.png)
Symposia
Research Methods and Statistics
Madelyn Frumkin, Ph.D. (she/her/hers)
Massachusetts General Hospital
Boston, Massachusetts, United States
Bryn Evohr, BA
Clinical Research Coordinator
Massachusetts General Hospital Center For Addiction Medicine
Boston, Massachusetts, United States
Julia Jashinski, MSW, LCSW
Associate Program Manager
Massachusetts General Hospital Center For Addiction Medicine
Boston, Massachusetts, United States
Cori Cather, MD
Director
Massachusetts General Hospital Center Of Excellence For Psychosocial And Systemic Research
Boston, Massachusetts, United States
Gladys Pachas, MD
Program Director
Massachusetts General Hospital Center for Addiction Medicine
Boston, Massachusetts, United States
A. Eden Evins, MD, MPH
Founding Director and Principal Investigator
Massachusetts General Hospital Center For Addiction Medicine
Boston, Massachusetts, United States
Jodi Gilman, PhD
Director of Neuroscience and Principal Investigator
Massachusetts General Hospital Center for Addiction Medicine
boston, Massachusetts, United States
Despite significant risk and limited evidence of long-term efficacy, over 20% of US adults with chronic pain use prescription opioids. Presciption Opioid Taper Support (POTS) interventions have been developed to help patients learn non-medication pain management strategies while slowly decreasing opioid dose. These interventions tend to be long (approximately 18-26 sessions) and incorporate cognitive-behavioral skills targeting several domains, including pain, sleep, and mood. We leveraged daily diary data collected during a 26-week POTS program to better understand potential within-person treatment mechanisms. Individuals with chronic non-cancer pain and on chronic opioid medications (N = 51, Mage = 58, SDage = 9.5, 63% Female, 88% White, 94% Non-Hispanic) each completed 149 surveys on average over the 26 weeks (SD = 32, Min = 68, Max = 195). Average adherence was 82% (SD = 16%), indicating feasibility of daily assessments during the behavioral intervention. We planned to use intensive longitudinal mediation models to understand whether non-medication pain management strategies (i.e., CBT skills) were associated with decreased opioid dose, and whether such relationships were mediated by changes in pain, sleep, and/or depression symptoms. We observed limited support for mediation effects in lagged models. For example, there was one participant for whom greater helpfulness of non-medication pain management strategies was associated with lesser next-day depressed mood (β = -0.281, 95% Credible Interval: -0.487, -0.086), which was associated with next-day opioid dose (β = 0.227, 95% CI: 0.109-0.341). However, this individual did not benefit overall from the intervention (pre-treatment morphine milligram equivalents = 66.46, post-treatment = 66.92). We observed greater variability in same-day associations between opioid dose and proposed treatment mechanisms, such that daily pain, sleep, and depression symptoms were each associated with same-day opioid dose for 32-49% of participants. We will discuss ongoing efforts to understand person-specific treatment mechanisms through intensive longitudinal data.