For a study, researchers sought to identify patterns of long-term opioid treatment (LTOT) and their correlates in patients with and without cancer and analyze the relationship between prescription patterns and risk for opioid overdose, taking into account the potential moderating influence of cancer.

A retrospective cohort research from 2010 to 2017 used latent growth mixture modeling and Cox proportional hazards modeling. The setting includes the electronic health record database of the Veterans Health Administration. Patients with incident LTOT with and without cancer were included (N=44,351 and N=285,772, respectively). Veterans in the cohort were ≥18 years old at the time of incident LTOT; had had VHA care, defined as ≥2 outpatient visits or ≥1 inpatient hospitalization during the year before cohort entrance; and had received LTOT, defined as 90 consecutive days’ opioid prescription. The pattern of opioid prescriptions was obtained as part of the exposure. The main outcomes comprised all opioid-related overdose episodes, including accidental and purposeful.

Low-dose, low-dose/de-escalating trend, moderate-dose, moderate-dose/escalating trend, and high-dose/escalating trend were the opioid reception patterns seen in patients without cancer and duplicated in patients with cancer. Controlling for variables, higher-dose and escalation trajectories substantially predicted the time to first overdose; the two low-dose patterns imparted comparable, reduced risk. The cancer group was unaffected effects of trajectories on time to overdose. People with cancer suffer the same risks of opioid overdose as patients without illness.