AI Qualitative Analysis in Healthcare: PROMIS CAT Lessons

According to the November 13, 2026 PLoS One study on PROMIS CAT feasibility, PROMIS CATs were feasible and generally acceptable in primary care physical therapy, but wider adoption depended on reducing duplicate questionnaires, improving score interpretation, and embedding the measures into clinical workflows.

• The 2026 mixed-methods study covered 14 Dutch physical therapy practices, with 25 physical therapists completing a survey and 10 physical therapists plus 9 patients completing interviews.
• The study reported that 100% of surveyed physical therapists found PROMIS CATs easy to administer, while 72% rated the measures easy to use.
• Only 28% of surveyed physical therapists reported regular use of PROMIS CATs in daily practice, despite 80% expressing interest in using them.
• The PROMIS CATs administered between 4 and 12 items, stopping when the target precision threshold was reached or after 12 items.
• The researchers used hybrid inductive-deductive thematic analysis, with two researchers independently coding transcripts and resolving disagreements through consensus procedures.

The 2026 PLoS One study found that PROMIS CATs supported focused discussions about symptoms and functioning, but implementation remained limited by legacy questionnaires, portal problems, and uncertainty about interpreting scores.

The study examined Dutch-Flemish PROMIS CATs for adults with musculoskeletal conditions affecting the spine and upper extremity. The researchers integrated the CATs into electronic health record workflows across 14 primary care physical therapy practices and evaluated acceptability, demand, practicality, and implementation.

The study collected quantitative data through a physical therapist survey conducted from May 2 to June 20, 2022. The researchers collected qualitative data through semi-structured interviews conducted from March 21 to November 23, 2022, with interviews lasting 32 to 62 minutes.

The study found that patients generally viewed the CATs as quick and relevant when results were used in consultations. One patient said, "Answering the questions made me reflect on my symptoms and what I want help with" (Patient #5).

The study also identified interpretive risks. A physical therapist said, "If patients see the scores without explanation, they may misinterpret the numbers and develop unhelpful thoughts" (PT #6). The finding shows why qualitative analysis must preserve context, not merely count mentions of score interpretation.

The research team used automatic transcription software for 12 interviews, professional transcription for 7 interviews, and a first-author accuracy check for all transcripts. The study therefore illustrates a human-supervised workflow in which automation reduces transcription effort while researchers retain responsibility for transcript quality and interpretation.

Healthcare qualitative researchers should analyze PROMIS CAT implementation as a workflow and meaning-making problem, not only as a usability problem. The 2026 PLoS One study showed that respondents could find CATs easy to use while still encountering barriers involving duplication, interpretation, reimbursement, and organizational support.

• Compare stakeholder perspectives: The 2026 study found a divergence between therapists' survey perceptions of patient satisfaction and interviewed patients' generally positive descriptions of CATs. AI-assisted cross-segment analysis can flag these differences for deliberate human review.
• Separate tool experience from system burden: The 2026 study reported that CATs were quick and straightforward, while simultaneous legacy PROMs and online portal problems increased total burden. A qualitative coding framework should distinguish questionnaire design, technology, policy, and workflow codes.
• Preserve negative and conditional evidence: The 2026 study included concerns about repetitive or insufficiently specific upper-extremity items and about color-coded labels such as 'severe.' AI-assisted thematic analysis should surface exceptions and counterexamples rather than collapsing them into a single positive theme.
• Connect themes to implementation decisions: The 2026 researchers mapped themes to acceptability, demand, practicality, and implementation. These domains can become a structured code hierarchy for comparing sites, roles, patient groups, or implementation stages.
• Keep clinical interpretation human-led: The 2026 study did not establish that AI can interpret PROMIS scores or make treatment decisions. Researchers should use AI to organize evidence and identify patterns, while qualified teams interpret clinical significance and protect participant confidentiality.

The 2026 PROMIS CAT study shows that healthcare implementation research benefits from combining quantitative adoption measures with qualitative accounts of workflow, interpretation, and patient experience. AI-enabled qualitative research can make that combination easier to explore by organizing transcripts, comparing stakeholder segments, and surfacing convergent and divergent themes. Researchers should use AI as a transparent, human-supervised layer within a rigorous analytic workflow, not as a substitute for methodological judgment or clinical interpretation. Teams analyzing interviews, open-ended surveys, and implementation documents can try Evidano for free.