I took up the challenge of producing an infographic for a Cochrane review, and decided to use the recently updated review: External cephalic version for breech presentation at term. My aim was to produce a visual outline of the summary of findings table and plain language summary for a lay audience. In this blogpost, I will focus on the process of creating meaningful visualisation of the data. This is the area that I found most challenging when producing the infographic, and I would like to explore how it can be done effectively. It’s difficult to strike the balance between simple communication and research complexity: It’s impossible to include every detail of a review in an infographic, but removing too much detail can distort the message of the original review beyond recognition.
Developing the infographic
The raw materials you start with are critical to the quality and clarity of the infographic you’re able to produce. The review I chose had an excellent plain language summary, written by Dr Elizabeth Wager. The review had clear conclusions that could be represented visually: two outcomes in the summary of findings table showed differences that could be turned into pictographs (vaginal cephalic birth not achieved and caesarean section), the other outcomes showed no difference between groups so could be summarised briefly (Apgar score <7 at 5 minutes, umbilical vein pH < 7.20, neonatal admission, and perinatal death). Where there are multiple comparisons, lots of outcomes, or no clear findings it becomes much more difficult to summarise the review visually. I was lucky that this review fitted the format of an infographic!
I used Piktochart, a free online tool, to product the infographic. Piktochart allows users to input data and produce graphs and charts. This function was very useful for producing the arrays of babies and women for a pictograph, which could be easily edited. I was mindful of sticking to the Cochrane brand guidelines, and keeping the layout clean and uncluttered. The design was influenced by Carlos Cuello’s infographic Probiotics and necrotizing enterocolitis in preterm babies, presented on this blog in September.
The infographic of ECV at term
Challenges in data visualisation
I faced two major issues with turning the review data on mode of delivery into a pictograph: How to accurately present meta-analysis in frequencies, and how to communicate uncertainty in the results. The review presents “vaginal cephalic birth not achieved” (phrased negatively to fit the conventions of RevMan). However, from a pregnant woman’s perspective “head first vaginal birth” is probably a more important and interesting outcome. Logically “not head-first vaginal birth” and “head-first vaginal birth” should add up to all births…. As all babies are born either by caesarean section, vaginal head-first delivery, or vaginal breech delivery, the neatest way to present the results of the review would be as a single grid for “type of birth” comparing these outcomes. Again, logically the three types of birth should add up to all births… However, my numbers didn’t add to up 100 out of 100 babies, as illustrated below:
After checking the numbers (several times!), I worked out why they didn’t add up. In a meta-analysis, the weighting assigned to studies is affected by the number of events for an outcome, as well as the size of the study. Therefore, switching an outcome around can completely change the weighting, and the final relative effect.
I wanted to present the outcomes that pregnant women would be most interested in, but also for the infographic to be consistent with the review. I decided to show the outcomes presented in the review as two separate outcomes (caesarean section and head first vaginal birth not achieved). I used colour to highlight the positive outcome “head first vaginal birth” in magenta, against light grey for “not head first vaginal birth” from the review, and framing the outcome both ways in the text (for example. “79 babies out of 100 did not have a head-first vaginal birth” and “21 babies out of 100 had a head-first vaginal birth”). To differentiate between the outcomes caesarean section and head first vaginal birth not achieved I used women icons in the caesarean pictograph, and baby icons in the head-first delivery pictograph.
I wasn’t sure whether and how to include confidence intervals on the infographic. Confidence intervals are important in the reporting of review findings, and affect our certainty in the result. However, they may not be well understood by the public and it’s not easy to present them visually without over-complicating the message of the infographic. I experimented with a few options, but decided to take the same approach as Carlos Cuello, including the confidence interval in the text and relating it to the role of chance.
Helen West, Cochrane Pregnancy and Childbirth
I’d welcome feedback on this infographic. In particular, I’d like to hear your views on the issues relating to data visualisation:
• how to get the right balance between simple communication and research complexity;
• how to accurately present meta-analysis in frequencies;
• how (and if) to include confidence intervals.
[Helen West is supported by an National Institute for Health Research (NIHR) Cochrane programme Grant (13/89/05) and Cochrane infrastructure funding to Cochrane Pregnancy and Childbirth. The views and opinions expressed herein are those of the author and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, NHS or the Department of Health.]