Prioritisation for Evidence Aid: choosing systematic reviews for the Evidence Aid database

Article type
Authors
Allen C1, Kayabu B2, Clarke M3
1Evidence Aid, UK
2Evidence Aid and Trinity College, Dublin, Ireland
3Evidence Aid and Queens University, Belfast, UK
Abstract
Background: Evidence Aid draws on knowledge from Cochrane Reviews and other systematic reviews which assess interventions/ actions which might have an impact on health in disaster settings.

Objectives: To identify Cochrane Reviews relevant to Evidence Aid, to provide reliable, up-to-date and timely access to evidence for natural disasters, healthcare emergencies or humanitarian crises.

Methods: In August 2012, The Cochrane Library contained 5168 full Cochrane Reviews and 2236 published Protocols. Each of these 7404 records was assessed to ascertain whether it might be relevant to Evidence Aid. For protocols and reviews published up January 2012, this assessment was done by three people working separately to categorise potentially eligible records as ‘High priority’, ‘Unsure’ or ‘Not Relevant’. Three Cochrane Review Groups were sent their selected records and provided feedback leading to the addition of a category of ‘Low priority’. Lists of records were then sent to all other Cochrane Review Groups for prioritisation. A more streamlined approach was adopted for the protocols and reviews published since the start of 2012, and the registered titles for a small number of Cochrane Reviews were also considered.

Results: All assessors categorised 135 records as ‘high priority’; prioritisation for 522 records was inconsistent. 91 records contained in the Special Collections were automatically marked as high priority and not assessed. This gives a total of 226 high priority records for the prototype for the Evidence Aid database which have since been assessed by the International Rescue Committee.

Conclusions: Prioritisation is important to Evidence Aid to ensure that the database contains relevant information. Evidence Aid will partner with aid agencies, NGOs and others to incorporate views and priorities. A workshop in 2013 will help to identify the highest priorities, using the approach developed and refined by the James Lind Alliance; results from this will be presented at the Colloquium.