What AI Qualitative Analysis Reveals About UTI Care

The General practitioners’ perspectives on treatment decisions and interprofessional care for women with uncomplicated urinary tract infections: An exploratory qualitative study, published in PLOS One on November 11, 2026, found that antibiotic-sparing UTI care depended on clinical judgment, patient expectations, trust, workload, and coordination across primary care.

The study used 15 virtual GP interviews conducted between June 10 and July 17, 2025, and thematic analysis produced six themes covering prescribing thresholds, relationships, practical constraints, practice staff, professional autonomy, and pharmacy-based care.

• According to the November 11, 2026 PLOS One study, 15 GPs participated, with a mean age of 46.6 years and 53.3% identifying as female.
• According to the study authors, interviews lasted a mean of 44 minutes, with a range of 32 to 54 minutes, and later interviews produced no substantively new content relevant to the research questions.
• The PLOS One researchers found that GPs generally supported antibiotic-sparing strategies, but prescribing thresholds varied according to symptoms, patient expectations, trust, time pressure, and access to follow-up.
• The study found that practice staff influenced UTI management through triage, urine testing, counselling, and care routines, while GPs disagreed about how much professional autonomy staff should have.
• The researchers concluded that clearer roles, appropriate training, and feasible communication structures could support more coordinated antibiotic-sparing care across general practices and community pharmacies.

The PLOS One research team examined how GPs make treatment decisions for uncomplicated UTIs in otherwise healthy, non-pregnant women aged 18 to 64 years, while also investigating the roles of non-physician staff and community pharmacies.

The researchers recruited GPs who served as teaching physicians at the Institute of Primary Care in Zurich, Switzerland, from a pool of approximately 500 registered teaching physicians. The recruitment period ran from April 23 to May 15, 2025, and the interviews took place between June 10 and July 17, 2025.

The study team conducted virtual semi-structured interviews, transcribed the recordings verbatim, and analyzed the material thematically using both deductive and inductive coding. The researchers developed an initial coding framework from three interviews, refined the framework through discussion, and applied the finalized codebook across the dataset.

The study authors described antibiotic-sparing care as a conditional process rather than a simple prescribing decision. GP10 summarized the implementation challenge by saying, “aspiration and reality are very far apart.”

The PLOS One researchers also found that trust could make conservative treatment easier to negotiate. GP14 explained, “If I already know her, I feel that the trust is there; she will come back if it does not get better.”

Healthcare researchers should analyze treatment decisions as part of a broader care pathway when interviews show that clinical choices are shaped by relationships, workflows, and professional boundaries.

The PLOS One study shows why surface-level frequency counts are insufficient for this topic: the same antibiotic-sparing strategy could be supported by trust in one account, constrained by workload in another, and limited by uncertainty about follow-up in a third.

Researchers studying clinical services, patient journeys, or antimicrobial stewardship should compare themes across roles and settings rather than treating the GP consultation as the only decision point. The study authors identified interactions among GPs, practice staff, patients, and community pharmacists as central to understanding treatment decisions.

UX researchers and service designers can use this type of qualitative evidence to map where expectations are formed, where information is lost, and where handoffs between professionals create uncertainty. The PLOS One study specifically reported that pharmacy-based care was often only partly visible to GPs.

AI-assisted analysis should support, not replace, researcher interpretation in healthcare research. The original study used researcher-led coding, iterative refinement, and reflexive discussion, and the authors noted that full transcripts could not be publicly shared because detailed qualitative data might permit participant identification.

The November 11, 2026 PLOS One study demonstrates how qualitative healthcare research can uncover the clinical, relational, organizational, and interprofessional forces behind treatment decisions. Its 15 interviews and six themes show why healthcare researchers need more than isolated keyword counts when analyzing complex service pathways.

AI qualitative analysis can help research teams structure transcripts, compare participant segments, identify recurring themes, and return to supporting evidence more quickly. Human researchers must still validate interpretations, account for contradictory perspectives, and protect sensitive participant data.

For teams analyzing healthcare interviews, open-ended surveys, or clinical service feedback, Try Evidano for free and explore a more efficient way to move from qualitative data to defensible research insight.