Healthcare researchers often need to identify patterns across interviews while preserving clinical context, disagreement, and participant voice. AI thematic analysis can help research teams organize large qualitative datasets, compare perspectives, and surface themes for human review. The November 11, 2026 PLOS ONE study by Andreas Plate and colleagues provides a focused example: researchers analyzed 15 interviews with general practitioners about antibiotic-sparing treatment for uncomplicated urinary tract infections. This article explains what the study found, why the findings matter for healthcare and qualitative researchers, and how AI-enabled analysis can support a transparent workflow without replacing researcher judgment. The discussion is research-focused and does not provide medical advice.
Key Takeaways
The November 11, 2026 PLOS ONE study found that antibiotic-sparing UTI care was shaped by clinical judgment, patient expectations, trust, workload, practice staff, and pharmacy relationships rather than by the GP prescribing decision alone.
AI thematic analysis can help healthcare researchers move from interview transcripts to a structured view of recurring themes, differences between participants, and evidence-backed quotations, while researchers remain responsible for interpretation.
- Between June 10 and July 17, 2025, the study team conducted 15 virtual interviews with GPs in Zurich, Switzerland.
- The interviews lasted an average of 44 minutes, with durations ranging from 32 to 54 minutes.
- The PLOS ONE researchers identified six themes covering prescribing thresholds, trust and expectations, practical constraints, practice staff, professional autonomy, and pharmacy-based care.
- 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 was published on November 11, 2026, and reported that no substantively new relevant content emerged after the 15th interview.
What the 2026 Healthcare Interview Study Found
The PLOS ONE study found that GPs generally supported antibiotic-sparing strategies, but their use depended on symptoms, previous self-treatment, patient expectations, trust, time pressure, access to follow-up, and concern about more serious illness.
The researchers used virtual semi-structured interviews, verbatim transcription, and thematic analysis informed by Braun and Clarke's framework. The analysis combined predefined interview domains with inductive coding developed from the interview material.
The study reported that practice staff influenced UTI care through triage, urine testing, telephone advice, counselling, and local workflows. The study also found that GPs disagreed about how much autonomy practice staff should have and how much oversight should remain with physicians.
The study found that pharmacy-based care was often only partly visible to GPs. Some participants expressed concern about incomplete medical histories and unclear communication, while others viewed pharmacists as an underused source of expertise.
"By now, the antibiotic-sparing approach is actually the approach of choice in our practice, the approach that we largely follow, " said GP10 in the PLOS ONE study, while adding that "aspiration and reality are very far apart."
GP5 told the PLOS ONE researchers that high workload could reduce the time available for explanation: "We cannot spend 10, 15, or 20 minutes explaining things to them."
The authors concluded that clearer professional roles, appropriate training, and feasible communication structures could support more coordinated antimicrobial stewardship in primary care.
Healthcare Interview Findings Snapshot
| Date or measure | Finding | Value | Research implication |
|---|---|---|---|
| June 10 to July 17, 2025 | Interview period | 15 GP interviews | The dataset captured multiple perspectives within one focused qualitative sample. |
| 2025 interviews | Average interview duration | 44 minutes | Thematic analysis had to preserve substantial contextual detail across transcripts. |
| 2025 participants | Interview duration range | 32 to 54 minutes | Transcript length and depth varied across participants. |
| November 11, 2026 | Themes generated | 6 themes | The researchers organized clinical, relational, practical, and interprofessional influences. |
| November 11, 2026 | Mean participant age | 46.6 years | The sample represented experienced primary-care professionals. |
| November 11, 2026 | Female participants | 53.3% | Participant characteristics provide context for interpreting the sample. |
Why AI Thematic Analysis Matters for Healthcare Researchers
How can AI help researchers analyze healthcare interviews?
AI can help researchers organize transcripts into candidate codes, themes, quotations, and cross-participant comparisons, while the research team validates interpretations against the original data.
The PLOS ONE study demonstrates why this support is useful: six themes emerged from 15 interviews, and the themes included both recurring patterns and tensions, such as support for role expansion alongside concern about professional autonomy.
AI analysis should support, not replace, reflexive human interpretation. The PLOS ONE researchers used multiple analysts, reviewed and refined the coding framework, considered divergent accounts, and retained responsibility for the final interpretation.
What should healthcare teams preserve during AI-assisted analysis?
Healthcare teams should preserve participant context, contradictory views, speaker attribution, and the distinction between what participants reported and what researchers inferred.
The PLOS ONE study shows the importance of preserving nuance because GP perspectives differed on workload, pharmacy collaboration, staff autonomy, and the conditions for antibiotic-sparing care.
Healthcare researchers should also protect sensitive interview data. The PLOS ONE authors did not publicly release full transcripts because detailed qualitative data could permit deductive identification despite pseudonymisation.
Which outputs are most useful after thematic analysis?
