Opioid use, opioid use disorder, and poisoning among children and adolescent cancer survivors

The aims of this study were to: investigate and compare opioid use patterns between pediatric cancer survivors and their matched controls; predict opioid use, chronic/persistent opioid use, and opioid use disorder (OUD); and describe and compare healthcare resource utilization (HCRU) and costs among children and adolescent cancer survivors. This is a retrospective cohort study using the 2007-2019 Optum Clinformatics® commercial claims database. The study cohort included cancer survivors aged 1-21 years diagnosed with leukemia, lymphoma, brain and other central nervous system (CNS) tumors, bone cancer or soft tissue sarcoma, or Wilms’ tumor. Two study designs were involved. For study 1, the index date was the date of the first cancer diagnosis. A control group was created using 3:1 exact matching on the index age, sex, geographic region, and Charlson Comorbidity Index score. For studies 2 and 3, the index date was the 30th day after the last observed active cancer treatment date. For study 1, 1,263 survivors and 3,789 matched controls were included. During the 3-year follow-up, 735 of 1,263 (58.2%) cancer survivors and 805 of 3,789 (21.3%) matched controls had [greater than or equal to] 1 opioid prescription. Among survivors, the unadjusted rates of chronic opioid use (COU), OUD, and opioid poisoning were 3.2%, 1.0%, and 0.4%, respectively. Compared to controls, cancer survivors were more likely to fill opioids (aOR=4.10, 95% CI=3.65-4.60), and have COU (aOR=8.7, 95% CI=4.4-17.2). For studies 2 and 3, a total of 1,181 survivors were included. There were 213 (18.0%) patients using opioids and 14 (1.2%) experienced persistent opioid use or OUD. Predictors for persistent opioid use or OUD post cancer treatment included higher age, anxiety diagnosis, earlier cancer diagnosis year, opioid use prior to and during cancer treatment, cancer type (brain/CNS), longer duration of cancer treatment, more emergency department (ED) visits and prescriptions. The AUCs of the predictive models for opioid use and persistent opioid use or OUD were 0.64 and 0.90, respectively. Compared to non-opioid users, intermittent and persistent opioid users were associated with more HCRU and higher healthcare expenditures. The findings could further help clinicians identify pediatric cancer survivors at higher risk of opioid use.