Artificial Intelligence Interventions in Chronic Pain Management: A Scoping Review (Preprint)

Abstract

Background There are indicators that artificial intelligence may have the potential to improve chronic pain care globally. However, research in this field is still evolving. Hence, it is necessary to synthesize the available evidence and evaluate its current scope.

Objective This scoping review presents a comprehensive synthesis of the available evidence, highlighting the types and delivery of biopsychosocial AI-based interventions, the chronic pain conditions managed, the population characteristics, and the context of care delivery.

Methods The literature search was conducted using MEDLINE, Embase via Ovid, and AMED (until July 25, 2023) in line with the PRISMA guidelines. Eligible studies were appraised using the relevant JBI checklist for each study type. The results were synthesized and reported through a narrative review. The review included English-language studies of any research design that reported a form of AI-based intervention for chronic pain management.

Results 32 RCTs and non-RCTs identifying 23 AI-based interventions were included. The results showed that many interventions targeted the management of one or more of the biopsychosocial components of chronic low back pain. These interventions were delivered through mobile apps, chatbots (text messages), computer-based or equipment-based algorithms, and robots. The interventions focused on exercises, CBT, and feedback.

Conclusions The use of artificial intelligence in delivering chronic pain interventions is developing rapidly. However, most interventions were not delivered by an interdisciplinary care team as recommended by the IASP. Furthermore, the report of many systems, without sufficient evidence to support effectiveness, may limit translation to practice. Thus, a joint effort of a team of expert pain clinicians and researchers is required to facilitate the uptake of AI-based interventions for chronic pain management.