Our study is certainly not without weaknesses. It is a single centre retrospective study with its attendant biases. However, given the current uncertainty on the optimum management of the severe COVID-19 respiratory failure, as well as the lack of robust data on CPAP in patients who are not fit for ITU escalation, it would be unethical to randomise patients, potentially depriving them from a widely accepted form of treatment. Firstly, one might argue that the level of care for patients who had ‘ward-based’ CPAP was not ‘on par’ with those who received CPAP in an ITU environment in terms of monitoring and ‘nurse-to-patient’ ratio, potentially having an impact on outcomes. However, it must be noted that both the patients whose CPAP was managed in the ITU during the initial phase of the pandemic did not survive either. Outside of ITU, the patients were managed in designated areas with expertise in dealing with non-invasive ventilation. The mortality was also similarly high in the prospective group who were treated later on during the pandemic, when the breadth of expertise in dealing with CPAP was broader. In our hospital, during the COVID-19 pandemic we established a 1:3 ‘nurse-to-patient’ ratio for our ‘Level 2’ areas, with continuous monitoring of vital signs and early warning score (EWS), allowing us to maintain high patient safety standards. To our knowledge, in recent published data referring to increased demand of care for COVID-19 patients in all areas, the ‘nurse-to-patient’ ratio has been either similar to ours or even less intense, even in ITU environments due to dilution of staff under the revised COVID recommendations, as well as the inevitable surge [5]. Additionally, one might argue that the patient selection might have been inappropriate, impacting outcomes. To minimise that possibility, we followed all current BTS and Intensive Care Society recommendations for CPAP patient selection and treatment strategies [6].