The study, led by Jerome Ndayisenga of the National Reference Laboratory at Biomedical Centre in Kigali, took the form of a secondary qualitative analysis drawing on semi-structured interviews conducted at two secondary-level hospitals: Nyamata Level Two Teaching Hospital and Nyarugenge District Hospital. The researchers recruited physicians involved in antibiotic prescribing through convenience sampling, continuing until data saturation was achieved. In the end, 24 physicians participated, all of them general practitioners under the age of 35. That demographic detail matters: these are the front-line clinicians who handle the bulk of patient encounters in district and teaching hospitals across Rwanda and much of sub-Saharan Africa, often with limited access to laboratory confirmation and without the safety net of specialist backup.

Using an inductive thematic analysis, the team treated individual codes as units of meaning and grouped them into subthemes and overarching themes. Three major themes emerged from the data, and together they sketch a portrait of clinicians caught between competing risks. The first theme, which the authors labeled physician uncertainty tolerance, captured the participants’ concerns about disease progression and the dangers of withholding antibiotics when a bacterial infection could not be ruled out. For a junior physician watching a febrile patient whose condition might deteriorate overnight, the fear of doing nothing can outweigh the abstract, population-level harm of an unnecessary prescription.

The second theme, avoiding unjustified prescription, revealed the defensive strategies some physicians deploy to resist that pressure. Participants described practicing watchful waiting, ordering additional laboratory investigations such as full blood counts and C-reactive protein tests, and consulting senior colleagues before committing to antimicrobial therapy. These behaviors represent deliberate attempts to buy time and gather evidence, converting an intolerable moment of uncertainty into a manageable, staged decision. Yet the study makes clear that such strategies are unevenly applied, depending on the individual clinician’s risk tolerance, the patient’s presentation, and the practical constraints of the hospital in question.

The third theme proved to be the most consequential for antimicrobial stewardship. Participants described feeling safer to treat with antibiotics, adopting an approach the authors characterized as treat first, adjust later. In this mode, physicians initiate definitive antibiotic treatment despite diagnostic uncertainty, planning to revisit the decision once laboratory results arrive or the clinical picture clarifies. The logic is clinically defensible on an individual basis, sepsis and other bacterial infections can kill within hours, but multiplied across thousands of encounters, it becomes a systematic engine of unnecessary antibiotic exposure.

What unites the three themes, the authors argue, is a form of systematic pressure: the urgency to decide when the evidence is not definitive. In secondary-level hospitals, where microbiology capacity is often limited and antimicrobial susceptibility tests may be unavailable or slow, the clinician’s uncertainty cannot simply be resolved by better data. Laboratory investigations such as full blood counts and C-reactive protein measurements provide probabilistic signals rather than definitive answers, and turnaround times may stretch beyond the window in which a treatment decision must be made. The result is a clinical environment in which prescribing becomes the default escape route from ambiguity.

The study’s grounding in Rwanda gives it particular weight in the global conversation about antimicrobial resistance. Low- and middle-income countries bear a disproportionate share of the AMR burden, driven by a combination of high infectious disease prevalence, constrained diagnostic infrastructure, and, in some settings, over-the-counter antibiotic availability. Secondary-level hospitals occupy a critical position in this landscape: they are the first point of formal medical contact for many patients, staffed largely by junior general practitioners, and equipped with laboratories that fall short of the reference facilities found in national hospitals. Understanding how prescribing decisions are made at this level is therefore essential to any realistic stewardship strategy.

The variability the researchers documented is itself a finding. While some physicians prioritized minimizing unnecessary antibiotic exposure, employing watchful waiting and repeated consultation to hold the line, others focused on preventing potential clinical deterioration, treating early and adjusting later. This heterogeneity means that two patients with similar presentations at the same hospital may receive markedly different antibiotic decisions depending on which clinician sees them. From a stewardship perspective, such inconsistency is a problem in its own right, because it undermines guideline adherence and makes prescribing patterns difficult to predict, monitor, or correct.

The authors’ central recommendation follows directly from their data: treatment guidelines should incorporate clearer, more explicit guidance on managing diagnostic uncertainty. Current guidelines tend to assume a level of diagnostic clarity that often does not exist at the point of care, leaving junior physicians to improvise their own risk calculations. Embedding structured decision pathways, for instance, specifying when watchful waiting is safe, which investigations should precede treatment, and how to sequence empirical therapy with planned review, could standardize prescribing practices and reduce the discretionary variability that the study observed. Such guidance would function as a cognitive aid, redistributing some of the decision-making burden from the individual clinician to the institutional protocol.

Beyond the Rwandan context, the study speaks to a growing recognition within antimicrobial stewardship research that prescribing behavior cannot be changed by guidelines and audit alone. Diagnostic uncertainty is a psychological and structural phenomenon, and it demands interventions that operate at both levels: rapid diagnostics that shrink the uncertainty itself, and training and protocols that help clinicians tolerate the residual uncertainty without defaulting to antibiotics. For the junior physicians at the center of this study, every ambiguous fever is a small ethical dilemma between the patient in front of them and the future patients who will inherit the consequences of resistance. The research suggests that the most effective stewardship programs will be those that acknowledge that dilemma, rather than pretending it can be resolved by willpower alone.

Subject of Research: Diagnostic uncertainty and antibiotic prescribing practices among junior physicians in low-resource settings

Article Title: Diagnostic uncertainties as driver of antibiotic prescription in low-resource settings: secondary qualitative analysis of junior physicians’ perspectives

Article References: Ndayisenga, J., Mugabo, E., Ntiyamira, S., Bugingo, J. P., Musabeyesu, G., Sibomana, O., Umuhoza, A., Bimenyimana, P. C., Igizeneza, A., Gahamanyi, N., Gashegu, M., Mukagatare, I., Ngabonziza, J. C. S., & Fitch, M. I. (2026). Diagnostic uncertainties as driver of antibiotic prescription in low-resource settings: secondary qualitative analysis of junior physicians’ perspectives. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15709-2

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15709-2

Keywords: antimicrobial resistance, diagnostic uncertainty, antibiotic prescribing, Rwanda, qualitative research, antimicrobial stewardship, junior physicians, secondary-level hospitals, low-resource settings, clinical decision-making, Diagnostic, uncertainties