What UTI Interviews Reveal: AI Thematic Analysis

The PLOS ONE study, published on November 11, 2026, found that antibiotic-sparing management of uncomplicated UTIs was shaped by clinical thresholds, patient expectations, trust, workload, practice staff, and pharmacy relationships.

The study shows why AI thematic analysis is useful in healthcare research: the important finding was not a single prescribing rule, but a network of recurring influences across the wider care pathway.

• Fifteen general practitioners took part in virtual semi-structured interviews conducted between June 10 and July 17, 2025, according to the 2026 PLOS ONE study.
• The researchers reported six themes covering prescribing thresholds, trust and expectations, practical constraints, practice staff, professional autonomy, and pharmacy-based care.
• The interviews lasted an average of 44 minutes, with a range of 32 to 54 minutes, according to the study’s results.
• The participating GPs had a mean age of 46.6 years, 53.3% were female, and their general-practice experience ranged from 5 to 25 years.
• The study concluded that clearer roles, appropriate training, and feasible communication structures may support more coordinated antibiotic-sparing care.

The 2026 PLOS ONE study found that antibiotic-sparing UTI care was a negotiated and context-dependent process rather than a simple prescribing decision. The researchers interviewed GPs who routinely managed uncomplicated UTIs in women and analyzed verbatim transcripts using thematic analysis informed by Braun and Clarke’s framework.

The study’s six themes connected clinical reasoning with relational and organizational conditions. GPs described considering symptom severity, duration, diagnostic findings, patient expectations, trust, access to follow-up, workload, weekends, holidays, and previous self-treatment.

One GP described the implementation gap directly: “By now, the antibiotic-sparing approach is actually the approach of choice in our practice... However, it is still often the case that aspiration and reality are very far apart.” The quotation appears in the study’s results and illustrates why qualitative analysis must capture tensions, not only recurring keywords.

The study also found that practice staff influenced care through triage, urine testing, counselling, and informal routines. GP views differed on whether staff should support care under physician oversight or manage selected cases with greater autonomy.

Pharmacy-based care was often only partly visible to GPs. One participant said, “I have no idea what’s going on, ” referring to limited information about pharmacy treatment decisions, antibiotics, dosage, and duration. The authors interpreted this limited visibility as a factor that could shape professional perceptions of pharmacy-led care.

Healthcare qualitative researchers should analyze treatment decisions as part of a care pathway when interviews reveal influences distributed across clinicians, staff, patients, and pharmacies. The PLOS ONE study demonstrates that a narrow code such as “antibiotic prescribing” would miss the role of trust, time pressure, patient self-treatment, professional boundaries, and information exchange.

Researchers can strengthen AI thematic analysis by combining deductive and inductive coding. The study began with predefined domains covering treatment decisions, non-physician staff, and pharmacies, then developed additional codes from the interviews and refined themes against contradictory and divergent accounts.

Healthcare teams should also preserve traceability from theme to evidence. The study used illustrative quotations and a finalized codebook, while the underlying transcripts were not publicly shared because detailed qualitative data could permit deductive identification of participants.

• Use segment comparisons to examine whether experienced and less experienced clinicians describe different prescribing thresholds.
• Separate stated intentions from reported practice, because the study identified a gap between antibiotic-sparing aspirations and routine constraints.
• Code enabling and limiting conditions together, including trust, follow-up access, workload, patient expectations, and communication structures.
• Retain verbatim quotations beside codes so decision-makers can inspect how an interpretation is grounded in participant language.
• Treat AI-generated themes as research aids that require human review, especially in clinical, sensitive, or potentially identifiable datasets.

The 2026 PLOS ONE study shows that antibiotic-sparing UTI care is shaped by more than the GP’s prescribing decision. Six themes revealed how clinical judgment, patient expectations, trust, workload, practice staff, professional autonomy, and pharmacy communication interact across the care pathway.

AI thematic analysis can help healthcare researchers manage this complexity by connecting codes, themes, quotations, and participant segments in a traceable workflow. Human researchers must still validate interpretations, protect sensitive data, and distinguish exploratory findings from clinical recommendations.

Teams analyzing interviews, open-ended surveys, or healthcare documents can try Evidano for free to accelerate qualitative synthesis while keeping research judgment at the center.