Systematic Reviews and Meta Analysis

To find articles on the accuracy of diagnostic tests, search for the condition being diagnosed and then the index test. For instance, if you wished to find studies that evaluated the effectiveness of computed tomography in diagnosing head injuries, you'd create a search for CT scanning and head trauma.

A large number of filters have been developed to focus search results on papers that describe tests of diagnostic accuracy. These include the PubMed Clinical Queries broad and narrow Diagnosis filters [1]. Recent work has shown that these filters are unlikely to reduce substantially the number of records reviewers need to screen while increasing the risk that relevant studies will be missed. A recent Cochrane methodological review found that none of the existing filters exhibited sensitivities exceeding 90% [2]. These authors conclude that such filters should not be used when performing systematic reviews of diagnostic tests.

In the real world, resources are limited, and merely searching for a condition and a test can produce large, unmanageable results sets. Hence it may be necessary to use a TDA filter. If so, every effort should be made to identify additional tests through independent means. These methods can include checking the references of relevant studies uncovered in the search and consulting subject experts. One of the better performing fliters, reported by Bachman [3] has 74-90% specificity (here in a form suitable for PubMed):

"Sensitivity and specificity"[MeSH] OR predict*[tw] OR diagnos*[tw] OR accura*[tw]

An independent filter has been evaluated for Embase [4] that shows a bit better sensitivity (86-97%) but it's not particulary precise:

sensitiv* OR detect* OR accura* OR specific* OR reliab* OR positive OR negative OR diagnos*

These filters will reduce the retrieval set to a third or so, but there is a risk of missing relevant articles.

GetTheDiagnosis.org may already have found what you need. This tool is a collaborative database that includes the sensitivity and specificity of history questions, physical exam findings, and lab and imaging tests.

Additional Reading:

1.   Haynes RB, McKibbon KA, Wilczynski NL, Walter SD, Werre SR; Hedges Team. Optimal search strategies for retrieving scientifically strong studies of
treatment from Medline: analytical survey. BMJ. 2005 May 21;330(7501):1179. PubMed PMID: 15894554; PubMed Central PMCID: PMC558012.

2.   Beynon R, Leeflang MM, McDonald S, Eisinga A, Mitchell RL, Whiting P, Glanville JM. Search strategies to identify diagnostic accuracy studies in
MEDLINE and EMBASE. Cochrane Database Syst Rev. 2013 Sep 11;9. Review. PubMed PMID: 24022476.

3.   Bachmann LM, Coray R, Estermann P, Ter Riet G. Identifying diagnostic studies in MEDLINE: reducing the number needed to read. J Am Med Inform Assoc. 2002 Nov-Dec;9(6):653-8. PubMed PMID: 12386115; PubMed Central PMCID: PMC349381.

4.   Bachmann LM, Estermann P, Kronenberg C, ter Riet G. Identifying diagnostic accuracy studies in EMBASE. J Med Libr Assoc. 2003 Jul;91(3):341-6. PubMed PMID: 12883560; PubMed Central PMCID: PMC164397.