Using intensive longitudinal data to investigate predictors of cannabis and opioid use among patients with chronic pain

Background: Prolonged prescription opioid use and access to unregulated potent synthetic opioids drives the Opioid Epidemic in the United States. Chronic pain, which affects over 50 million U.S. citizens, significantly contributes to this crisis, as about 10% of patients with chronic pain eventually develop opioid use disorder. Interventions that reduce prescription opioid use may lead to fewer adverse health outcomes, OUD, and fatal overdoses. Over 24 million U.S. citizens currently use cannabis as an alternative or complementary pain management strategy. Yet, research on the effectiveness of cannabis for chronic pain management is limited by retrospective recall bias, cross-sectional designs, and a lack of information on daily use patterns. 
Method: To address these limitations, we conducted an intensive longitudinal ecological momentary assessments (EMA) study with patients with chronic pain (n = 129) who used opioids and cannabis. Participants provided an average of 78 (SD = 30) smartphone-based EMAs from four randomly prompted surveys per day across 30 days (10,062 observations). Surveys assessed participants' use of opioids, cannabis, and other substances, pain severity, pain catastrophizing, affective and emotional states, and cravings for opioids and cannabis. We used a novel, algorithmic analysis strategy called Group Iterative Multiple Model Estimation to determine the presence and direction of emotional, affective, and cognitive pain reactivity mediators influencing opioid and cannabis use.
Results: The results revealed a main pattern beginning with stress triggering pain, initiating opioid cravings that progressed to opioid use, followed by cannabis use. Moreover, pain resulted in pain catastrophizing, marked by rumination, helplessness, and magnification, which led to a sequence from opioid craving to use and subsequently to cannabis use. Finally, a parallel process emerged, whereby opioid cravings led to cannabis cravings and opioid use led to cannabis use – suggesting that cannabis use occurred secondary to opioid use among these patients.
Conclusions: Our findings provide information on moment-to-moment predictors of opioid and cannabis use among patients with chronic pain. These critical sequences reveal a nuanced relationship between chronic pain, opioid cravings and use, as well as cannabis use. Findings can inform interventions that target pain catastrophizing to potentially prevent the necessity for opioid use by curbing opioid cravings. Future research is needed to investigate why cannabis cravings and use follow opioid cravings and use and how to leverage this pattern to reduce harms from opioid use. Ultimately, these novel behavioral patterns may be useful to inform future harm reduction recommendations and treatment plans to improve the quality of life for millions of patients with chronic pain.