Diagnostic Scales

The evaluation of diagnostic tools, either scales based on clinically acquired information from the history and physical examination, or electrodiagnostic tests, requires a clear consensus on a reference standard against which the performance of these diagnostic tests can be compared. This type of consensus still does not exist with respect to carpal tunnel syndrome. It is recognized that electrodiagnostic testing has long been considered to represent a reference standard but this assumption is untenable because these tests clearly have false positive and negative results. Beyond this there simply is no consensus supporting any single diagnostic tool as a reference standard. Where clinical diagnostic scales are taken as the reference standard, electrodiagnostic tests may demonstrate poor sensitivity and specificity. The same is true of clinical diagnostic scales when electrodiagnostic tests are taken as the reference standard. Agreement between electrodiagnostic tests and clinical diagnostic tests, regardless of which is taken as the reference standard, is also complicated by the binary nature of the comparison. Electrodiagnostic data is, by and large, continuous in nature and so establishing a hard cutoff point to compare to clinical diagnostic scales seems potentially arbitrary. At least one of the clinical diagnostic scales, the CTS-6, attempts to address this by defining the diagnosis in probabilistic terms as a continuous variable. Given this set of circumstances the Workgroup sought to evaluate the role of clinical diagnostic tests and electrodiagnostic testing in the evaluation of CTS in the context in which they are used, in other words, in clinical settings where a patient presents with complaints that might be attributable to this condition.
 
There were two clinical diagnostic tests studied in high quality investigations, the Katz Hand Diagram and the CTS-6. The Boston Carpal Tunnel Scale, a status instrument most frequently used to measure outcomes of treatment for CTS was also evaluated in two high quality studies.
 
In comparison to electrodiagnostic testing Katz et al demonstrated high sensitivity (0.96) and good negative predictive value (0.91) for the “classic”, “probable” or “possible” designations however, positive predictive value and specificity were low. This indicates that, using electrodiagnostic testing as a reference standard, the Katz Hand Diagram used in this way had more value as a “rule out test”. Sensitivity decreased and specificity increased if comparison to electrodiagnostic tests was made only using “classic” or “probable” results. Sensitivity decreased further and specificity was commensurately increased when only “classic” results were compared to electrodiagnostic testing. Defined using only “classic” or “classic” or “probable” results the Katz Hand Diagram was considered weak or poor as either a “rule in” or “rule out” test. Vanti made similar observations using AANEM electrodiagnostic definitions for CTS in demonstrating that the “classic” or “probable” results functioned as a strong “rule out” test.
 
Graham took a different approach to evaluating the respective roles of electrodiagnostic testing and the CTS-6, an instrument that expresses the probability of CTS. The pre-test probability of CTS was established using the CTS-6 and then the post-test probability after electrodiagnostic testing was estimated using likelihood ratios established with two electrodiagnostic standards for CTS, one lax (with higher sensitivity and lower specificity) and one stringent (with lower sensitivity and higher specificity). This study showed that the changes in probability after electrodiagnostic testing, using either electrodiagnostic definition, were small and probably below a clinically relevant standard. This suggests that the most appropriate setting for electrodiagnostic testing is where there is uncertainty about the clinical diagnosis.
 
There were two high quality studies evaluating the Boston Carpal Tunnel Syndrome Questionnaire (Wainner, Naranjo). Both of these studies used electrodiagnostic tests as the reference standard. The results were consistent in both studies in showing that this instrument functioned as either a weak or poor “rule in” or “rule out” test. This may have been due to the fact that the scale was actually developed as a status instrument rather than as a diagnostic scale.