Background: About 5% of NSCLC cases present with anaplastic lymphoma kinase (ALK) fusions and are candidates for an ALK tyrosine kinase inhibitor (ALKi). Evidence on real-world adherence and persistence of ALKi is limited. Objectives: Estimate adherence and persistence of in NSCLC. Methods: This retrospective study included adult (³ 18 years old) NSCLC patients who had an ALKi claim (crizotinib, alectinib, ceritinib, brigatinib) during 7/1/2015-12/31/2018 in commercial claims databases. We categorized ALKi use into ‘first ALKi’ (the first observed ALKi) and ‘subsequent ALKi’ (any following ALKi). The initiation date of each ALKi was set as index date. To ensure complete capture of ALKi use, we included patients who had continuous activity or health plan enrollment for ³12 months pre-index and ³3 months post-index. Adherence was measured by medication possession ratio (MPR). Persistence (time from index to treatment discontinuation (earlier of treatment switch or a ³60-day gap)) and time to next treatment, were compared among ALKis using Cox models to adjust for patient characteristics. Results: This study analyzed 1482 patients whose first ALKi was alectinib or crizotinib (n=445 and 1037, respectively) and 880 patients with subsequent ALKis (604 alectinib, 142 brigatinib, 134 ceritinib). Adherence during treatment period (95-97%) and proportions of MPR ³0.8 (92-95%) were similar across all ALKis. Real-world median time to discontinuation was 27.4 vs 8.9 months for alectinib vs crizotinib in the first ALKi cohort. Patients using alectinib as first ALKi were 46% less likely to discontinue ALKi than crizotinib patients (adjusted hazard ratio (aHR) [95% CI]: 0.54 [0.44-0.65]). For subsequent ALKi use, the risk of discontinuation for alectinib was 64% lower vs ceritinib (aHR [95% CI]: 0.36 [0.27-0.49]) and was 34% lower vs brigatinib but not statistically significant (aHR [95% CI]: 0.66 [0.42-1.02]). Similar patterns were observed for the time to next treatment. Conclusions: Previous data from phase 3 studies have demonstrated superiority of alectinib over crizotinib in first line ALK-positive NSCLC patients. In this real-world study, alectinib was associated with longer treatment persistence and comparable adherence to other ALKis. Further research with larger sample sizes and longer follow-up is warranted to link ALKi persistence with real-world clinical outcomes.