Update: November 2015 BMJ Cates Plot available here 'Clinical importance cannot be ruled out using mean difference alone'
Since 1999 Visual Rx has been used to create smiley face plots to visually communicate the risks and benefits of treatments. Notable current uses include a series of articles from the National Prescribing Centre, including a recent patient information guide about the lipid modification in response to the following NICE guidance:
“People should be offered information about their absolute risk of CVD and about the absolute benefits and harms of an intervention over a 10-year period. This information should be in a form that:
- presents individualised risk and benefit scenarios
- presents the absolute risk of events numerically
- uses appropriate diagrams and text.”
Visual Rx has also been used to produce book illustrations (for example Evidence Based Rheumatology) and NNT illustrations in Cochrane Airways Group reviews.
The smiley face diagram made its first appearance in the BMJ in an article communicating risk; since then we changed the colours of the faces to red, yellow and green as red and purple are not easy to distinguish for those who are colour blind, and it seemed that traffic light colours might be more familiar to users.
There is also an interesting recent article on communicating risk by David Spiegelhalter in the Annals of Family Medicine in which he refers to Visual Rx as one of the ways of trying to make risks accessible to patients. Links to many further examples are available here.
With an increasing number of publishers and individuals reproducing the images we feel it necessary to start referring to them consistently, and for want of a better name we are going to call it a Cates Plot!
The essence of the current Cates Plot is the use of 4 face categories to visually indicate the following:
- People not affected by a treatment (green faces for those with a good outcome and red for those with a bad outcome)
- People for which treatment changes their category from a bad outcome to a good outcome (yellow faces)
- People for which treatment causes an adverse event and changes their category from a good outcome to bad outcome (crossed out green faces)
This way of demonstrating treatment effects is in line with the Summary of Findings (SoF) Tables that are being introduced to Cochrane reviews. Theses SoF Tables compare the risks of an event in 100 or 1000 patients who receive treatment with the same number who are not treated. The concept used in Visual Rx of using pooled Odds Ratios or Relative Risks with the associated confidence interval to calculate an absolute effect of treatment for particular baseline risks have been adopted in the SoF tables, but the tables do not yet contain a visual display, such as the Cates Plot. Visual Rx originally produced a 100 face Cates Plot but we have updated Visual Rx to optionaly display 1000 faces in response to requests from those dealing with rarer events.
We hope that the concept will continue to prove useful in explaining risks and benefits in absolute terms (i.e. the number that benefit, or are harmed out of 100 or 1000 treated). Note that it is important to include a timescale for the risk (10 years in the case of the CVS risks), as this will alter the numbers who benefit or suffer harm.