Developing the Systematic Review Team
Establish a team with appropriate expertise and experience to conduct the systematic review.
Information adapted from Institute of Medicine's Standards for Systematic Reviews.
It's helpful to create and sign a memorandum of understanding (MOU) for everyone to outline expectations at the beginning. Here's an MOU template that can be adapted for your uses.
Developing the Question
To develop your question, start with an interest, then complete a few scope searches in relevant databases to see the breadth of research out there on your topic or on an adjacent topic. When you are familiar with the amount of research out there, you can then narrow or broaden your question. For example, a systematic review on diabetes intervention methods' effectiveness would be much too broad; a systematic review of randomized controlled trials on the use of EHRs for CAUTI prevention is too narrow (or can be adapted). The question will also depend on where you plan to publish. Publishing a Cochrane review with zero included studies is perfectly normal, but most journals would like to see a good number of included studies.
In an effort to let others know the research you are working on, you can register your protocol. Cochrane reviews and Joanna Briggs Institute have all of their protocols registered through their system. If you are not performing a Cochrane or JBI review, you can register your protocol with PROSPERO. Many journals also publish protocols.
Decide as a team what tools to use to stay organized. If your team includes people to whom you do not have physical access.
Documentation of the entire process is critical. Be sure to identify what needs to be documented and how you will do so before beginning. The goal is to keep records in the most systematic way possible so that all of your work can be reproduced. That means you should keep detailed records of the exact search you used for each database and that all your searches should have an end date so that the results can be reproduced exactly every time.