A certified qualitative AI tool was feasible and time-saving for analyzing structured interview data, though reliability decreased for topics requiring higher abstraction.
Certified qualitative AI tools are feasible and time-saving for analysing structured interview data in nursing interventions but require human-in-the-loop verification for complex synthesis.
Abstract Background Artificial intelligence (AI) is increasingly used in nursing to support clinical decisions, automate documentation, improve monitoring, and optimise workflow. Few studies evaluate nursing interventions with AI, and none have applied AI to evaluate programmes aimed at improving self-care in rural heart failure (HF) patients. This study assessed the feasibility of using a certified qualitative AI tool for such evaluation. Methods We conducted a secondary analysis of a prior qualitative study of a telephone problem-solving support intervention with 29 rural HF dyads (patient and care partner) designed to promote HF self-care. Two researchers independently analysed interview transcripts of those who completed the programme (N 50), to identify perceptions of acceptability, usefulness, agreement levels, and participant suggestions to inform intervention refinement. Secondary analysis was performed using a certified AI-powered web application that supports qualitative analysis with automated summarisation, structured outputs, cross-case comparison grids, and query-driven review, within a human-in-the-loop workflow requiring researcher validation. Results Automated topic identification mapped ten core areas corresponding to participant experiences, perceived benefits, implementation issues, and contextual factors related to self-care support. Two researchers verified all but one topic; the programme structure and suggestions for improvement topic were poorly summarised by AI alone. A refined prompt generated detailed insights showing participants found personalised telephone support beneficial for engagement, improved self-care, and communication. Suggested structural improvements included shorter, more frequent sessions, topic-based content, tailored materials (for example "cheat sheets" for survey questions), and flexible scheduling to reduce burden and enhance engagement. Conclusion Certified qualitative AI tools appear time-saving and reliable for analysing structured interview data and detecting key themes in nursing interventions. However, reliability decreases when higher abstraction is needed or when interview content lacks specific descriptive detail. These tools should be used with researcher verification for complex qualitative synthesis.
Durante et al. (Wed,) conducted a other in Heart failure (n=50). Certified qualitative AI tool vs. Human researchers was evaluated on Feasibility of using a certified qualitative AI tool for evaluation. A certified qualitative AI tool was feasible and time-saving for analyzing structured interview data, though reliability decreased for topics requiring higher abstraction.