A true systematic review explores a narrowly defined question through a reproducible and comprehensive search of the literature (published and unpublished) with no language or date restrictions. It is expected that teams will perform quality assessments, risk of bias analyses, and usually a meta analysis. Systematic Reviews are utilized to change real world practices through acquiring and synthesizing the most complete and robust evidentiary base to make evidence-based and informed decisions.
According to IOM, PRISMA, and Cochrane standards, to write a truly comprehensive systematic review, librarians should be involved in designing the search, executing the search, compiling the citations, creating a search strategy appendix, and writing the appropriate portion of the methods section.
If the systematic review methodology is not appropriate for your question or the breadth and depth of the research you intend to undertake, there are other review types that may be better suited for your purposes.
Conducting a Systematic Review
Systematic Reviews require a large investment of time and human labor to complete - often at least a year and a team of 3-6 people. The IOM and PRISMA guidelines list key components in detail (see "Guidelines & Best Practices", but, in sum, you should:
1. establish a team including experts in the needed clinical content area, review and searching methodologies, and content analysis. Consider using the IOM/HMD guidelines to focus your planning of the SR life cycle.
2. Manage bias with financial, professional, and other conflict of interest disclosures.
3. Formulate a topic including rationale for the need for a SR, develop a clinical question, and refine the rationale and question based on user input.
4. Create an SR protocol based on recognized standards like the PRISMA guidelines that include pertinent details such as your question, inclusion/exclusion criteria, analytical processes, and outcome measures.
5. Submit the protocol to an established SR registrar like PROSPERO (a multi-disciplinary registry including health sciences and social sciences).
6. Work with an expert in searching to create, refine, and perform the literature review in consultation with your team members - recognizing that this is an iterative process and often takes months to complete. The search is translated and documented for each database including the date searched to ensure maximum reproducibility.
7. Once the final search is run and results compiled you will need to perform blinded assessments of the articles for inclusion in the SR based on your pre-determined inclusion/exclusion criteria listed in the protocol. You will need to perform a statistical analysis of your assessments to reduce bias and ensure reviewer inter-reliability.
8. Include tables in your SR including the literature search flow diagram, list of included articles with rationale, and any other relevant analysis.