Test clusters to improve diagnosis (of shoulder pathology)
Clinicians face a range of problems when attempting to establish a firm diagnosis, among which the lack of high quality evidence or even the lack of evidence at all, and – if evidence is present – the absence of robust, easily interpretable, methodological characteristics.
Although sensitivity and specificity are relatively straightforward measures of diagnostic accuracy, their calculation only takes into account either those with (in the case of sensitivity) or without (specificity) a specific pathologic condition. In the positive likelihood ratio and negative likelihood ratio, both subjects with as well as without a certain condition are reflected. In order to produce stronger likelihood ratios, tests are clustered. Moreover, clustering of tests resembles clinical practice more closely than the use of a single test to confirm or reject a hypothesis.
6 solid test clusters are available concerning shoulder pathology: for impingement, rotator cuff tears, anterior instability and labral tears. Most notably, distinguishing between impingement and rotator cuff tears remains difficult due to overlap in tests included in the clusters, and in both clusters for rotator cuff tears, older age is a component.
> From: Hegedus et al., Phys Ther Sport (2014) (Epub ahead of print). All rights reserved to Elsevier Ltd. Click here for the Pubmed summary.