AI Qualitative Analysis: Psychological Safety in Medications

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents. The primary keyword for this post is "AI qualitative analysis psychological safety" and this article explains how AI-enabled methods can accelerate synthesis of interview data about psychological safety in medication conversations for researchers and healthcare teams. According to the PLOS ONE study published on August 20, 2026, 21 adults were interviewed or participated in focus groups between May and August 2024 about their experiences discussing medications with healthcare providers. This post translates the PLOS ONE findings into reproducible qualitative research actions using AI tools and links each recommendation to specific study methods and statistics so teams can implement a secure, research-grade Evidano workflow.

Key Takeaways

According to the PLOS ONE study, patients report that respectful communication, clear medication information, and invitation to participate in decisions create psychological safety during medication conversations (PLOS ONE). The PLOS ONE authors summarize this as a need for environments where patients feel "heard, respected, and able to contribute" (El-Kotob et al., 2026).

21 participants were interviewed between May 1 and August 31, 2024, in Ontario, Canada, according to the PLOS ONE study (El-Kotob et al., 2026).
The PLOS ONE study published on August 20, 2026, reports that 19 of 21 participants were aged 40 or older and 14 identified as White, which limits demographic transferability.
The PLOS ONE authors use the Patient Psychological Safety (PPS) framework with three constructs: Patient Belonging, Patient Learning, and Patient Participating, and they report examples such as the participant quote, "I want to be a part of the decisions that are made about my care" [P03].
A concise analytic workflow (structured coding by PPS constructs, Excel data display matrices, and iterative reviewer consensus) maps directly to AI-enabled thematic synthesis for faster, reproducible outputs.

What Happened and how the study was done

The PLOS ONE study examined patient experiences of psychological safety during medication-related clinical encounters in primary and community care (El-Kotob et al., 2026).

According to the PLOS ONE methods section, researchers recruited adults in Ontario who had taken at least one prescribed medication for three months or longer and collected qualitative data via semi-structured interviews and focus groups conducted virtually between May and August 2024.

The PLOS ONE authors coded transcripts using a secondary deductive analysis guided by the Patient Psychological Safety (PPS) framework, which includes the domains Patient Belonging, Patient Learning, and Patient Participating (El-Kotob et al., 2026).

The PLOS ONE research team processed 21 participant transcripts, used a data display matrix in Microsoft Excel, and resolved coding differences through weekly analyst meetings until consensus was achieved.

Findings Snapshot

Date	Metric	Value	Implication
May–Aug 2024	Data collection window	Interviews and focus groups conducted	Use comparable time-bound collection to justify saturation claims
Aug 20, 2026	Publication date	PLOS ONE article published (El-Kotob et al., 2026)	Cite the study when discussing psychological safety evidence
21	Sample size	n = 21 participants	Small qualitative sample, appropriate for in-depth PPS analysis
19	Age distribution	n = 19 participants ≥ 40 years old	Consider additional recruitment strategies for younger cohorts
14	Race/ethnicity	n = 14 participants identified as White	Findings may not generalize to racially diverse populations

Implications for qualitative researchers and health services teams

Answer: The PLOS ONE study implies qualitative teams should code for belonging, learning, and participation to capture psychological safety dynamics. The PLOS ONE authors recommend structuring analysis using the PPS framework to identify how communication and inclusion affect medication decision-making (El-Kotob et al., 2026).

Design: According to PLOS ONE, recruit purposively to cover intersectional identities because the 2026 study sample skewed older and White (n = 19 and n = 14 respectively).
Interview guides: The PLOS ONE study shows open prompts about how patients felt "heard" or "dismissed" produce rich data; include direct probes on stigma and perceived judgment.
Analysis: The PLOS ONE team used a deductive matrix in Excel; researchers can replicate that structure in AI platforms to map codes to PPS domains for faster cross-case synthesis.
Equity: The PLOS ONE discussion urges future work to examine social determinants such as race, gender, and disability when studying psychological safety; incorporate purposive sampling and cross-segment analysis.

How Evidano Helps

Problem: Slow synthesis of interview transcripts → Solution: Rapid thematic mapping

Evidano accelerates coding by ingesting transcripts and applying thematic and deductive codebooks aligned to frameworks like the PPS framework.

According to the PLOS ONE methods, the research team used a deductive matrix; Evidano automates initial code suggestions, produces code frequency and co-occurrence tables, and exports a reproducible matrix that matches that approach.

Use the Evidano features page to see automated coding, code hierarchy visualizations, and export options for matrices identical to the Excel display the PLOS ONE authors describe.

Problem: Manual transcription and PII risk → Solution: Secure speech-to-text with redaction

Evidano provides transcription with custom dictionaries and PII redaction to match the PLOS ONE study practice of audio-recording and transcribing interviews securely.

Teams replicating the PLOS ONE methods can reduce transcription time and protect participant confidentiality through automated redaction; see the Evidano speech-to-text feature for details.

Problem: Comparing subgroups and intersectional identities → Solution: Cross-segment analysis

Evidano supports cross-segment and frequency analyses so researchers can test the PLOS ONE finding that psychological safety varied with social identities (El-Kotob et al., 2026).

Evidano produces tables and visualizations that let teams compare codes by age, gender, race, or other demographic buckets to surface where belonging or exclusion concentrates.

FAQ: AI qualitative analysis psychological safety

How can AI qualitative analysis improve psychological safety research?

Answer: AI qualitative analysis speeds coding and surface pattern detection while preserving the interpretive steps researchers need for rigor. The PLOS ONE study used a deductive matrix approach, and AI can reproduce that matrix at scale to identify which excerpts map to Patient Belonging, Patient Learning, and Patient Participating (El-Kotob et al., 2026).

AI tools reduce time spent on initial coding and generate frequency and co-occurrence outputs that researchers then review, which mirrors the PLOS ONE consensus process but shortens iteration cycles.

Is AI reliable for sensitive interview material about stigma or shame?

Answer: AI is reliable as an assistive tool when combined with human review and privacy safeguards. The PLOS ONE authors report sensitive themes such as stigma and dismissal, and best practice is to maintain human-led interpretation of AI-generated codes for nuance and context (El-Kotob et al., 2026).

Researchers should apply PII redaction, trained codebooks, and analyst consensus after AI suggestion to preserve ethical and interpretive rigor.

What sample size did the PLOS ONE study use and does AI change sample needs?

Answer: The PLOS ONE study analyzed 21 participants collected between May and August 2024, and AI does not remove the need for purposeful sampling. The PLOS ONE authors highlight that n = 21 provided depth for PPS constructs but called for broader sampling for transferability (El-Kotob et al., 2026).

AI can make larger-sample qualitative analysis feasible, but study design should still prioritize purposive recruitment to capture intersectional experiences.

What ethical considerations apply when using AI on patient interviews?

Answer: Use encryption, controlled access, and explicit consent that covers AI-assisted processing. The PLOS ONE team followed University of Toronto ethics approvals and restricted sharing of identifiable raw data, underscoring the need for institutional review and consent for AI workflows (El-Kotob et al., 2026).

Researchers should also document AI steps in methods sections and preserve raw transcripts under secure governance to match transparent reporting standards.

Conclusion & Next Steps

The PLOS ONE study shows that respectful communication, clear medication information, and invitation to participate are core drivers of psychological safety in medication conversations (El-Kotob et al., 2026). AI-enabled qualitative analysis makes it possible to replicate the study's deductive PPS coding at scale, compare subgroups, and speed the path from transcripts to actionable recommendations.

Researchers and health services teams implementing similar studies should combine secure transcription, a PPS-aligned codebook, and AI-assisted thematic synthesis to maintain rigor while accelerating insights.

For teams ready to pilot an AI-enabled qualitative workflow that follows the methods used in the PLOS ONE study, explore Evidano features and start a secure trial. Try Evidano for free

Topics

AI qualitative analysis psychological safety
qualitative analysis of patient psychological safety
AI thematic analysis healthcare
AI-enabled qualitative research

Keep reading

Browse all articles