Creating succinct and digestible syntheses of data in systematic reviews of complex interventions can be a struggle. Details of the included studies are usually reported in tables and web appendices with an additional summary in the text of the paper. It could be beneficial to readers of some reviews, for example those with complex interventions or those with a broad review question, if the characteristics of the studies were presented graphically, so readers could understand the range and variety of included study and type of intervention quickly.
To use a graphical figure to present data describing the included studies for a systematic review on interventions for reducing the effects of financial hardship on mental health.
I prepared bar charts to display descriptive numerical data, e.g. age and duration of unemployment. I used square bars to represent relative sample size and included pictorial illustrations of study population features. I assembled these visual elements and overlaid them to create a single figure of the main descriptors of the included studies.
The new figure illustrated the variability of the studies in terms of participant type, study size and the intervention (Figure 1). For example, it was clear there were six interventions and that one, job-club, had been trialed five times with large variation in sample sizes (range 16 to 1771). The figure was published in the journal Psychological Medicine, entered for a competition on data visualisation, and tweeted and retweeted on Twitter, which increased the altmetrics. A figure to summarise the studies included in a systematic review of interventions for children exposed to domestic violence was also prepared for use in presentations.
While tables with complete details describing the included studies are necessary for the accurate reporting of systematic reviews and to comply with PRISMA reporting guidelines. Graphical visualization of study data is a useful addition to reviews of complex interventions or with broad inclusion criteria. They can help readers quickly grasp the main elements of the included studies and the image can be tweeted with a link to the paper, to improve dissemination.
Patient or healthcare consumer involvement: None yet, as this was a pilot idea.