prescoping and search advisor

Prescoping is the process of doing quick and more basic searches to establish a good overview and understanding of the evidence base relevant to your research question. You should keep a brief record of which key words you search for, what type of relevant hits they return and any irrelevant hits linked to each key word. While this can take time, it is often very valuable later on as it enables you to have confidence that your research question does have a sensible evidence base to review and that your analysis plans are likely to work. Try to think about being able to these questions from your prescoping:

• Is there evidence available on the research question?

• How large is the evidence base for the research question?

• What type of studies are available? RCT/CT, within participant, case studies, technical reports, quant and/or qual data?

• What key words are used by the articles that are useful to the research questions?

• What key words are more efficient (bring up useful hits with low number of irrelevant results) and are there any terms that are safe not to use (inefficient and any relevant hits are found using more efficient key words)?

• What are they interventions/populations/controls/comparisons/outcomes used in the evidence base?

• Are there gold standards in the relevant field, or is there uncertainty?

Overall you are trying to inform decisions around what is a sensible research question, establishing an efficient search strategy and ensuring your overall scope answers the main question, is not overly complex and will be managable within your resources.

Check out the data sources guide for ideas on where to search.

Remember that by protocol planning you should have a good idea of your review's research question and scope.

By this point you should have a good idea of the question and scope of your review. If you need to, you can do more prescoping of the final question to help inform the final systematic search strategy and inclusion criteria. In this stage you need to be confident that you can justify decisions around your inlcuded populations, interventions, control/comparisons, outcomes and study types. You might also want to check you are using all the needed key words for your final search and that your search strands will all be efficient. This is also a last chance to check the search details before you set and run your full search, so it is the last chance to find and fix problems ahead of the main review. Remember that most problems are easier to fix at planning and protocol stage than once the review has started, so it is worth investing time to do thorough pre-scoping.

Some reviews will sift through 1000s of hits, so don't panic at first if the hits a higher than your might expect if this is your first review. The number of hits links to the research question and scope set during planning and protocol development and should be based on thorough prescoping to avoid the number of hits during the main search being a surprise. If you think you have a problem, you might want to revisit the protocol step, conduct more prescoping and develop a more refined question, scope and search strategy. It is difficult to define what is "too many" in this context. Here are some questions to think about if you are feeling there are more hits than you can cope with.

1. Is your question and scope sensible?

   1. Is your question too broad? Do you need to refine it? Should it be more than one review?

2. Is your search effcient?

   1. Did you check for key terms that return a high number of irrelevant hits? Any terms that do so should be checked for if they return any relevant hits and if those same hits are returned by a more efficient term. If a more efficient term returns all the relevant hits of an ineffcient term then you can justify not using inefficient terms to help control hits.

3. Are you looking in too many simulations?

   1. If you have astronaut data, then do you also need to go into simulation environments?

   2. If you have human data, then do you also need to go into animal models?

4. Is the evidence base simply large and so a large number of hits is to be expected?

   1. If you have what you feel and can justify is single research question with a reasonable scope and an efficient search, then the evidence base and scope may be large and the number of hits might be appropriate. However, it could be worth checking with your review team's topic area expert and methods expert to boost confidence.

5. What if the number of hits is too much for our resources but the question and scope is good?

   1. This is difficult to answer as in an ideal world you should try to ensure you have the resource available to conduct the review. Remember you can likely do initial screening on title and abstract alone and systems such as Rayyan can help you search for key inclusion and exclusion words that can help speed up the screening process. If you begin to consider initial screening by title only instead of title and abstract then beware that the quality of the review could be lowered and you need to be completely transparent in your methods how you performed screening and might need to justify decisions. The Cochrane handbook chapter on searching for and selecting studies should also be consulted to help you make the best decisions.

For this question you are best reading the global methods guide, the Cochrane handbook. Chapter four provides advice on what a systematic search is. Remember to check the data sources guide for ideas on where you might want to search for space biomedicine.