The sample sizes vary dramatically across sections (n=407 for preparedness, n=376 for confidence, n=324 for attitudes, n=328 for interprofessional roles). Why were these incomplete responses included at all if “each section was analyzed if completed in full”? This suggests inconsistent inclusion criteria. More importantly, the decreasing n suggests systematic non-response to later sections, likely indicating respondent fatigue or disengagement, which introduces significant bias that you fail to address.
You state the questionnaire was distributed to 1602 students (100% of eligible students), with 518 responses (32.3%) and 407 valid (25.4%). However, on page 3, you report that “teachers were requested to provide the number of students to whom the questionnaire had been distributed”; this means the 1602 figure is an estimate, not an actual count. How can you claim a 25.4% response rate when you don’t actually know how many students received the questionnaire? This is a fundamental methodological flaw.
1. With a final response rate of only 25.4%, self-selection bias is unavoidable, respondents were likely more interested in deprescribing, artificially inflating positive attitudes and perceived preparedness. Furthermore, Table 1 reveals severe clustering: 15 out of 24 universities contributed fewer than 10 students, while just 3 universities (Lyon, Paris Cité, Clermont-Ferrand) supplied nearly 36% of the sample. This is not a nationally representative sample. How did you account for this clustering, and why did you not perform weighted analyses or sensitivity tests excluding these small, unrepresentative subgroups before claiming “national” generalisability?
2. You state: “58% of the respondents disagree with the statement ‘deprescribing is safe,’ yet 72% disagree with ‘deprescribing has no benefit’.” However, Table 2 contains no item measuring “deprescribing has no benefit.” The closest item is “Deprescribing PIMs is valuable to patients”—where 97.5% agreed, not 72%. This appears to be a complete misreading of your own results. Please clarify which item you actually meant; if this is an error, it undermines the reliability of your discussion and suggests careless data interpretation.