Toward Personalized Neuromodulation: Nomogram Models Forecasting Long-Term Response After Temporary Peripheral Nerve Stimulation.

INTRODUCTION

Peripheral nerve stimulation (PNS) delivered for 60 days has emerged as a promising short-term neuromodulation therapy for chronic pain, with early outcomes showing favorable results. However, the ability to predict which patients will sustain long-term benefit remains limited, given existing predictive models have not fully integrated psychosocial and behavioral factors that are increasingly recognized as key modulators of response in chronic pain management. To address this gap, we aimed to identify psychosocial predictors of sustained pain relief at 12 months after 60-day PNS and develop clinically relevant nomograms to guide patient selection and prognostication.

MATERIALS AND

METHODS

This prospective, multisite observational cohort study enrolled 110 patients who underwent temporary PNS therapy for chronic trunk and/or limb pain. Baseline assessments included demographics, clinical characteristics, and validated psychosocial instruments including the Pain Catastrophizing Scale, Pain Self-Efficacy Questionnaire, Oswestry Disability Index, and Physical Activity Vital Sign. The primary outcome was ≥50% pain reduction at 12 months. Predictors were evaluated using univariate logistic regression, and two nomograms were developed: one incorporating core clinical features and another integrating psychosocial factors. Model performance was assessed using the area under the curve (AUC), internal validation with 1000 bootstrapped samples, and penalized regression through Least Absolute Shrinkage and Selection Operator.

RESULTS

A total of 110 participants completed follow-up. Initial 60-day response (odds ratio [OR] 11.09), high physical activity (OR 7.71), and high pain self-efficacy (OR 6.67) were strong positive predictors, whereas pain catastrophizing, insomnia, anxiety, and high baseline disability predicted nonresponse. The clinical nomogram achieved excellent discrimination (AUC = 0.91), whereas the psychosocial-enhanced model showed robust performance (AUC = 0.97). Adverse events were rare and mild, and supplemental treatments were more common in nonresponders.

CONCLUSION

Treatment response at 60 days, functional capacity, and psychosocial context meaningfully influence long-term outcomes after temporary PNS. The resulting nomogram models show promise for supporting individualized patient selection and hypothesis generation in future predictive neuromodulation research. However, these findings should be considered preliminary pending external validation in independent cohorts.