LAUSR

would be this: In what percentage of cases could this variability lead physicians to change their therapeutic decision (e.g., the indication of CPAP or alternative treatments)? To make this calculation, the authors would need access to individual patients’ clinical data. We do not know howmany of the studied patients were treated with CPAP (or alternatives), but even without this information, the authors could make an estimate based on the indications for treatment specified in the different international guidelines (5). On the other hand, the authors could offer additional clinically relevant information by assessing theminimumnumber of consecutive sleep studies needed tominimize their variability and thus reduce costs. The authors observe that themore sleep tests that are performed, the more the AHI stabilizes, according to an analysis of the different areas under the receiver operating curve (ROC) curves at differentmoments in time. Thus, it can be seen that, although the variability was very high in the first days, it wasmuch lower after 1 week and evenmore so after 2 weeks, although the results obtained onDays 7 and 14 were very similar. The authors could calculate the relevant differences between the various areas under the ROC curves from clinical and statistical viewpoints. This information would be of enormous clinical relevance, as it would indicate theminimumnumber of days required for sleep tests to obtain an optimal balance between the least variability that would allow a minimumnumber of clinically acceptablemisclassifications (and, above all, a minimumnumber of relevant therapeutic changes) and lower costs, less time tomake an accurate diagnosis, and fewer resources. Therefore, we believe that it would be very interesting and enriching for the study, and for clinicians who care for patients with OSA, if the authors could contribute these suggested new analyses to their already excellent study.