Evaluation of clinical data and regulatory submission characteristics for neuromodulation devices approved via the 510(k) pathway (1996-2024).

Abstract

Objectives: The 510(k) pathway is prevalent for medical device approval, yet its integration of clinical data, especially in neuromodulation devices, remains underexplored.

Methods: We evaluated neuromodulation devices approved via the 510(k) pathway from 1996 to 2024. Using Natural Language Processing, we traced predicate networks and analyzed clinical data inclusion. Published clinical data were also reviewed.

Results: Among 715 medical devices, 31 neuromodulation devices were identified, with 30 having summary documents. The predicate network included 160-related 510(k) approvals; 48 devices (30%) included clinical data, mainly from randomized trials. Nearly 30% of files were untraceable or lacked an FDA summary. Most approvals were based on a single predicate. Seven percent were subject to recalls, all approved without clinical data. Median time-to-approval was significantly longer for devices with clinical data than those without (157 vs. 110 days, p = 0.008). Clinical data from 36 publications included 32 trials and 4 observational studies, with a median Cohen's d of 0.83 (0.27-1.71) and NNT of 4.8 (2.4-7.9). Trials with differing indications from predicates had larger sample sizes.

Conclusions: A significant shift toward clinical data inclusion in the 510(k) pathway was observed, particularly for neuromodulation devices. Challenges include balancing safety standards while streamlining approval process.