Payor-driven denials and delays for cancer-related supportive care medications.

Abstract

78 Background: Prior authorization (PA) requires providers to obtain payor pre-approval for a prescribed treatment in order for it to be covered financially. While PA can be used to ensure adherence to evidence-based guidelines and to contain payor costs, it often results in delayed or denied care, which is problematic for patients with cancer who are seeking urgent management of symptoms. This analysis sought to describe PA-related denials and delays for supportive care medications among patients with cancer. Methods: All outpatient prescriptions at an urban comprehensive cancer center requiring PA from 2021-2022 were reviewed to characterize PA for supportive care medications. Prescriptions included long-acting opioids (buprenorphine, fentanyl, hydromorphone, methadone, morphine, oxycodone, oxymorphone); non-opioid analgesics (duloxetine, pregabalin); and anti-emetics (aprepitant, dronabinol, granisetron, ondansetron, scopolamine). Prescription-level outcomes included decision status (approval vs. denial) and days to insurance decision. Bivariate analyses identified variations in PA denials by drug class and payor type. Among a subset of PA approvals for which data were available, we conducted bivariate analyses to evaluate associations of payor type and drug class with length of delay. Results: In the analysis period, 3304 supportive care medication prescriptions required PA: 2489 (75%) long-acting opioids; 131 (4%) non-opioid analgesics; 684 (21%) anti-emetics. Of these, 246 (8%) prescriptions were denied by the payor. The proportion of denied PAs for non-opioid analgesics (n=17; 13%) was greater than the proportion for opioids (n=182; 7%) and anti-emetics (n=17; 2%: χ2=6.5, df=2, p=.04). The proportion of denials among Medicare beneficiaries was greater than the proportion of denials among patients with private insurance, Medicaid, or Medicare Advantage plans (χ2=12.6, df=3, p=.006). Among denials, length from prescription entry to PA final denial ranged from 0-78 days (median=2.0), with no variation by payor type (F=0.85, p=.47). Among approved prescriptions with available data (n=851), 349 (41%) had same-day approval, with delays ranging from 1-54 days (median=2.0). Medicaid patients faced longer delays than patients from other payors (F=4.98, p=.002), with no difference in length of delay by drug class. Conclusions: The acute burden of cancer-related pain and nausea requires immediate mitigation, often with medication. PA can have a negative impact on the medication management of cancer-related pain and nausea: over one-half of supportive care medications had a ≥1 day delay, and 8% were denied altogether. Denials, or even brief delays, of provider-prescribed medication to ease symptom burden may lead to downstream increases in patient morbidity and payor costs, including symptom-related emergency room visits and hospitalization. Urgent efforts to reform PA are needed.