AI Qualitative Analysis in Physical Therapy

The November 13, 2026 PLOS ONE study found that Dutch-Flemish PROMIS computerized adaptive tests were feasible and generally acceptable in primary care physical therapy, while interviews exposed workflow, interpretation, and questionnaire-burden barriers.

AI qualitative analysis is relevant to this type of mixed-methods research because researchers must synthesize interviews, survey responses, implementation themes, and differences between patient and clinician accounts.

• In 2022, researchers collected survey responses from 25 physical therapists and interviewed 10 physical therapists and 9 patients across 14 Dutch primary care physical therapy practices.
• The 2026 study found that 80% of surveyed physical therapists were interested in using PROMIS computerized adaptive tests, but only 28% reported regular daily use.
• The 2026 study found that 100% of surveyed physical therapists considered PROMIS computerized adaptive tests easy to administer, while only 40% believed the tests took little time.
• The qualitative interviews identified a gap between clinician perceptions and patient accounts: surveyed therapists reported limited patient satisfaction, while interviewed patients generally described the tests as quick, relevant, and simple.
• AI-enabled thematic and cross-segment analysis can help researchers compare these divergent perspectives while keeping the original interview evidence available for review.

The 2026 PLOS ONE study concluded that PROMIS computerized adaptive tests were feasible in primary care physical therapy, but broader implementation required better training, technical integration, reimbursement, and replacement of legacy questionnaires.

The researchers used a concurrent mixed-methods design. According to the study, quantitative data came from a survey of physical therapists, while qualitative data came from semi-structured interviews with physical therapists and patients. Two researchers independently performed open coding, compared codes in consensus meetings, and mapped inductively developed themes onto Bowen’s feasibility domains of acceptability, demand, practicality, and implementation.

The study reported that 76% of surveyed physical therapists believed PROMIS computerized adaptive tests contributed to treatment evaluation, and 72% believed they contributed to quality of care in 2022. However, only 48% reported always or mostly discussing CAT results with patients.

The interviews added context that percentages alone could not provide. One patient said, "Answering the questions made me reflect on my symptoms and what I want help with" (Patient #5). One physical therapist said, "I can’t always link the questionnaire scores directly to my treatment choices" (PT #3).

The study also identified a practical implementation tension. Physical therapists valued adaptive questions because one CAT could potentially replace several longer questionnaires, but participants experienced additional burden when PROMIS CATs were administered alongside legacy PROMs.

Healthcare researchers should analyze patient and clinician evidence separately before combining the findings, because the 2026 PROMIS study showed that the two groups did not always describe the experience in the same way.

The study’s divergence around patient satisfaction illustrates why qualitative analysis should not treat a survey percentage as a complete account of user experience. Only 24% of physical therapists surveyed in 2022 believed patients were satisfied with PROMIS CATs, while interviewed patients generally described the CATs as quick, simple, and relevant.

UX and implementation teams should distinguish product barriers from workflow barriers. The 2026 study reported that physical therapists found the CATs easy to access and administer, but interviews identified portal logins, forgotten passwords, incomplete online forms, uncertainty about T-scores, and the simultaneous use of legacy PROMs as obstacles.

Researchers should preserve contradictory evidence rather than forcing premature consensus. A therapist’s concern that "Using words like ‘severe’ might unintentionally worry patients or make the problem feel bigger" (PT #7) represents a different implementation question from a patient’s statement that the CAT "zoomed in on what mattered to me, instead of asking irrelevant things" (Patient #4).

AI qualitative analysis can support this work by clustering related statements, comparing segments such as patients and clinicians, identifying co-occurring barriers, and surfacing representative quotations. Researchers must still review source excerpts, define the codebook, and make the final interpretive decisions.

The 2026 PROMIS physical therapy study shows that mixed-methods healthcare research depends on more than survey percentages: interviews explain why a tool can be acceptable, easy to administer, and still difficult to implement routinely.

AI qualitative analysis can help physical therapy researchers organize transcripts, compare patient and clinician segments, identify implementation themes, and connect qualitative findings with survey evidence. Human researchers remain responsible for methodological decisions, ethical safeguards, and interpretation.

Teams analyzing patient interviews, PROM feedback, or healthcare implementation data can Try Evidano for free to accelerate evidence organization while keeping qualitative review central to the research process.