AI Pain Care Tool May Increase Access, Reduce Costs for Chronic Pain Patients
A cognitive behavioral therapy intervention for chronic pain, personalized for patients using artificial intelligence, achieved similar results as standard therapies.
Source: Getty Images
By - New research published in JAMA Internal Medicine shows that an artificial intelligence (AI)-based cognitive behavioral therapy intervention for chronic pain (AI-CBT-CP) had similar outcomes to standard CBT-CP, which could result in increased access and reduced costs.
According to the study, CBT-CP is an effective alternative to opioids for chronic pain management. However, CBT-CP requires multiple sessions with a therapist, and there are not enough therapists to provide these services due to the ongoing clinician shortages and healthcare workforce issues. Because therapists are scarce, many patients have severely limited access to CBT-CP or fail to complete their treatment.
With relatively few other options available, many chronic pain patients may turn to opioids. Opioid prescription and misuse have come under scrutiny amid the ongoing opioid epidemic. Some research suggests that opioid overdoses increased significantly over the course of the COVID-19 pandemic.
Clinicians have advocated for enhanced patient-centered care for chronic pain patients for years, and this research contributes to those efforts. To determine if AI-CBT-CP is a viable alternative to its standard counterpart, the researchers compared the performance of both interventions in 278 patients with chronic back pain from the Department of Veterans Affairs health system.
The AI-CBT-CP intervention is designed to customize patient treatment using reinforcement learning and interactive voice response (IVR), an automated system often used in customer service call centers that combines pre-recorded messages with an engagement interface that allows users to provide and access information without someone else on the line.
All patients were offered 10 weeks of CBT-CP as part of the study. However, the AI-based intervention group was asked to provide daily feedback about their therapy via IVR calls. Using the data from these calls, the reinforcement learning algorithm made weekly recommendations to personalize each patient’s therapy in the form of either a 45-minute or 15-minute therapist-delivered telephone session or an individualized IVR-delivered therapist message.
Treatment outcomes for all participants were measured at three months and six months following therapy completion using the Roland Morris Disability Questionnaire (RMDQ). Patients were also asked about their pain intensity and pain interference. The research team used clinical guidelines to identify meaningful improvements in these areas.
The average difference in RMDQ scores between the AI-based intervention group and the standard CBT-CP group was -0.72 points at three months and -1.24 points at six months following treatment completion. At six months, 37 percent of AI-CBT-CP patients had clinically meaningful improvements compared to 19 percent of those in the standard CBT-CP group. Using AI-CBT-CP also required less than half of the therapist's time as standard CBT-CP.
These findings indicate that AI-CBT-CP is non-inferior to standard CBT-CP and requires considerably less therapist time, the research team concluded. Thus, AI-CBT-CP has the potential to allow more patients to receive CBT-CP interventions using the same number of therapists, which would increase access and decrease costs.
