When writing results and conclusions of a systematic review, there are many ways to communicate the results, and authors use a variety of phrases. Phrases such as “possibly large but non-significant results” can be confusing to readers and users. There are a limited number of approaches to wording the results, and they appear only to focus on effect size and statistical significance.
To develop standard narrative statements that can be used to write the results and conclusions of a systematic review; and to evaluate the acceptability of the standard narrative statements to producers and users of systematic reviews.
We based this work on research previously conducted with consumers to develop standard wording in plain language summaries. Interviews, and a randomised controlled trial were conducted, and the preliminary list of standard statements was well-understood. We presented the standard statements to review authors and guideline developers in three workshop focus groups. We provided five examples to participants which included the analysis for an outcome from a systematic review, the certainty of evidence (using the GRADE approach), and the importance of the effect to key stakeholders. We asked participants if the proposed wording was acceptable and understandable. We then conducted an electronic survey of key stakeholders via SurveyMonkey, including statisticians, review authors, guideline developers, and healthcare professionals. Again, we provided multiple examples for feedback about acceptability and ease of understanding.
The preliminary list was revised with feedback from over 50 people who attended workshops. Participants agreed that statements should be based on certainty of evidence. However, they were concerned about making decisions about the importance of an effect and the limited choices for wording. We developed suggestions for how to determine the importance, and for more options.The survey results will inform the final list.
We will have a list of narrative statements that authors of knowledge syntheses can use to communicate results. The list will include multiple options that are acceptable and understandable to producers and users.
This research builds on work for writing plain language summaries.