Second, in our study, there was a certain amount of bias in the three-group classification by the type of bystander CPR. In our study, conventional CPR tended to be associated with the greatest extent of improved outcomes, followed by no CPR and then chest-compression–only CPR. Although it would seem impossible that no CPR would have a better prognosis than chest-compression–only CPR, there are several conceivable causes for this finding in our study. One reason is the difference in baseline characteristics between the chest-compression–only CPR group and the no CPR group. The chest-compression–only CPR group had more factors contributing to a poor outcome; this group had more elderly people, fewer witnesses present, and more cases wherein advanced pre-hospital airway management was performed. There may also be other hidden differences between the two groups. Another cause could be that chest-compression–only CPR performed during the study period was not as effective as the recommended and validated chest-compression–only CPR: the chest-compression– only CPR performed before it was widely recommended and prevailed was considered to be low-quality CPR performed by a bystander who was not sufficiently familiar with CPR techniques. To more accurately assess the relationship between the type of bystander CPR and the prognosis of OHCA caused by respiratory disease, further prospective studies will be required that align the baseline characteristics in each of the bystander CPR groups in a state where the concept of chest-compression–only CPR has been disseminated to the general population.