OBJECTIVES:  To analyse the approach of The Agency for Health Technology Assessment and Tariff System (AOTMiT) in Poland to handle crossover (treatment switching) in oncology clinical trials in comparison to the British (NICE), Canadian (pCODR) and Australian (PBAC) HTA Agencies practice. METHODS:  Recommendations issued by AOTMiT for cancer drugs between 2012 and 2014, with substantial rate of crossover in pivotal studies (>50%) were reviewed and compared to appraisals carried out by NICE, pCODR and PBAC. RESULTS:  Four cases were analysed: sorafenib, crizotinib, sunitinib and pazopanib. Crossover ranged between 54 and 75% and caused uncertainty in survival estimation. Crossover correction resulted in more favourable OS estimates and lower ICER/ICUR values. In the AOTMiT appraisal documents economic models based on adjusted estimates in the majority of cases were considered correct, despite noted limitations. Methodological limitations of the clinical study, including crossover, were listed among the reasons for the negative recommendation in case of crizotinib. The remaining HTA agencies pointed out the uncertainty of adjusted survival estimates, however, the cost-effectiveness outcomes based on the adjusted estimates were approved. Progression-free survival estimates were also accepted as a substitute of a biased OS. CONCLUSIONS:  In Poland standards on the approach to the treatment switching are still not established, whereas other HTA agencies consider the adjusted estimates used in economic models as acceptable. This could hinder the consistent interpretation of test results, and thus their utilization in economic models and furthermore result in a delay of reimbursement of effective therapies in Poland.