Indexing of public health databases - report
Systematic reviews and other critical assessments of the messages of research for policy and practice are becoming increasingly important to the practice of evidence-based healthcare and health policy. To produce these reviews researchers are increasingly searching databases extensively to systematically identify research evidence. Researchers are seeking research evidence about specific topics and which have used one or more specific research designs. Search strategies thus focus on capturing subject topics, such as the prevention of alcohol addiction, and also study designs, such as trials or cohort studies. Being able to search effectively by study design can potentially reduce the volume of records that need to be assessed for relevance and so save time and money.
However, searching effectively for study design is dependent on a number of factors:
- The explicitness of the description of the study design given by the author of the research in the title and abstract. Good descriptions enable effective searching on title and abstract textwords.
- The degree of standardisation of the vocabulary used to describe a study design. Diverse vocabulary creates longer search strategies.
- The availability of abstracts in a given database. Abstracts enable more sensitive searching than might be possible on title alone.
- The availability of additional ways to search for records through a controlled vocabulary or thesaurus in a given database. Controlled vocabularies assist searchers by offering single search terms to capture many synonyms.
- The availability of suitable indexing terms to capture study design within the controlled vocabulary (thesaurus) used by a given database. A controlled vocabulary may not focus on all aspects of the subject and methods of research.
- The consistency of indexing achieved by indexers of a given database. If indexing is not of high quality the indexing may just be seen as an additional tool for the searcher rather than the preferred single route into the database.
An assessment of these factors needs to be made whenever approaching a database to conduct a search because databases differ in terms of their searching options and facilities. Researchers wishing to identify research evidence as economically as possible need to be aware of the different valuable features of databases that may help them to tailor a search more precisely to their research question.
Within the field of systematic reviews of public health interventions the range of study designs of interest can be large and finding evidence by study design can be problematic for the reasons described above. Information professionals from the Health Development Agency, Health Evidence Bulletins Wales and the Centre for Reviews and Dissemination have collaborated to explore the options for study design searching in some of the major databases of interest when researching public health interventions. The focus of this work has been to identify for a given database
- the availability of suitable controlled vocabulary
- the volume of potential research available by specific research design from searching by using both any controlled vocabulary and text words
This document is intended to save researchers' time and to indicate the most promising databases and search approaches for particular study designs. It is not intended to offer exhaustive search strategies for all study designs and its suggestions are indicative rather than comprehensive. Many issues of searching for study design in health and social care databases remain to be explored. As well as being a practical tool, this document provides some raw material to inform further research. Databases and research methods do not stand still and therefore this information will need to be updated to reflect changes in terminology, indexing practice and indexing terms. This document will be updated on an annual basis to reflect any changes.
Report of the HDA/CRD/Cardiff University(HEBW) Collaborative Group.
Prepared by Geoff Powell (Health Development Agency), Julie Glanville and Lisa Mather (Centre for Reviews and Dissemination) and Alison Weightman (Health Evidence Bulletins Wales).
Last updated August 2004
Many thanks to Bethan Evans, Dorothy Connell, Julia Bichard, Teresa Stevenson, Bao Luong and Astero Kanaris for their help in the preparation of this document.
This page was last updated: 30 April 2009