4.1
The evidence for the accuracy of Neuropad in diagnosing preclinical diabetic peripheral neuropathy (DPN) comprises longitudinal observational studies that mainly compared testing in terms of neuropathy scoring systems (most commonly the neuropathy disability score). The committee was aware that the external assessment centre (EAC) had rejected the study by Tsapas et al. (2014; a meta-analysis identified by the company) because of overlapping populations in the studies included and differences in the reference standards used, and had, instead, done its own meta-analysis. The results from the EAC's meta‑analysis showed that Neuropad has a sensitivity of 89.4% (95% confidence interval [CI] 83.2 to 93.5) and a specificity of 60.3% (95% CI 50.9 to 69.0), when using a neuropathy disability score of 5 or more as a reference standard for the diagnosis of DPN. Based on this, the committee concluded that Neuropad demonstrates good sensitivity but poor specificity as a diagnostic test for DPN. Although no direct comparative data were available for 10 g monofilament, the committee and EAC agreed that it was appropriate to use the sensitivity (84%) and specificity (83%) estimates for 10 g monofilament that were used in NICE's medical technologies guidance on VibraTip. The committee concluded, therefore, that the current evidence for Neuropad is insufficient to support its effectiveness as an alternative test to 10 g monofilament for detecting DPN.