Abstract
Chronic, or long-term, opioid use is associated with higher risk of addiction and death as well as poor inflammatory bowel disease (IBD) outcomes. Recent data estimate that up to 20% of adolescents and young adults (AYA) with IBD may be chronic opioid users. Models to predict chronic opioid use among AYA with IBD may support identification, clinical management, and follow-up of patients at high risk of opioid dependence. We aimed to develop a clinical predictive model based on administrative data for chronic opioid use among AYA with IBD.
We performed a retrospective analysis of AYA patients between 15-29 years treated at a tertiary academic center for IBD management from 3/2018-12/2021. Univariate regression was performed to identify individual predictors of chronic opioid use. Multivariate regression was performed using variables with p-values <0.15 in univariate regression. Variables significant after adjustment were included in a simple scoring system. Point estimates for chronic opioid use risk probabilities were derived using the percentage of chronic opioid users in our analytic cohort assigned that prognostic score. 95% confidence intervals were estimated using a bootstrapping technique with 1000 replications.
Our final cohort included 575 unique AYA patients with IBD. Patients were, on average, 23 years old (IQR 20-26) and 69% of the cohort was female. 41% had a diagnosis of Crohn’s disease, 25% with ulcerative colitis and 64% with indeterminate colitis. Patients’ mean age at index IBD diagnosis was approximately 22 years and over half (56%) had undergone prior IBD surgery. Following adjustment, eleven factors were associated with the highest odds of chronic opioid use, including: public insurance coverage, non-digestive or non-gastrointestinal (GI) surgery, outpatient prescription receipt for weaker opioids (i.e. codeine/tramadol), and concurrent depression and post-traumatic stress disorder (PTSD) diagnoses (Figure 1). Prognostic scores ranged from 1-9 (mean: 5, SD=2.7). Patients were stratified into three risk categories by prognostic score (1-4; 5-7; 8-9) with associated risk scores of 52.1% (95% CI: 48.9-55.2%), 68.6% (66.5-71.1%), 86.2% (76.6%-94.4%), respectively (Figure 2).
We developed a practical scoring system to stratify AYA with IBD by chronic opioid use risk. Future directions include understanding the relationship between these risk factors and their influence on chronic use. Risk factors most associated with chronic opioid use among AYA with IBD included: public insurance coverage, undergoing a non-digestive surgical procedure, and having concurrent diagnoses of depression and PTSD. Findings will directly inform targeted opioid reduction initiatives for this vulnerable population.
Figure 1 Final risk factors associated with development of prognostic score.
Figure 2 Percent of chronic opioid users by prognostic score.
