What AI Qualitative Analysis Reveals About Cancer Care

According to the PLoS ONE study published on October 28, 2026, healthcare providers at Ocean Road Cancer Institute identified three major barriers to psychosocial cancer care: inadequate training, an unsupportive organizational environment, and difficulty building trusting patient-provider relationships. AI qualitative analysis can help healthcare researchers organize interview evidence around these themes while preserving quotations and context.

• The PLoS ONE researchers interviewed 18 healthcare providers between April 20 and May 19, 2022, including 15 nurses and three social workers.
• The PLoS ONE study found that none of the 18 participants had attended a formal course specifically focused on psychosocial care.
• The PLoS ONE researchers reported that each interview lasted between 30 and 48 minutes and continued until information saturation was reached.
• The PLoS ONE analysis identified three connected levels of constraint: provider capability and workload, organizational coordination and resources, and patient-provider communication.
• The PLoS ONE study supports training, dedicated coordination systems, clear protocols, and multidisciplinary referral pathways as practical areas for improvement.

The PLoS ONE study used an exploratory qualitative design to examine healthcare providers’ perceptions of barriers to psychosocial cancer care at Ocean Road Cancer Institute in Dar es Salaam, Tanzania. The researchers used purposive sampling, conducted face-to-face semi-structured interviews, transcribed the recordings verbatim, and analyzed the transcripts thematically.

The PLoS ONE research team recruited 18 participants who had worked in their current unit for at least six months and were directly involved in caring for people with cancer. The participant group included 15 registered nurses and three social workers, with professional experience ranging from 3 to 34 years.

The PLoS ONE researchers identified three overarching themes. First, providers described insufficient training, staff shortages, heavy workloads, and limited time for psychosocial conversations. Second, providers described fragmented coordination, unclear roles, inadequate resources, and a lack of dedicated psychosocial care structures. Third, providers described communication challenges, patient hesitancy, mistrust, and difficulty establishing therapeutic relationships.

One nurse told the PLoS ONE researchers, “I have not received training on psychosocial care; most of the training I have been receiving is on pain management.” Another participant described the workload constraint this way: “the time to talk with patients in detail is minimal.” These quotations show why qualitative analysis must retain participant language rather than reducing findings to theme labels alone.

The Tanzania study indicates that oncology service improvements should address individual skills and organizational systems together. The PLoS ONE researchers concluded that training alone may not resolve barriers when staff shortages, unclear referral pathways, competing responsibilities, and limited resources remain in place.

Healthcare research teams can use AI qualitative analysis to move from raw interviews to an auditable evidence map. A useful workflow is to code provider capability, workload, coordination, resources, trust, communication, and patient disclosure as related but distinct concepts. Researchers can then compare how nurses and social workers describe the same barrier, inspect representative quotations, and identify whether a theme reflects widespread discussion or a small number of especially detailed accounts.

The PLoS ONE study also demonstrates why context matters. The research took place at one national tertiary cancer referral hospital, and the authors stated that the findings may not represent every cancer setting in Tanzania. AI-assisted synthesis should therefore support, not replace, reflexivity, peer debriefing, methodological review, and local interpretation.

For clinical and public health researchers, the ethical boundary is important: AI-generated themes are research outputs for review, not diagnoses of patients or healthcare workers. Researchers should remove identifying information, protect recordings and transcripts, and verify every important interpretation against the source data.

The October 2026 PLoS ONE study shows how qualitative interviews can reveal connected barriers that statistics alone may not explain, from missing training and excessive workload to fragmented coordination and mistrust. AI qualitative analysis gives healthcare researchers a practical way to organize these patterns, compare participant segments, retrieve supporting quotations, and maintain a clearer audit trail. Human researchers must remain responsible for interpretation, ethics, confidentiality, and decisions about how findings should inform care. To analyze interview transcripts, surveys, and healthcare documents with an AI-assisted qualitative workflow, try Evidano for free.