Systematic Reviews (in the Health Sciences)

Publishing a systematic review is similar to publishing any other research article: you need to have a plan for how to publish before starting the project.

• Look for potential journals to which you would like to submit, and use the editorial policies of these potential journals to help scope and shape the research project. Not all journals accept systematic review manuscripts. Some journals will have specific rules that may help you make choices about your systematic review project.
• Systematic review manuscripts are required to include more detailed information about their methodologies as compared to other research methods, so authors typically need to record (write down) methodological decisions made and the justifications for these decisions throughout the project planning and conduct stages. Inclusion of this information allows readers to judge the methodological strength of the overall study. Instead of making their own reporting standards, many journals rely on outside standards. One common standard is produced by PRISMA, and this can be a useful standard to follow during your project development phase if unsure which standard you'll need to use to write. The PRISMA checklist and documentation available on the standards tab provides explicit instructions for what information to retain during the process of planning and documenting your review, and how to format this information in your manuscript.

One of the largest and most common challenges in writing a systematic review manuscript for publication that most journal's editorial policies do not allow authors to follow PRISMA standards for documentation. PRISMA standards require that decisions and rationales be presented in the text alongside the results; this can take several hundred to several thousand words. Systematic review results are thorough, in-depth, and nuanced; this also requires significant lengths of text and tables, graphs, or images to express. Most journals have text length limits of 5,000-8,000 words and 1-5 figures. It's simply not possible to provide both the methods and the results in detail in the same manuscript.

Authors are faced with a difficult choice and will need to choose whether to follow standards and document the whole project, or follow a journal's policies and exclude vital information. Some options to consider that strike a balance:

• Contact an editor prior to submitting a manuscript to ask about policy exceptions to a manuscript. Perhaps a few extra hundred words or an extra figure could be permitted for a manuscript describing a complex project.
• Contact an editor prior to submission to ask if an online-only appendix could be included for the manuscript. This online appendix could explain the methods in great detail and leave the bulk of the manuscript to discuss the results. Readers could view both files to get a full picture of your work.
• Consider building your own website or finding a free website like Open Science Framework that can host files for you. You can post a file describing the methods in detail, then include the URL for this website in your manuscript alongside your results. This does mean you need to maintain this website in perpetuity or select a free site that is likely to remain active in the long term. Additionally, not all journals allow inclusion of URLs in manuscripts, so review the journal policies carefully or contact an editor to determine if this is possible.
• Consider making the corresponding author responsible for maintaining a copy of the methods. Write the manuscript, and include only the general methods. Add a sentence explaining that full methods can be obtained by contacting the corresponding author.

Many journals will instruct systematic review authors to "document the search" or explain what citation databases, registers, websites, etc., were searched and how these searches were conducted. The goal of documenting a search is twofold: first, to allow readers to assess whether the sources consulted were appropriate and if the searches were conducted in a thorough and detailed manner; and second, to allow a reader to replicate the search themselves (to either replicate the research, or extend the research question to include results from a new timeframe, new geographic area, new language of publication, new age group of subjects, etc.).

Any documented search needs to include these elements:

• Names of all databases, registers, websites, etc searched
• For citation databases, the name of the platform on which the database was searched
• Date(s) the searches were conducted
• Exact key words or subject headings included in the search strings
• All Boolean logic statements (AND, OR, NOT) and all parentheses, quotation marks, commas, and other punctuation marks included in the search strings
• Any limits or filters applied to limit your search to specific dates, languages, publication types, or other characteristics of the publication or the characteristics of individuals involved in the research (their ages, their genders).
• How many results were exported from each database, register, or website searched.

There are several ways to document a search-- in text, as a figure, in a table format-- and typically a journal's policies will dictate the formatting in the final manuscript.

Watch this tutorial to learn how to document a search to PRISMA standards.