Validating a Pain Stability Algorithm for Multiple Sclerosis-Induced Neuropathic Pain: Implications in Pain Management

Objectives : To validate three novel algorithms used to categorize longitudinal pain variability in multiple sclerosis (MS)-induced neuropathic pain (NPP) against clinical opinion of variability and, utilizing the superior algorithm, categorize overall variability in our cohort. Methods : Daily visual analogue scale (VAS) pain scores were collected over 28 days from patients with relapsing-remitting MS (RRMS) and NPP (n=29). Three newly developed “Pain Stability Algorithms” categorized each patient as either stable or unstable with respect to pain variability over 28 days. Algorithms were validated by comparing categorization results to blinded ratings of two clinical pain specialists, who also categorized pain ( stable or unstable ) following review of daily VAS scores. Analysis conducted between pain specialist ratings and algorithm rating results determined the algorithm that best reflected the clinical opinion of variability. Regression analysis determined which baseline patient characteristics – age, EDSS, baseline pain, pain duration and years with MS - predicted categorical pain variability labeling identified by the superior stability algorithm. Results : Stability Algorithm C agreed superiorly with clinical opinion. This algorithm detected unstable patients with sensitivity and specificity of 0.86 and 0.93 respectively and agreement of 90% to the blinded clinical ratings. Using Stability Algorithm C, 33% of the cohort displayed unstable MS pain. No significance between baseline characteristics and categorical variability assignment ( stable or unstable ) observed. Discussion : A clinically validated algorithm was developed to categorize pain stability in MS-induced NPP, and classified approximately 33% of patients with clinically unstable daily pain. No baseline patient characteristics were significant in predicting pain stability ratings.