Better Care Insights: AI Thematic Analysis in Healthcare

The SYNC-RF protocol shows that AI thematic analysis in healthcare can support mixed-methods research by organizing interview evidence about referral barriers, communication gaps, and follow-up experiences alongside survey and facility data. The protocol was published in PLOS ONE on November 3, 2026, and describes a three-phase programme for developing an integrated referral and follow-up model in India.

• According to the PLOS ONE protocol, the study will assess 20 Primary Health Centres, including 10 rural and 10 urban facilities, to examine referral capacity and follow-up systems.
• According to the PLOS ONE protocol, the patient survey will recruit 600 adults aged 30 years and above who have at least two chronic conditions, with 30 participants targeted at each selected facility.
• According to the PLOS ONE protocol, the qualitative component will include purposive interviews with 10 patients, 10 caregivers, 10 healthcare providers, and 4 to 6 state or district health officials.
• According to the PLOS ONE protocol, interviews are expected to last 30 to 45 minutes and will be conducted in English or a local language, then transcribed and translated where necessary.
• According to the PLOS ONE protocol, the research team expects expert consultation to refine the SYNC-RF model by mid-2027, with final programme reporting expected by the end of 2027.

The SYNC-RF protocol uses three linked phases to connect global evidence, district-level health system assessment, and model development. According to the PLOS ONE authors, Phase 1 is a systematic review, Phase 2 is a convergent mixed-methods assessment in Jodhpur district, Rajasthan, and Phase 3 develops and refines the System for Navigation, Coordination, Referral, and Follow-up model.

The PLOS ONE authors state, "This study will use a sequential, multistage mixed-methods research design conducted in three interlinked phases." The design combines a systematic review, facility assessments, a cross-sectional survey, and in-depth interviews before integrating findings through a joint display matrix.

The protocol defines reflexive thematic analysis as the method for analyzing interview transcripts. According to the PLOS ONE protocol, researchers will repeatedly read transcripts, apply deductive and inductive codes, group codes into sub-themes and broader themes, and discuss interpretation within the research team.

The qualitative sample is designed to capture different experiences of healthcare continuity. According to the PLOS ONE protocol, interviews will include adults with multiple long-term conditions, caregivers, healthcare providers, and district or state health officials, while the final sample will be guided by information power and data saturation.

The PLOS ONE authors explain the intended outcome directly: "The integrated model developed through this research is expected to strengthen continuity of care, improve coordination across various levels of the health system, and improve health outcomes for those living with MLTC." The statement describes an expected contribution of a protocol, not a result that has already been demonstrated.

AI thematic analysis can help healthcare researchers move from raw interviews to a structured evidence map, while researchers retain responsibility for interpretation, reflexivity, and ethical decisions. The SYNC-RF protocol is especially suitable for this workflow because it combines multiple stakeholder groups, multiple languages, and quantitative and qualitative data.

• According to the PLOS ONE protocol, referral initiation, counter-referral, communication, patient navigation, and follow-up tracking are important model domains that can become an initial coding framework.
• According to the PLOS ONE protocol, researchers will compare quantitative findings with qualitative themes through a joint display matrix, making cross-source organization a central analytic requirement.
• According to the PLOS ONE protocol, divergent findings will be retained for interpretation rather than automatically resolved in favor of one data source, so AI outputs should be treated as prompts for inquiry rather than final conclusions.
• According to the PLOS ONE protocol, local-language interviews will be translated into English while original-language transcripts are retained, making translation review and culturally sensitive coding essential.
• According to the PLOS ONE protocol, the study has ethical approval from the Centre for Chronic Disease Control and the University of Leicester, which reinforces the need for privacy-aware handling of patient and provider narratives.

AI does not independently establish clinical effectiveness, causal relationships, or generalizability. According to the PLOS ONE protocol, the study is limited by its single-district setting, cross-sectional survey design, possible underrepresentation of people who do not use primary care, and potential loss of cultural nuance during translation.

For healthcare researchers, the strongest use of AI is therefore operational and analytical: accelerate transcript preparation, identify candidate patterns, compare stakeholder segments, and make supporting excerpts easier to review. Human researchers should validate codes, examine contradictory cases, protect participant confidentiality, and connect themes to the study framework.

The SYNC-RF protocol demonstrates how healthcare researchers can combine systematic review evidence, facility assessment, patient surveys, and stakeholder interviews to study continuity of care for multiple long-term conditions. AI thematic analysis can make this design more manageable by organizing transcripts, comparing stakeholder segments, supporting multilingual workflows, and surfacing evidence for mixed-methods integration. The PLOS ONE protocol also makes the boundary clear: AI can accelerate analysis, but researchers must validate themes, preserve cultural meaning, protect participant data, and interpret divergent evidence. Teams that want a structured environment for healthcare interviews, surveys, documents, and cross-segment analysis can try Evidano for free.