

A “citizens’ jury” of 25 women, assembled this week to provide advice for the drafting of a new leaflet on breast cancer screening, has reached consensus on some of the tricky issues.

The leaflet is being rewritten after criticism that it conveyed a falsely optimistic message and in the light of the Marmot review of breast cancer screening, which found that women invited to mammographic screening were three times as likely to be “overdiagnosed” as they were to have their lives saved.1

The jury did not draft a new leaflet or the covering invitation letter that will accompany it but did consider some of the central issues to be considered by those who will draft them. For example, the jury preferred the term “overtreatment” to “overdiagnosis” by a majority of 21 to four, on the grounds that it was easier to understand, and also preferred expressing benefits in terms of lives saved rather than deaths avoided, though by a smaller majority.

One juror said that lives saved set a more upbeat message than deaths avoided. Put to the vote, 13 women favoured lives saved and three preferred deaths avoided, while eight believed that both terms should be used­­—and one neither.

The women, who had listened to two days of presentations about breast cancer and the screening programme, were broadly content with the terms benefits and risks, preferring them to alternatives such as pros and cons, but a substantial minority (seven) preferred “disadvantages” to “risks.”

Benefits and risks should both be measured by reference to the number of women actually screened, not the number invited to screening, the women said, and a clear majority favoured using the same denominator throughout the leaflet to express these benefits and risks. What that denominator should be was not so important to them as ensuring that it was used consistently, but the jury seemed to lean to 250 as the favoured number. Percentages were not favoured.

Defining what was meant by overdiagnosis, even if the term itself was not to be used, produced a consensus on the sentence, “Screening detects many cancers that may never have become a problem in the patient’s lifetime.”

There was agreement, though the question was not put to a vote, that the leaflet should begin by citing the numbers of lives saved by breast cancer screening, 1300 a year according to the Marmot review, followed by the caveat that a small number of women would suffer overtreatment after their diagnosis.

But which matters most, to reassure or to be accurate? The majority (15) wanted the leaflet to do both, while three opted for reassurance as the first priority and seven for accuracy. The drafters, led by Amanda Ramirez of King’s Health Partners, may not have found this advice quite so helpful. Joanne Rule, the former chief executive of the charity Cancerbackup, who chaired the discussions, acknowledged that on this point the jury did not reach consensus.

There was a similar division over how to list the benefits and harms. Should they be expressed together in the same sentence or separately? Four jurors voted for the first option, eight for the second, but 12 voted for firstly expressing the benefits and harms separately and then together. (One juror did not vote on this occasion.) The jury was, however, almost unanimous in its view that it was necessary to make it clear that there was uncertainty about benefits and risks both.

The women who made up the jury, volunteers recruited on the street or outside community centres by the Office for Public Management, spent two days listening to a series of witnesses.2

Malcolm Reed of Sheffield University explained the nature of breast cancer and its surgical treatment, while Alison Jones, a consultant oncologist at the Royal Free Hospital in London, covered non-surgical treatments and outcomes.

Patsy Whelehan, a radiographer from Dundee University, described the process of screening, and Mike Michell, clinical director of the breast screening programme at King’s College Hospital and National Training Centre, spelt out the nature of any subsequent diagnostic tests.

Angela Coulter, a health policy analyst who specialises in patient involvement, discussed the problems of communicating complex health issues, and David Spiegelhalter of Cambridge University outlined the challenges of conveying and displaying risks. Roger Felton, a graphic designer, discussed different design options for the leaflet.

There are many ways of displaying risks and benefits graphically, but the one that appealed most to the jury was a depiction of women who were screened as a series of dots or schematic figures arranged in an array, with those representing diagnosis and overdiagnosis highlighted in a different colour. This was preferred to a pie chart by a majority of 20 to five.

Ramirez said that the jury had exceeded expectations. “We’ve learnt more than we expected,” she said. “Now we need to work very hard to pull together the recommendations you’ve made to produce a new leaflet by next March.”

The process will involve a draft being sent out for expert comment, any modifications incorporated, then testing on a sample of women to make sure that they can understand it and that it makes sense to them.

Finally it will go to an advisory committee, chaired by Mike Richards, the national cancer director, which will be responsible for signing off the final document.

Notes

Cite this as: BMJ 2012;345:e8047

Footnotes

  • bmj.com News: Mammography results in substantial overdiagnosis of breast cancer, concludes study (BMJ 2012;345:e7910, doi:10.1136/bmj.e7910); Observations: The NHS breast screening programme needs independent review (BMJ 2011;343:d6894, doi:10.1136/bmj.d6894); Observations: An independent review is under way (BMJ 2011;343:d6843, doi:10.1136/bmj.d6843); Analysis: Screening for breast cancer—balancing the debate (BMJ 2010;340:c3106, doi:10.1136/bmj.c3106).

References