Text

Background: Harvest plots are a graphical method for the presentation of complex and heterogeneous evidence in health-related evidence syntheses. Their design and use are not standardized.

Aim/scientific question: To provide a systematic overview of the use and methodological characteristics of harvest plots in the context of health-related evidence syntheses.

Material & Methods: We used a meta-research approach to capture existing research and publication practices. The study protocol was prospectively registered on Open Science Framework (10.17605/OSF.IO/75V6X). We included evidence syntheses on a health-related research question that used and reported a harvest plot for data analysis and synthesis. The search was carried out in PubMed and Embase. We performed a forward citation search of the original 2008 publication, in which the harvest plot was incepted. The selection and extraction of data was mainly carried out by one reviewer and double-checked by a second reviewer, with no formal quality assurance of the included studies.

Results: We included 201 harvest plots to our analysis, published between 2008 and 2025. Study characteristics show that narrative harvest plot descriptions were mainly provided as part of the figure (n=191; 95%). The design of the studies that used a harvest plot was mainly a systematic review (n=186; 93%) including interventional (n=122; 61%) and observational (n=103; 51%) studies. The included harvest plots mainly provide information about intervention or exposure in the row (n=80; 40%), effect direction in the column (n=170; 85%). Bar color indicates mainly study quality (n=27; 13%) or outcome measurement (n=25; 12%). General top or bottom information mainly indicated a study reference (n=114; 57%).

Discussion: Our results revealed large heterogeneity and a lack of standardization of the design, illustration, and characteristics when providing a harvest plot in health-related evidence synthesis. Methods guidance in combination with a digitalized tool or program code and reporting guidance may contribute to a higher degree of standardization and facilitate interpretation of harvest plots.

Take Home Message: Although harvest plots are being increasingly used, their design varies widely. They are usually limited to just a few dimensions, such as the effect direction, intervention characteristics, and/or the study quality, even though the method offers more possibilities. There is an urgent need to harmonize and standardize through methodological guidance and reporting standard, preferably supported by digital tools.

Competing interests: JH and GM authored publications eligible to this meta-research study. All authors declare no conflict of interest.