A Systematic Review of the Role of Wearable Devices and Artificial Intelligence Applications in Assessing Functional Outcomes after Lumbar Fusion.

Abstract

Background context: As the population ages, rates of lumbar spine disease have risen, and lumbar fusion surgeries have become more prevalent. There has been a corresponding emphasis on value-based cost reductions and outcomes research to identify which patients benefit from fusion. While wearable remote-monitoring devices such as goniometers have been used for some time in other medical fields, these seem to have yet to attain wide usage in spine surgery.

Purpose: We aimed to conduct a systematic review of the PubMed database in accordance with PRISMA guidelines to characterize the use of wearable devices to describe functional outcomes before and after lumbar fusion surgery. We discuss the role of artificial intelligence and its applications in terms of predictive analytics incorporated into such portable devices for evaluating outcomes of lumbar fusions.

Study design/setting: Systematic review of studies evaluating the use of wearable devices for functional outcomes in lumbar fusion surgery. The review was conducted using the PubMed database and followed PRISMA guidelines.

Methods: We included all relevant articles and excluded lumbar spine surgeries without fusion (i.e. microdiscectomy), review articles and editorials, proof-of-concept studies, biomechanical analyses, and technical notes.

Results: Our initial search generated 5283 citations, of which 9 articles with 813 patients were ultimately included. 5/9 (55%) studies included steps per day as a primary outcome. All studies were pre-post in design. Data collected included vitals, positional data, step counts, diet and sleep data, incision photos, pain scores, and serial patient reported outcome measure administration. Benefits of wearable devices with and without artificial intelligence/predictive analytics included patient education, reduced ER visits, reduced in-person visits, continuous data collection, earlier identification of complications, and wearable devices that do not require FDA device approval. Drawbacks of wearable devices with and without artificial intelligence/predictive analytics included concerns for data security, uncertain cost-effectiveness, lack of standard protocols, heterogeneity of devices, and susceptibility to placebo effect. Overall, studies including wearable devices with and without artificial intelligence/predictive analytics showed that lumbar fusion patients recovered functionally more slowly (i.e. when compared to discectomy patients) but had good long-term functional outcomes.

Conclusions: Our review suggests wearable devices enhance post-operative monitoring for lumbar fusion surgery by providing real-time, objective data to optimize rehabilitation and functional recovery. As digital health tools evolve, integrating predictive analytics driven by artificial intelligence and through wearable devices may further refine personalized rehabilitation strategies, improve long-term outcomes, and provide other benefits.