Background

Systematic reviews often assess multiple outcomes (benefits and harm). An ongoing challenge is how to present results for multiple outcomes in a clear and concise way, in order to facilitate judgments of the overall balance of benefits and harms. Dot plots are a data visualization method to display findings across multiple outcomes in a visually pleasing format. In the March 2021 web conference titled “Methods Symposium: Advanced Methods and Innovative Technologies for Evidence Synthesis” funded by the Agency for Healthcare Research and Quality (AHRQ; R13HS027701), focusing on data visualization for adverse events in randomized trials, Dr. Rachel Phillips presented a dot plot as an example data visualization method.¹ In her example, the dot plot summarized serious adverse events from an individual randomized trial.² The dot plot shows the relative risk with 95 percent confidence intervals as well as the absolute rates in each arm. Having both outcomes in the same plot is useful for interpreting the magnitude of effect. The dot plot also shows the number of adverse events in each group.

A Stata module developed by Phillips and colleagues is available to produce such dot plots (https://ideas.repec.org/c/boc/bocode/s458735.html).³ However, the Stata module developed by Phillips and colleagues is intended for display of data from an individual study; it is not designed for use with pooled data. We were unable to locate a module to produce dot plots for pooled data in Stata or another statistical package. In a personal communication, Dr. Phillips stated that she was not aware of such a module being available. In addition, dot plots do not need to be restricted to harms and also could be adapted to summarize findings for continuous as well as dichotomous outcomes. Although commercially available data visualization software can produce plots that display findings for multiple outcomes, this requires purchasing/having the software and uploading the data; in addition, the default plots in currently available data visualization software packages do not display the data (e.g., both the absolute rates and relative risks) in the same format as the dot plots. The availability of a statistical package module to easily produce dot plots would enable Evidence-based Practice Centers (EPC) and other systematic reviewers to more easily include such figures in reports and other products summarizing the findings for multiple outcomes in a single figure, without having to use additional commercial software. This would enhance the usability of EPC and other systematic reviews.

Objective

The purpose of this methods project was to develop a tool for a standard statistical package (Stata) to create dot plots to summarize pooled data for multiple outcomes in systematic reviews. Our tool could be used for outcomes that are benefits as well as harms. We also attempted to adapt the dot plots to display pooled data for continuous outcomes.