OBJECTIVES:   Guideline recommended initiated therapy (GRI) for mNSCLC has been associated with lower outpatient costs. We evaluated the relationship between GRI and total healthcare costs and utilization. METHODS:   mNSCLC patients that initiated infusion therapy in 2013-14 were identified from Humana claims and treatment authorizations. A claims-based model predicting disability was used to approximate performance status (PS) from procedure, diagnosis, and durable medical equipment codes. Patients receiving GRI with >1 cycle of a National Comprehensive Cancer Network recommended infused first-line therapy based on age, PS, or targeted therapies regardless of age and PS, were included in the study. Index date was date of first infusion. P-values were generated from chi-square and t-tests. Costs were presented as mean differences. A generalized linear model controlling for pre-index costs, age, gender, region, comorbidity burden, PS, plan type, months of follow-up and treatment regimen assessed the relationship between GRI and cost. RESULTS:   Of 1,458 patients, 347 (23.8%) were non-GRI. Non-GRI were older (71.9 + 7.4 vs 70.0 + 7.5 years, P<0.0001), more frequently dual-eligible for Medicare/Medicaid (23.1% vs 10.5%, P<0.0001), and had low-income subsidy (30.3% vs 16.7%, P<0.0001). Pre-index, non-GRI had higher inpatient (+$2,517, P=0.0096), emergency room (ER) (+$226, P=0.010), and lower outpatient costs (-$1,739, P=0.0002). Post-index, non-GRI had more inpatient stays (49.2% vs 58.5%, P=0.0026) and ER visits (55.0% vs 62.8%, P=0.0102). GRI had greater post-index total (+$8,807, P=0.0005), oncology-related (+7,449, P=0.0018), physician office (+5,612, P=0.0006) and pharmacy costs (+1,501, P=0.0157). In a multivariable model controlling for patient-level characteristics, no cost differences by GRI were found. CONCLUSIONS: