Primary care researchers need clear methods to study clinician attitudes to large language models. The primary keyword for this post is "LLM adoption qualitative analysis" and this post explains, for qualitative researchers and UX teams, what the PLOS One protocol specifies and how AI-enabled qualitative research platforms speed rigorous thematic analysis. According to PLOS One, Pokharel et al. published the study protocol on August 7, 2026, to explore primary care physicians' perceptions of LLM adoption in clinical practice.
Key Takeaways
According to PLOS One, Pokharel et al. (2026) published a qualitative study protocol on August 7, 2026, that will use semi-structured interviews and the Technology Adoption Behavior (TAB) framework to examine how family physicians in British Columbia perceive LLM adoption. The protocol states recruitment is planned from August 1, 2025 to January 31, 2026 and analysis will combine deductive and inductive thematic coding using NVivo.
- The protocol was published on August 7, 2026, according to PLOS One (Pokharel et al., 2026).
- Pokharel et al. plan purposive sampling of physicians with three LLM-use categories: 50+ uses, 5–50 uses, and fewer than 5 uses, described in the Methods section of PLOS One.
- The protocol reports that the number of Canadians without a family doctor rose from 4.5 million in 2019 to about 6.5 million in 2023, a statistic the authors use to justify focusing on primary care (PLOS One).
- The research team received ethical approval from the UBC Behavioral Research Ethics Board under protocol H25-01658, as stated in PLOS One.
- Pokharel et al. say the study will use the TAB framework to unify perceived usefulness, perceived ease of use, and task–technology fit in data collection and analysis (PLOS One).
What Happened: Study design and measures
What happened: Pokharel et al. published a qualitative study protocol on August 7, 2026 in PLOS One to examine primary care physicians' perceptions of LLM adoption in clinical practice.
According to PLOS One, the study uses a descriptive qualitative approach with semi-structured interviews conducted in British Columbia, Canada, and the authors plan purposive sampling to include physicians with frequent (50+), occasional (5–50), and minimal (<5) LLM use.
According to PLOS One, the protocol centers the Technology Adoption Behavior (TAB) framework to integrate perceived usefulness, perceived ease of use, and task–technology fit when designing interview prompts and coding rubrics.
According to PLOS One, interviews will be recorded, transcribed verbatim, deidentified, and analyzed iteratively with NVivo using both deductive TAB-based codes and inductive emergent codes.
Findings Snapshot
| Date | Metric | Value | Implication |
|---|---|---|---|
| August 7, 2026 | Protocol publication | PLOS One | Public protocol establishes methods and timeline for qualitative data collection |
| August 1, 2025 to January 31, 2026 | Planned recruitment window | 6 months (anticipated) | Dataset expected to be collected within a defined, short timeframe to capture contemporary clinician views |
| 2019 → 2023 | Canadians without family doctor | 4.5 million → 6.5 million | Authors cite access pressures to justify studying primary care clinicians' technology adoption |
| Study design | LLM-use categories | 50+, 5–50, <5 lifetime uses | Purposive sampling to capture variation in experience |
| Ethics | Review board | UBC Behavioral Research Ethics Board, H25-01658 | Confirms planned protections: consent, deidentification, audio destruction |
Implications for qualitative researchers and health services teams
Answer: The PLOS One protocol shows how to design a timely, theory-informed qualitative study of clinician technology attitudes and offers concrete design elements researchers can reuse.
According to PLOS One, Pokharel et al. combine deductive TAB-derived codes with inductive coding, which implies teams should pre-map likely constructs such as perceived usefulness and task fit, then remain open to emergent themes during analysis.
According to PLOS One, purposive sampling across predefined LLM-use strata (50+, 5–50, <5) provides a practical model for ensuring variation in experience; qualitative researchers can adapt those cutoffs to other clinician populations to capture novice to expert perspectives.
According to PLOS One, reflexive memos and a multidisciplinary team composition (medical trainee, practicing physician, educators, economist, health systems researchers) are central to richer interpretation and can be implemented in other qualitative LLM adoption studies.
How Evidano Helps
Problem: Slow synthesis of interviews to policy-ready recommendations
Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.
Evidano feature mapping: Use Evidano's thematic analysis and hierarchical coding to convert semi-structured interview transcripts into organized codes and themes faster than manual spreadsheets, matching the PLOS One plan to combine deductive and inductive coding.
Evidano outcome: Researchers following the PLOS One protocol can shorten the time from transcription to theme summary while preserving reflexive memos and audit trails via Evidano's codebook export, which supports methodological transparency.
Problem: Manual transcription and PII handling for recorded interviews
According to PLOS One, interviews will be recorded, transcribed, and deidentified; Evidano supports automated transcription with custom dictionaries and PII redaction to align with that workflow.
Evidano feature mapping: Use Evidano's speech-to-text workflow to transcribe interviews, apply a custom medical dictionary, and redact personal identifiers before coding, matching the ethics protections described in the protocol (Evidano speech-to-text).
Problem: Need to compare subgroups (e.g., LLM experience strata)
According to PLOS One, the protocol samples physicians across three LLM-use strata; Evidano's cross-segment analysis lets teams compare theme frequencies and co-occurrence networks by those exact strata.
Evidano feature mapping: Exportable visualizations and segment-level frequency tables help translate qualitative findings into briefings for policymakers and professional bodies, as Pokharel et al. plan to do in their dissemination strategy.
Problem: Secure, research-grade data handling
Evidano maintains encrypted data storage and does not use customer data to train third-party models, supporting the ethical deidentification and data handling described in the PLOS One protocol (Evidano data security).
FAQ: LLM adoption qualitative analysis
What is the Technology Adoption Behavior (TAB) framework and why does the PLOS One protocol use it?
Answer: The TAB framework unifies perceived usefulness, perceived ease of use, and task–technology fit to explain technology uptake.
According to PLOS One, Pokharel et al. developed TAB to integrate established models like TAM and task–technology fit, and they use TAB to structure interview prompts and deductive codes for thematic analysis.
How should researchers sample clinicians when studying LLM adoption?
Answer: Purposive sampling across experience strata provides analytic variation and richer comparative insights.
According to PLOS One, the authors plan to sample family physicians in British Columbia with three experience categories (50+, 5–50, <5 lifetime LLM uses) and to aim for sufficiency rather than statistical representativeness.
Can AI tools safely speed transcription and coding while preserving rigor?
Answer: Yes, when tools include PII redaction, custom medical dictionaries, and audit logs to preserve transparency.
According to PLOS One, Pokharel et al. plan deidentification and audio destruction after transcription, and Evidano provides speech-to-text with custom dictionaries and PII redaction features to support those ethics-aligned workflows.
What are common risks qualitative teams should flag when studying clinician attitudes to LLMs?
Answer: Key risks include bias amplification, overreliance on LLM outputs, and shifting clinician roles, which require careful probing and reflexivity.
According to PLOS One, the authors highlight possible harms such as bias propagation and false information and they recommend reflexive memos and a multidisciplinary team to mitigate interpretive risks.
Conclusion & Next Steps
The PLOS One protocol by Pokharel et al. (2026) provides a clear, theory-informed blueprint for conducting a timely qualitative study of primary care physicians' perceptions of LLM adoption, with planned recruitment between August 1, 2025 and January 31, 2026 and analysis using NVivo and the TAB framework.
Researchers and health services teams can reuse the protocol's sampling strata, TAB-derived deductive codes, and reflexive team practices to accelerate robust qualitative inquiry.
Evidano can operationalize the protocol's workflow by automating secure transcription, supporting deductive + inductive thematic coding, and producing segment comparisons and visualizations to inform policy briefings.
To try this workflow on your interview transcripts or open-ended survey data, Try Evidano for free.
Topics
- LLM adoption qualitative analysis
- qualitative analysis of LLMs
- primary care LLM adoption
- technology adoption behavior framework
- AI-enabled qualitative research
Keep reading
- Commentary on NewsPhysician Perceptions of LLMs: Adoption in Primary CarePractical explainer of physician perceptions of LLMs from a PLOS One protocol (7 Aug 2026). Learn methods, timelines, and how AI-enabled qualitative research accelerates insight.
- Commentary on NewsFaster Safety Reporting: AI qualitative analysisShow how AI qualitative analysis for safety reporting reduces trial burden after April 2026 UK reforms. Practical steps, quotes from PLOS One, and an Evidano demo.
- Commentary on NewsNew Insights: Qualitative Analysis of Adolescent MotherhoodAI-enabled guide to qualitative analysis of adolescent motherhood in Pakistan, based on a July 30, 2026 PLOS One study. Learn methods, findings, and how Evidano speeds synthesis.