Overview - Diagnostic Questions

This kind of question arises frequently for most radiologists when they encounter an image demonstrating an unusual finding that is not immediately familiar to them, but for which they have a differential diagnosis. This knowledge gap is often best filled by an article on rad-path correlation. Unfortunatly this was not an area of interest for EBM physicans when they worked on critical apprasial of literature. There is, therefore, no relevant EBM literature on searching for, or analysing, articles on Rad-Path correlation. It is an area for further research in EBR. Because it is important to radiologists, we have included it on this site.

In our experience, the journal ‘Radiographics’ is a very useful source of information. Failing that, a search of the ‘RSNA Index to the Imaging Literature’ can provide rapid access to relevant literature. A wider search of primary resources can be performed if a suitable article is not found. Image databases maintained by organisations (Eurorad), RSNA, Radiology Central or others (Aunt Minnie) may be helpful, although the quality of the results becomes progressively less predictable. If you know what the disease is, but want up-to-date information on its appearances with a particular modality, either the RSNA Index to the Imaging Literature or the Clinical Queries section of PubMed can give quick ‘hits’. The bottom line, at present, is that Rad-Path textbook resources are still very valuable for many cases.

Comparison of imaging methods in a specific clinical condition

This is a more classical type of evidence-based question. In this situation, you need to find the best available evidence about the superiority of one imaging method over another in resolving clinical dilemmas or about the power of imaging signs to reliably confirm or exclude suspected disease processes. The search must therefore be wider than that performed for a ‘rad-path’ correlation. Regrettably, there is very little secondary evidence available for diagnostic radiologists as yet. We will concentrate on search tips for the primary literature. The APPRAISE section of this site describes the hierarchy of evidence within which individual retrievals are ranked.

Search Engines

Search engines can be classified as ‘raw’ or ‘filtered’, ‘free’ or ‘subscription-based.’

When to use which search engine?

In our experience to date, many ‘knowledge gap’ questions arising in daily practice are readily answered by a ‘filtered’ search. These usually return a small number of highly relevant papers within a few seconds and can, for example, be used to answer a few questions before finishing a reporting session.

If the problem is one of Departmental strategy / policy or if it is centered around an academic exercise to be submitted for publication, a ‘filtered’ search alone is insufficient. It will likely exclude some relevant citations, such that all of the best available evidence will not be retrieved. However, when attempting a comprehensive search, we find it useful to start at a filtered site to get a few relevant papers. The ‘view citation’ function in PubMed, which is on the DISPLAY pull-down menu with SUMMARY and ABSTRACT, lets us see what MeSH terms have been used to index these. This information is helpful in building a good PICO question for use in PubMed or OVID. Another useful starting point is a standard textbook of radiology. If you look in the relevant chapter, find the relevant section and check the references, they are often too old for current use. However, you can look them up on PubMed and (again using the 'citation' display function) find out what MeSH terms were used to index them. We find these tips reduce the time needed to design a satisfactory search.

Detail about the technical aspects of a technique in a specific clinical condition.

Our experience to date has been that the non-radiologic literature is typically light on technical details. We suggest starting with the RSNA Index to the Imaging Literature and proceeding to Knowledge Finder as a starting strategy.