3.1 The evidence for Neuropad assessed by the external assessment centre (EAC) comprised 18 studies, of which 13 were full text articles and 5 were abstracts. Of the 18 studies, 17 investigated the diagnostic accuracy of Neuropad against a reference standard and 1 reported its ability to predict the risk of diabetic foot ulceration. In addition to examining diagnostic accuracy, 1 study looked at the reproducibility of results when using Neuropad and 3 assessed the association between Neuropad testing and developing foot ulcers. The most common reference standard used was the neuropathy disability score. All the studies were prospective observational, cross-sectional or longitudinal cohort studies. For full details of the clinical evidence, see section 3 of the assessment report.
3.2 The EAC considered that the 2 published UK studies (Ponirakis et al. 2014 and Quattrini et al. 2008) were fully relevant to the scope. The EAC also did a meta-analysis of 5 diagnostic accuracy studies that used a neuropathy disability score of 5 or more as the reference standard: Freitas et al. 2009, Kamenov et al. 2010, Liatis et al. 2007, Manes et al. 2016 and Tentolouris et al. 2008.
3.3 The EAC used its meta-analysis results comparing Neuropad with a neuropathy disability score of 5 or more with the results obtained for 10 g monofilament. It concluded that Neuropad may be more sensitive than 10 g monofilament testing in detecting diabetic peripheral neuropathy (DPN), but have lower specificity. In addition, the EAC noted that the current care pathway includes interventions that are triggered only by clinically apparent DPN, which would be regarded as moderate or advanced, so the benefit of detecting preclinical DPN in the current care pathway is uncertain.
3.4 Neither the company nor the EAC identified any relevant published economic evidence. The company submitted a Markov model with 2 comparisons: Neuropad testing compared with 10 g monofilament testing, and Neuropad testing compared with Neuropad testing then 10 g monofilament testing. The time horizon of the model was 3 years. The EAC made a number of changes, including: adding the implications of false-negative and false-positive results; adding a death state; extending the time horizon to 10 years; and adding a third comparison of Neuropad testing with no testing. For full details of the economic evidence, see section 4 of the assessment report.
3.5 The EAC disagreed with a number of the sources used to generate parameter values in the company's model. It also noted discrepancies between the values used in the model and those quoted in the referenced sources. Moreover, the EAC considered that the cost of 10 g monofilament testing in the model had been overestimated, because it included the cost of the reusable holder. For full details of the EAC's changes to the company's economic model, see sections 4.2 and 4.3 of the assessment report.
3.6 Results from the EAC's revised model showed that Neuropad testing incurs additional cost over a 10‑year time horizon compared with all other comparators:
£775 per patient compared with 10 g monofilament testing
£1,075 per patient compared with Neuropad testing then 10 g monofilament testing
£1,792 per patient compared with no testing.
The EAC did sensitivity analyses which showed that Neuropad testing alone was not cost saving in any considered scenario.