Background: In the last five years, different immunotherapy (IO) clinical trials have improved overall survival (OS) for stage IV non-small cell lung cancer (NSCLC) without driver mutation. However, it is necessary to check whether outcomes of clinical trials are reproduced in the general population, the true objective of any scientific advance. Methods: We conducted a retrospective observational study selecting all patients with stage IV NSCLC without driver mutation treated in the Medical Oncology Department at Puerta de Hierro University Hospital between 2016 and 2019. Data cut off was June 30, 2021. Our goal was to evaluate treatment changes for NSCLC in clinical practice in our institution with OS in the real-world as the main outcome, in addition to assessing quality of care. Results: Between 2016 and 2019, 317 patients with stage IV NSCLC were diagnosed, 260 without driver mutation and 189 of them received systemic treatment. We observed a higher proportion of stage IVB patients in 2018 and 2019 compared to 2016 and 2017. A progressive greater use of IO was observed, except for a small decrease in 2019 compared to 2018. At the date of analysis, 77.8% of patients had died. With a median follow-up of 40.8 months, we observed a gradual increase in OS, with medians of 12.3, 13.6, and 18.9 months for patients diagnosed in 2016, 2017, and 2018, respectively, with a decrease to 10.4 months for patients diagnosed in 2019. Survival rates over time show the same trend. IO (any line) was significantly associated with a reduced risk of death: HR 0.35 (95% CI: 0.23 - 0.58) in multivariate analysis. The deterioration of the results of the patients diagnosed in 2019 is manifested in the median survival and in the survival rates at different points from 6 months onwards, which should correspond to the excess mortality in the year 2020 in our studied population. Conclusions: Our study shows that IO reduced risk of death: HR 0.35 (95% CI: 0.23 - 0.58) in stage IV NSCLC, observing an increase in overall survival from a median of 12.3 to 18.9 months for patients diagnosed in 2016 and 2018, suggesting that the progressive use of IO for NSCLC in the years 2016, 2017 and 2018 has transferred the benefit observed in clinical trials to the general population. The deterioration of outcomes for patients diagnosed in 2019, with a median OS of 10.4 months and decreased survival at 6, 12 and 18 months, could suggest the negative impact and excess mortality in the pandemic year 2020.