065 Pubmed vs. Google Scholar: A Database Arms Race?

Abstract

Background The patient perspective in guideline development is of vital importance. To find out what this perspective entails, different methods may be considered, such as participation of patients or their representatives in guideline development groups or in focus group discussions, or by conducting patient surveys addressing specific problems and needs. In addition, a review of the literature in the early stages of guideline development can provide relevant information. Literature search filtres for patients' perspectives and preferences applicable for Medline (OVID), PubMed, and Embase were developed and validated in 2012. The specificity was 98% but the sensitivity was only 90%. Objectives To verify the sensitivity of the filtres by means of a newly available 'gold standard'. Methods We re-estimated the sensitivity of the search filtres by using the references of a recent Cochrane Review, Interventions for providers to promote a patient-centred approach in clinical consultations 2012;(12):CD003267, as a gold standard. Results The search filtres for patients' values and preferences retrieved 72 (Medline (OVID/Pubmed) and 67 (Embase) titles, respectively, out of 73 references included in the Cochrane Review (mean sensitivity 96%). Discussion Applying filtres for patients' perspectives and preferences retrieved almost all references. Minor adaptations to the Embase filtre were needed to enhance the sensitivity without compromising the specificity. Validation of filtres is an iterative process, illustrating that filtres are dynamic tools.