Volume 200 - Issue 4

Finally . . . an evidence-based tool to find primary health care evidence

Authors:  Jennifer J Tieman, Ruth M Sladek and Ellen McIntyre

Med J Aust 2014; 200 (4): 207-208. || doi: 10.5694/mja13.11345
Published online: 3 March 2014
An improved capacity to retrieve relevant evidence helps support policy and practice

To the Editor: If primary health care is to maximise its contributions to population-based outcomes, there must be optimal capacity to retrieve the evidence so critical to informing Australian policy and practice.1 We report a tool called the PHC (Primary Health Care) Search Filter, which retrieves the underlying corpus of relevant published literature in MEDLINE. The PHC search filter was developed using established research methods, incorporating a gold standard comparison approach.2 This useful and simple search tool offers a sophisticated mechanism for retrieving relevant documents pending the development of operational machine-based filtering.

A dual independent review of references of 10 systematic reviews commissioned by the Australian Primary Health Care Research Institute and deemed relevant to primary health care by the Expert Advisory Group (EAG) provided a gold standard set of MEDLINE-indexed articles for testing. These citations were randomly divided into three sets: the term identification set (TIS), filter development set (FDS) and filter validation set (FVS). The TIS was used to generate a list of candidate MeSH (Medical Subject Headings) and text word search terms based on frequency analyses. The highest frequency terms were individually and collectively tested for retrieval effectiveness in the FDS. The best performing search string was then tested in the FVS. This search, comprising 12 terms known as the PHC search filter, successfully retrieved 154/200 citations (77.0%) in the FVS and 415/529 citations (78.4%) from the full gold standard set.

The filter was then translated for use in PubMed through the Primary Health Care Research and Information Service website (http://www.phcris.org.au) to facilitate quick, real-time and free access to relevant citations. A post-hoc relevance test of 500 items retrieved by the search filter in PubMed, dual-reviewed by the EAG, demonstrated 88.4% relevance. Testing in a mixed set of known relevant and non-relevant records could provide a further estimate of the filter’s precision and sensitivity. The PubMed version comprises the terms listed in the Box and is articulated in full in an unpublished report available on request. Readers can access the search filter at http://www.phcris.org.au/phcsearchfilter. The filter can be run alone or combined with other useful search topics such as chronic disease management, continuity of care and heart failure.

This filter was developed to support policy and practice with respect to primary health care in Australia, and builds on earlier work in relation to palliative care as part of the federally funded CareSearch project (http://www.caresearch. com.au). Collectively these strategies demonstrate how evidence-based approaches to finding the evidence can be implemented into online knowledge networks.


More like this

General medicine Perspective 17 November 2025 Open Access

The CURE Asthma roadmap

Gary P Anderson, Anthony Flynn, Phil G Bardin, John D Blakey, Shyamali C Dharmage, Paul Foster, Peter G Gibson, Adam Jaffe, Alan James, Christine R Jenkins, Sundram Sivamalai, Peter D Sly, Guy B Marks, Vanessa M McDonald, Judy Wetttenhall