OBJECTIVES: LOT rules for determining treatment patterns utilizing electronic medical record (EMR) data are a challenging endeavor for cancer with active and maintenance lines.  This study evaluated the impact on treatment duration of assigning different rules to determine active and maintenance LOT in a study cohort of lung cancer patients with stage IIIB and IV NSCLC.   METHODS: Patients who were diagnosed from October 1, 2012 to December 31, 2014, 18 years or older and treated with bevacizumab and/or pemetrexed in combination with taxane/platinum in first line treatment were included in this retrospective study from a community oncology EMR dataset.  Mean and SDs were calculated for treatment duration on line one active and maintenance for three scenarios: 1) platinum doublet +/- Bev (PDB) until all drugs discontinued; 2) PDB with maintenance counted when regimen reduced to a single agent; 3) PDB counted active through five administrations (standard LOT1 4-6) with maintenance counted when platinum discontinued. RESULTS: The study cohort consisted of 626 patients, with a mean age (at first LOT) of 62.6 and 44% females.  Scenario 1 mean active duration was 4.32 months (+/- 4.19).  Scenario 2 mean active duration was 2.84 months (+/-2.09) and mean maintenance duration was 4.02 months (+/- 4.09).  Scenario 3 mean active duration was 3.39 months (+/- 2.62) and mean maintenance duration was 3.90 months (+/- 4.39).  CONCLUSIONS: This research illustrated the impact of different rules on calculating the duration of therapy per line.  Consensus on the rules utilized should be agreed a priori to the research and should be consistent across different studies in the same cancer diagnosis.  Difficulties in comparing treatment duration outcomes can be avoided proactively when conducting further comparative effectiveness research.