Background: Regulatory efforts in response to the United States opioid epidemic have resulted in a decrease in opioid prescribing among oncologists and non-oncologists. Understanding how decreased opioid prescribing affects cancer-related pain can provide key insights into the downstream effects of these efforts. In this study, we explored temporal trends in opioid use, gabapentinoid use, and pain-related emergency department (ED) visits in older adult patients with advanced cancer and those without cancer. Methods: We queried the Surveillance, Epidemiology, and End Results (SEER)-Medicare database from 1/1/2012 to 12/31/2017 to identify patients aged 66 years or older diagnosed with advanced solid tumor cancer (defined as stage IV and/or metastatic disease). A cohort of patients without a cancer diagnosis was identified using the Medicare 5% random sample. The three dependent outcomes assessed were opioid use, gabapentinoid use, and pain-related ED visits. Multivariable logistic regression was used to identify predictors of these outcomes and to calculate the predicted probability of each outcome for patients with and without cancer. We determined the absolute change in the predicted probability of each outcome from 2012 to 2017. Results: A total of 294,113 patients were included in the cohort; 45,899 (15.6%) with cancer and 248,214 (84.4%) without cancer. Between 2012 and 2017, the absolute change in predicted probability of opioid use declined less in the cancer [-6.1%; 68.6% to 62.5%] compared to the non-cancer cohort [-8.2%; 35.8% to 27.6%] (p<.001 for difference). Gabapentinoid use increased more in the cancer [3.9%; 10.8% to 14.7%] than non-cancer [1.8%; 10.5% to 12.3%] cohort (p<.001 for difference). There was no statistically significant change in pain-related ED visits in patients with cancer [0.6%; 12.4% to 13.0%] and a decrease in patients without cancer [-0.5%; 2.6% to 2.1%] (p=0.005 for difference). Conclusions: Our findings showed that opioid use decreased to a lesser degree among older patients with advanced cancer compared to patients without cancer, whereas gabapentinoid use increased to a greater degree among patients with cancer. There was no corresponding rise in pain-related ED visits, suggesting that these shifts in medication use may not have adversely affected this important outcome.

*p-value <0.001.