We developed a structured clinical documentation support (SCDS) toolkit that standardizes polyneuropathy patient data collection to conform to Best Practices for evaluating patients with polyneuropathies. Part of this project was to design a practical, EMR embedded neuropathy scale, that can be easily used in all newly diagnosed polyneuropathy cases.

The NSNIS was used in 2,215 new patients evaluated for a diagnosis of polyneuropathy. Out of these, the diagnosis of polyneuropathy was established in 1,182. The NSNIS is highly predictive for a diagnosis of polyneuropathy (AUC 0.717).   For pattern of involvement, the NSNIS performed better in cases of large fiber polyneuropathy (AUC 0.81) compared to cases with small fiber polyneuropathy (AUC 0.63), although it performed better in patients with autonomic symptoms (AUC 0.768). Prediction of anatomic distribution of findings was as follows: isolated polyneuropathy (AUC 0.713), added radiculopathy (AUC 0.778), added mononeuropathy (AUC 0.713). When looking at potential etiologies, the NSNIS was more predictive of a hereditary cause (AUC 0.848) compared to idiopathic (AUC 0.566) and acquired etiologies (AUC 0.657).   NSNIS performed equally well in axonopathies (AUC 0.718) compared to primary myelinopathies (AUC 0.708).