Concordance of provider treatment patterns with NCCN oncology treatment guidelines (NCCN Guidelines) for metastatic non-small cell lung cancer (mNSCLC).

e21574 Background: mNSCLC clinical treatment guidelines are rapidly evolving, however, their influence on prescriber use of anti-cancer treatment regimens is not known. This retrospective study evaluated whether US patients with mNSCLC received first line (1LOT), second (2LOT), or third (3LOT) line treatment regimens that were concordant with NCCN Category 1-2A recommendations. Methods: mNSCLC patients who initiated treatment between 2014-2017 were identified using medical and pharmacy claims from the 100% Medicare Fee-for-Service (FFS) Part A/B/D sample and multi-payer Inovalon MORE2 Registry. Claims-based algorithms identified mNSCLC and LOT following diagnosis; patients with other primary cancers or < 6 months follow-up were excluded. Treatment regimens were compared to active NCCN Guidelines at time of treatment to determine if regimen was concordant with Category 1-2A recommendations. Genetic mutation/biomarker status was not available. Results: 6,523 patients with mNSCLC met criteria for analysis (FFS/Medicare Advantage:78.6%; commercial:14.0%, managed Medicaid:7.4%; mean age:69.3 [SD:9.2]). For 1LOT, 81% received platinum-based doublet regimens; most common monotherapy included erlotinib, nivolumab, and pembrolizumab; 18% of 1LOT patients received maintenance therapy. For 2LOT, approximately one-third received nivolumab monotherapy, one-third received platinum-based doublets, and remainder received other regimens. Most common 3LOT included monotherapy nivolumab, pemetrexed, paclitaxel, and docetaxel. NCCN Guidelines concordance varied: 1LOT (90.5%), 2LOT (58.5%), 3LOT (71.9%). Non-concordant regimens included nivolumab as 1LOT and early adoption of platinum agent+(pemetrexed or paclitaxel) in 2LOT, and chemotherapy as 2LOT in anaplastic lymphoma kinase rearrangement-positive or epidermal growth factor receptor mutation-positive disease. Conclusions: In Medicare and non-Medicare patients with mNSCLC, 1LOT closely followed NCCN Guidelines. Most variability was observed in 2LOT where 58.5% regimens were concordant with Category 1-2A recommendations. Non-concordance may be driven by lack of effective second-line regimens and early adoption of innovative therapies.

Treatment, logistical, and cost burden of prior authorization in community practice.

164 Background: Prior authorization (PA) process is complicated and labor intensive and with limited resources available to community practices, significant opportunity to reduce the costs of therapy exists by transforming this decision support process while being compliant with guidelines Methods: We retrospectively evaluated 123 patients between November 2016 and June 2016 across our three rural oncology practices with no dedicated in-house pharmacy. An established PA workflow involved two physicians, 1 nursing staff and 2 clerical staff involved in the process to manage PA for high cost imaging (PET/CT/MRI) or after a high-risk/cost prescription was initiated by a medical provider to a local pharmacy, specialty or mail-order pharmacy. Two Physicians determined strength of concordance with National Comprehensive Cancer Network (NCCN) guidelines. We examined approval rates, method of communications, time to approval and whether these variables were associated treatment delay and insurance provider Results: Eight decision steps were necessary to obtain a PA. Forty-nine Medicare (39%), 21 managed Medicaid (17%), 29 commercial insurance (23.5%) and 14 commercial insurances (11%) required PA mostly for targeted therapy (29%), oral chemotherapy (28%), PET/CT (24%). Approval rate was 94.3% on first attempt with only 1.6% being denied after an appeal process. 89% concordance with NCCN guidelines was seen. Median time to start treatment was 4 working days (Mean 4.7; Range 0-18 days). PA requests were associated with greatest delay in approval time with managed Medicaid (7 v 1.51 days; p < 0.0001) compared to Medicare, when specialty pharmacy was used (5.2 v 2.3 days p < 0.03) and when fax was used versus online (4.1 v 2.3 days p < 0.02). Conclusions: The impact of the PA process on care delivery and cost reduction are not clear. Current process is particularly burdensome on small practices especially given the high approval rates. Significant opportunity exists to reduce the inefficiency of this system using a computer algorithm-based PA using real-time online decision support system using NCCN guideline concordance