Search for a command to run...
Older adults with delirium and bacteriuria or pyuria without systemic signs or symptoms of [urinary tract] infection are frequently treated with antibiotics despite a lack of evidence that antibiotics improve delirium resolution [1, 2]. A randomized controlled trial (RCT) is needed, but there is no consensus on the acceptable difference in delirium resolution between those who do and do not receive antibiotics. We conducted a survey of healthcare providers to identify an acceptable non-inferiority margin to guide the design of an RCT. We developed an electronic survey via Survey Monkey (San Mateo, CA) to inform the design of an RCT evaluating whether routine care without antibiotics in older adults with delirium and bacteriuria or pyuria is non-inferior to routine care with antibiotics using a primary outcome of delirium resolution at Day 7. We recruited a convenience sample of Canadian academic and community healthcare providers in the fields of general internal medicine, geriatric medicine, and infectious diseases in November 2025 through electronic mailing lists of relevant divisional and professional societies. The survey assessed acceptable non-inferiority margins for the trial's primary outcome—resolution of delirium by Day 7. Using a reference rate informed by the pilot phase of this trial, in which 40% of older adults with delirium experience resolution by day 7 [3], clinicians were asked which resolution rate (32%, 30%, 28%, 26%, 24%, 22%, 20%, or free text) would be acceptable for the no-antibiotic arm. A second question assessed thresholds for antibiotic-associated harms (e.g., Clostridioides difficile infection) that would justify cessation of the trial. Ethics approval was obtained from the University of Toronto's Human Participant Research Ethics Board (protocol # 49005). Among 177 healthcare providers who received the introductory email, 59 individuals (33%) opened the survey with 50 (28%) ultimately completing it. Table 1 shows the characteristics of survey participants, most of whom were staff physicians (45/50, 90%), followed by residents (3/50, 6%) and pharmacists (2/50, 4%) and nearly three-quarters practiced in academic settings (37/50, 74%). The practice setting of the 50 respondents was 40% geriatric medicine (20/50), 44% infectious diseases (22/50), and 16% general internal medicine (8/50). Most respondents (48%, n = 24/50) selected “≥ 30% delirium resolution,” corresponding to a 10% non-inferiority margin, which was followed by 26% who selected a margin less than 10%, 12% who selected one between 10% and 20% and 14% who selected a margin of 20% (Figure 1). Stratifying responses by specialty showed similar results. For antibiotic-associated harms, respondents most frequently endorsed stopping the trial if the absolute increased risk was ≥ 2% (20/50, 40%), with almost as many (18/50, 36%) choosing a threshold of ≥ 4% and 14% (7/50) choosing ≥ 3% (Table S1). Our survey provides a healthcare provider-derived estimate of an acceptable non-inferiority margin for resolution of delirium for a clinical trial, with an absolute difference of 10% being the most commonly selected margin. Additionally, most healthcare providers selected low thresholds to stop the trial based on antibiotic-associated harms (≥ 2%), further emphasizing a desire to avoid preventable antibiotic risk. The 10% non-inferiority margin for delirium resolution by day 7 mirrors those used in recent infectious disease randomized trials [4, 5]. Assigning a similar acceptable margin to delirium resolution by day 7 supports the idea that healthcare providers view delirium as an important condition with severe consequences. Our finding of a 10% acceptable difference in treatment effectiveness is also aligned with the FDA's guidance on the design of non-inferiority drug trials [6]. The willingness to accept a decrease in treatment effectiveness to avoid antibiotic risk may reflect the growing awareness that antimicrobial resistance is a major threat to public health and antibiotic harms are consequential. These concerns have fueled a growing body of evidence demonstrating that shorter durations of antibiotics are as effective as longer durations [7]. Whether this extends to withholding antibiotics for patients with delirium and bacteriuria or pyuria and no signs of a genitourinary tract infection is an important and unanswered question [8]. Limitations of our study include the fact that this was a convenience sample survey of Canadian healthcare providers with lower-than-expected representation from GIM, which raises the issue of volunteer bias and may impact generalizability. While the response rate was modest, it is comparable to other published surveys of clinicians [9]. Our findings support a non-inferiority margin of 10% for delirium resolution by day 7 and a ≥ 2% threshold to stop the trial based on antibiotic-associated harms for a planned multicentre randomized trial. These results can also be used to inform non-inferiority margins for delirium resolution in future clinical trials. All authors listed have contributed significantly to study conception and design and/or data acquisition and/or analysis and interpretation of the data. All authors were involved in the manuscript's preparation, revision and approval of version to be published. Study concept and design: M.F., C.K., N.M.S., C.R.-R., O.B., E.M. Acquisition of subjects and/or data: M.F., C.K., G.P., E.H.-S. Analysis and interpretation of data: M.F., C.K., G.P., E.H.-S., C.R.-R., O.B., E.M. Preparation of manuscript: All authors. This work was supported by Physicians' Services Incorporated Foundation (#R24-23), Sinai Health Foundation's Healthy Ageing Initiative, and Mount Sinai Hospital Department of Medicine Research Fund. The sponsors were not involved in the design, methods, data collection, analysis, or preparation of the survey or paper. The authors declare no conflicts of interest. Table S1: Responses to non-inferiority margin and risk of antibiotic harm questions. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.