Introduction and Objectives: Litigation for health care, also known as health judicialization, is frequent in Brazil. It involves recourse to the court system to access health services. The study aimed to evaluate whether cancer patients in Belo Horizonte, Minas Gerais, Brazil, increased their overall survival by increasing access to certain drugs or treatments through litigation, controlling for the effect of demographic and disease-related variables. Methods: Patients with breast, prostate, brain, lung, or colon cancers from 2014 to 2019 were included. A retrospective cohort study was conducted. Survival analysis was performed using the Cox proportional hazards model. Results: In the multivariate analysis, litigation was significantly associated with increased survival in cancers of breast (HR=0.51, 95%CI 0.33-0.80), prostate (HR=0.50, 95%CI 0.30-0.85), colon (HR=0.59, 95%CI 0.38-0.93), and lung (HR=0.36, 95%CI 0.22-0.60). Five-year survival rates of patients who sued for treatment were 97.8%, 88.7%, 59.3%, and 26.0%, compared to median survival of 95.7%, 78.7%, 41.2%, and 2.4%, respectively, among patient that did not resort to court action. The study suggests that litigation for access to cancer treatment may represent a step forward in obtaining more effective treatment. This study main limitations are the lack of patient clinical information and data regarding to the patients’ quality of life. The study also found that many cases involved clAims that could have been solved by administrative rather than legal action. Some clAims thus reflect the lack of adequate administrative procedures. Conclusions: When based on scientific evidence, access to new therapies, combined with other technologies already available, can favor patient survival. Access to new therapies through litigation may increase health inequalities since low-income patients have limited access to legal recourse against the State to meet their needs. The timely approval of new effective therapies can mitigate the judicialization of cancer treatment.