More than just a search: using past search strategies and search summary tables to improve future identification of implementation studies

Article type
Year
Authors
Rogers M1, Bethel A1, Thompson Coon J1, Abbott R1, Lang I1
1University of Exeter, UK
Abstract
Background: Implementation science is an emerging field for which the parameters and boundaries are still being constructed. This lack of clarity means that a common language is lacking and reporting is often poor, making it hard for studies to be located. In recent years there has been a rapid growth in the number of implementation studies. A recent update search for systematic reviews of implementation strategies in healthcare returned as many records from 2014 to 2016 as the original search, which covered the previous 13 years. Locating these studies is challenging: there is a large variation in terminology amongst authors and implementation science experts; there is no particular database where these types of studies are likely to be located and there is often disagreement about what constitutes an implementation study, leading to confusion about what terms should be included in the search strategy

Objectives: To analyse the terms used to describe or enable implementation in health settings, within published literature, in order to inform future search strategies for these types of study.

Methods: Using three completed searches for implementation studies in healthcare, care homes and dementia for systematic scoping reviews, we examined the impact of the terms used in each search strategy against the studies included in the reviews. We completed search summary tables for the three reviews to identify which resources were the most effective for locating and returning implementation studies

Results: A total of 13,965 titles and abstracts were screened, resulting in 684 full text articles for inclusion across the three projects. Analysis of the search results indicates that some search terms and controlled vocabulary were more effective than others in retrieving the included implementation studies. The search summary tables indicate which databases are most likely to hold the relevant literature.

Conclusions:This work will provide evidence towards how search strategies for locating implementation should be constructed, and provide guidance on how to search for implementation studies in the future in terms of resources and terminology.