The most useful outputs include a transparent code hierarchy, theme summaries, supporting quotations, frequency views used cautiously, and comparisons across roles, sites, or participant groups.
For a study like the PLOS ONE research, useful comparisons could include GPs who favored delegated practice-staff roles versus GPs who preferred physician oversight, or participants with strong versus limited pharmacy relationships.
The PLOS ONE authors reported that their analysis combined deductive and inductive elements, making a structured code-to-theme workflow particularly relevant for this type of research.
How Evidano Supports AI Thematic Analysis Healthcare Workflows
Problem: Manual synthesis slows the move from transcripts to themes
Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.
Evidano can organize interview transcripts into thematic, content, frequency, and cross-segment analyses, giving healthcare researchers a structured starting point for reviewing recurring patterns and differences.
Researchers can learn more about the platform's qualitative analysis capabilities on the Evidano features page.
Problem: Multilingual or inconsistent transcripts complicate comparison
Evidano provides transcription with custom dictionaries and PII redaction, which can help research teams prepare sensitive interview material for analysis.
Evidano also provides translation with custom dictionaries, supporting consistent terminology when healthcare research includes multiple languages.
The PLOS ONE study used verbatim transcription and pseudonymised data, so transcript quality, terminology control, and privacy safeguards are directly relevant to comparable workflows.
Problem: Researchers need to interrogate evidence without losing the source context
Evidano provides AI chat over uploaded documents and analyses, allowing researchers to ask targeted questions about themes, participant differences, and supporting evidence.
A healthcare research team could use document chat to examine how participants described workload, patient expectations, or interprofessional communication, then verify each answer against the transcript and coding structure.
Evidano states that data is encrypted and never used to train third-party models; healthcare teams should still apply their own institutional governance, consent, and access-control requirements.
FAQ: AI Thematic Analysis Healthcare Research
What is AI thematic analysis in healthcare research?
+
AI thematic analysis in healthcare research is the use of artificial intelligence to help identify, organize, compare, and summarize patterns in qualitative data such as interview transcripts and open-ended survey responses.
The November 11, 2026 PLOS ONE study used thematic analysis to identify six themes in 15 GP interviews, illustrating the type of complex qualitative dataset that AI tools can help structure.
How many interviews were included in the PLOS ONE UTI study?
+
The PLOS ONE UTI study included 15 general practitioners who were interviewed between June 10 and July 17, 2025.
The researchers reported an average interview length of 44 minutes, with interviews ranging from 32 to 54 minutes.
What themes did the healthcare qualitative study identify?
+
The PLOS ONE study identified six themes: antibiotic-prescribing thresholds, trust and patient expectations, practical constraints, the role of practice staff, professional autonomy, and pharmacy-based care.
The themes showed that antibiotic-sparing UTI management was shaped across the broader primary-care pathway.
Can AI replace human researchers in qualitative healthcare analysis?
+
AI should not replace human researchers in qualitative healthcare analysis because interpretation requires contextual judgment, reflexivity, ethical oversight, and responsibility for claims.
The PLOS ONE research team refined its coding framework through discussion, reviewed the complete dataset, considered divergent accounts, and retained responsibility for analytical decisions.
How should researchers protect healthcare interview data?
+
Researchers should protect healthcare interview data through informed consent, pseudonymisation, restricted access, secure storage, and careful review of whether quotations or combinations of details could identify participants.
The PLOS ONE authors withheld full interview transcripts because detailed qualitative data could permit deductive identification despite pseudonymisation.
Conclusion and Next Steps for AI Qualitative Research
The November 11, 2026 PLOS ONE study shows how 15 GP interviews can reveal the clinical, relational, organizational, and interprofessional factors behind treatment decisions.
AI thematic analysis can help healthcare researchers organize complex interview material, compare participant perspectives, and retrieve evidence efficiently, provided that researchers validate outputs and preserve context.
Evidano brings thematic analysis, cross-segment analysis, transcription, privacy-aware workflows, and AI chat over research documents into one qualitative research platform.
Researchers can Try Evidano for free to explore an AI-enabled workflow for healthcare interviews, open-ended responses, and qualitative documents.
Topics
- AI thematic analysis healthcare
- AI qualitative data analysis
- thematic analysis of healthcare interviews
- qualitative research in primary care
- AI interview analysis
Keep reading
- Commentary on NewsQualitative UTI Research: Six Themes AI Can SurfaceExplore qualitative UTI research through six themes from a 2026 PLOS ONE study, and learn how AI thematic analysis helps healthcare teams act faster.
- Commentary on NewsWhat AI Thematic Analysis Reveals About UTI CareLearn how AI thematic analysis helps healthcare researchers interpret UTI interviews, identify care barriers, and turn qualitative findings into action.
- Commentary on NewsAI Qualitative Analysis in Healthcare: UTI CareSee how AI qualitative analysis in healthcare turns 15 GP interviews into themes on UTI care, antibiotic stewardship, and interprofessional research. Try it.