Visual Abstract Generator

ByYuki Furukawa

The Visual Abstract Generator builds a one-page graphical summary of a study from a form or from a short YAML block. It is designed around the structure of a systematic review: population, intervention versus comparator, and a summary-of-findings table in which each outcome carries an effect estimate and a GRADE certainty rating. Everything is drawn in your browser and exported as PNG or PDF.

How to use

There are two ways in. Fill in the form field by field, or paste Markdown with a YAML front matter block: everything between the opening and closing --- lines is read into the form, and anything below it is kept as a free-text note. The YAML route is more convenient if you produce several abstracts in one house style, since a file can be copied and edited rather than re-entered.

  • Study details — title (up to 120 characters), authors, journal, year, DOI, and the name of the person who made the figure, printed in the footer.
  • Summary — the key finding (up to 250 characters) and, optionally, limitations, placed at the foot of the canvas with a warning mark.
  • Design — study type, a detail line such as "19 RCTs", the number of trials and the number of participants.
  • Population and comparison — descriptions, detail lines for age or sex distribution, short labels for intervention and control, and images with background removal and cropping within the slot.
  • Outcomes — up to five, shown as a summary-of-findings table. Each takes a binary or continuous type, a risk ratio or odds ratio, a direction of effect (favours intervention, favours control, or unclear), a GRADE certainty rating (high, moderate, low or very low), control and intervention event rates written as "24 per 100", and a primary-outcome flag.
  • Custom sections — up to three, for anything the standard fields do not cover.

Instead of the table you can drop in a figure — a forest plot or flow diagram exported from your analysis software — which occupies the same slot. A doughnut chart is a third option, but only under the JAMA-like template; with the BMJ-like template the choice is between the summary-of-findings table and a figure. Templates come in those two styles with rounded corners optional; palettes are Classic, Crimson, Forest, Violet, Slate or a custom set of primary, intervention and control colours; fonts are Calibri, Inter, Arial, Times New Roman or Georgia. A preview toggle switches between desktop and phone widths, worth checking before posting. PNG export has three sizes — SNS at 1024 px, SNS HD at 2048 px, and Journal at 300 DPI — and PDF export is either square at 90 × 90 mm or A4 portrait.

Method and primary sources

The layout mirrors a GRADE summary-of-findings table rather than a generic infographic. Each outcome row carries an effect measure, the absolute event rates in both arms and a certainty rating, which is what separates a visual abstract of a systematic review from a decorative summary of its conclusions: the reader sees not only the direction of the effect but how much confidence to place in it. Capping the table at five outcomes and asking you to mark one as primary follows the same reasoning.

The tool derives none of these numbers. It draws what you supply, so effect estimates should come from your analysis and certainty ratings from your GRADE assessment; for the imprecision domain, the OIS Calculator on this site implements the thresholds of GRADE Guidance 34 [2]. A visual abstract also does not replace the abstract itself: PRISMA 2020 [1] specifies what a systematic review abstract must report, and the figure should be consistent with it. A figure showing only the favourable outcomes of a review whose abstract reports mixed results is a reporting problem, not a design choice.

How to cite

A visual abstract is a figure, so it is normally credited in the caption or the post rather than in a reference list. A caption might read:

Visual abstract prepared with the Visual Abstract Generator (Furukawa Y. Visual Abstract Generator. https://yukifurukawa.jp/va/). Effect estimates and certainty ratings are taken from [source study]; certainty was rated using GRADE.

Always name the study the figure summarises and, if it is not your own work, make clear that the figure is a third-party summary; the creator field prints your name in the footer, which handles this on the image itself. The tool is maintained by Yuki Furukawa, a psychiatrist specialising in insomnia who works on systematic reviews and meta-analysis.

Limitations

  • This is a drawing tool. It does not calculate effect sizes and does not rate certainty; both are entered by you, and an error in the input becomes an error in the figure.
  • At most five outcomes and three custom sections. Reviews with many outcomes need a selection, made on the principles you used to choose primary outcomes in the protocol.
  • Binary effects are expressed as risk ratios or odds ratios; other measures have to be conveyed in the detail text.
  • The templates are styled after journal conventions but are not endorsed by any journal, and meeting a specific journal's figure requirements remains your responsibility.
  • Images you add are your responsibility with respect to licensing and consent for publication; the tool places them in the figure but cannot check their provenance.
  • Compression is the main risk to accuracy: a visual abstract has room for a headline finding, not for caveats, which is why the limitations field exists and is worth using.

References

  1. Page MJ, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. PubMed
  2. Zeng L, et al. GRADE Guidance 34: update on rating imprecision using a minimally contextualized approach. J Clin Epidemiol. 2022;150:216-224. PubMed