Article type
Abstract
Background: Forest plots are the principal tool for presenting the results of multiple studies addressing the same synthesis question in systematic reviews. The plots usually display study effect estimates and their confidence intervals and may display the meta-analysis estimate. Typically, studies are not displayed on the forest if it is not possible to calculate the effect estimate and its confidence interval. Yet, often, at least some information is known about these studies (eg, whether the direction of the intervention effect is favorable or harmful), and inclusion of this information may facilitate a more structured and complete assessment of the evidence available for a particular synthesis question.
Objectives: To develop a forest plot to display results from all included studies.
Methods: We have developed a prototype forest plot that provides a more complete display of results from studies eligible for each synthesis by including results for studies for which effect estimates (and confidence intervals) are not calculable. We will undertake testing to gather users’ feedback on their understanding of the plot, its usefulness, problems, and suggestions for improvement. Using a "think-aloud walkthrough" method, we will ask participants to articulate their thoughts as they try to interpret the information presented on the plot. Participants will be interviewed remotely. These interviews will be recorded, and an observer will take notes. The plot will be iteratively revised, first from interviews with methodologists and then with end users (approximately 6 of each).
Results: Our prototype forest plot presents available information from all studies eligible for a particular synthesis question (Figure). For studies wherein the effect estimate is not calculable, we display a range of plausible effect estimates consistent with the available information. We will present at the Summit findings from the user testing and the final forest plot.
Conclusion: Building on guidance in Chapter 12 of the Cochrane Handbook, our forest plot integrates all available information from studies eligible for a particular synthesis.
Relevance and importance to patients: Forest plots that display information from all studies may facilitate improved assessment of the evidence, with the potential to improve evidence underpinning health care decision-making for patients.
Objectives: To develop a forest plot to display results from all included studies.
Methods: We have developed a prototype forest plot that provides a more complete display of results from studies eligible for each synthesis by including results for studies for which effect estimates (and confidence intervals) are not calculable. We will undertake testing to gather users’ feedback on their understanding of the plot, its usefulness, problems, and suggestions for improvement. Using a "think-aloud walkthrough" method, we will ask participants to articulate their thoughts as they try to interpret the information presented on the plot. Participants will be interviewed remotely. These interviews will be recorded, and an observer will take notes. The plot will be iteratively revised, first from interviews with methodologists and then with end users (approximately 6 of each).
Results: Our prototype forest plot presents available information from all studies eligible for a particular synthesis question (Figure). For studies wherein the effect estimate is not calculable, we display a range of plausible effect estimates consistent with the available information. We will present at the Summit findings from the user testing and the final forest plot.
Conclusion: Building on guidance in Chapter 12 of the Cochrane Handbook, our forest plot integrates all available information from studies eligible for a particular synthesis.
Relevance and importance to patients: Forest plots that display information from all studies may facilitate improved assessment of the evidence, with the potential to improve evidence underpinning health care decision-making for patients.