Article type
Year
Abstract
Background: Writing efficient search strategies for systematic reviews requires defining, a priori, the topics of a review within a search. A difficulty arises when a population or set of interventions is unclear or unknown: how best to operationalise these points in the context of the search?
Objectives: This paper presents a search designed for qualitative evidence on 'hard-to-reach’ groups at risk of tuberculosis. It explores the methods we used to construct a search filter for this ill-defined population and whether the collaborative approach we used was an effective method for defining difficult-to-define concepts.
Methods: Information specialists extensively tested potential terms and themes in a variety of resources and drafted the filter through database searching, citation chasing, expert/advisory contact, grey literature searching and working collaboratively with the review team. Hard-to-define populations were then analysed as a combination of specific risk groups (e.g. prisoners) coupled with risk factors (e.g. drug use) to ensure a highly sensitive population filter. The original search combined three themes (TB + Population + Qualitative terms). To test the efficiency of the filter, we re-ran our original search in Medline without the population filter. The results of the new search (TB + Qualitative terms) were screened to see if any populations were retrieved which had not been defined in the filter.
Results: The new search yielded 5454 studies not located in the initial search. No additional populations were identified after screening the extra studies. The sensitive population filter appeared to capture all relevant populations, suggesting that the time invested in creating the search strengthened the conceptual understanding of the population under review.
Conclusions: Dialogue between information specialists, review teams and topic experts is an effective method for operationalising poorly defined concepts. A well defined search filter benefits both the retrieval of evidence and the scope of the review.
Objectives: This paper presents a search designed for qualitative evidence on 'hard-to-reach’ groups at risk of tuberculosis. It explores the methods we used to construct a search filter for this ill-defined population and whether the collaborative approach we used was an effective method for defining difficult-to-define concepts.
Methods: Information specialists extensively tested potential terms and themes in a variety of resources and drafted the filter through database searching, citation chasing, expert/advisory contact, grey literature searching and working collaboratively with the review team. Hard-to-define populations were then analysed as a combination of specific risk groups (e.g. prisoners) coupled with risk factors (e.g. drug use) to ensure a highly sensitive population filter. The original search combined three themes (TB + Population + Qualitative terms). To test the efficiency of the filter, we re-ran our original search in Medline without the population filter. The results of the new search (TB + Qualitative terms) were screened to see if any populations were retrieved which had not been defined in the filter.
Results: The new search yielded 5454 studies not located in the initial search. No additional populations were identified after screening the extra studies. The sensitive population filter appeared to capture all relevant populations, suggesting that the time invested in creating the search strengthened the conceptual understanding of the population under review.
Conclusions: Dialogue between information specialists, review teams and topic experts is an effective method for operationalising poorly defined concepts. A well defined search filter benefits both the retrieval of evidence and the scope of the review.