Surgery · Medical Literature
Exploring patient and NHS staff acceptability of artificial intelligence-supported surgical wound monitoring within the WISDOM feasibility study: a multi-centre qualitative interview study
Abstract
Objective Artificial intelligence (AI) is beginning to be used within digital surgical wound monitoring to facilitate implementation at scale. AI acceptability is essential to its success. This study aimed to explore patient and staff acceptability of an AI-based digital surgical wound monitoring platform. Design Qualitative interviews Setting Two hospitals performing cardiac surgery in England. Participants 20 patients undergoing cardiac surgery and 10 clinical staff participating in a randomised feasibility trial of surgical wound monitoring with AI. Interventions Semi-structured interviews were conducted focusing on participants’ experiences or perceptions of AI-based digital surgical wound monitoring. Data were analysed, guided by the theoretical framework of acceptability. Primary measure Patient and staff acceptability of AI within surgical wound monitoring. Results Patients and staff were supportive of the use of AI within surgical wound monitoring and felt safety was improved, access was increased and efficiency was improved. AI monitoring was perceived to allow the early detection and treatment of surgical wound complications and facilitate the implementation of monitoring at scale for all patients. Some participants were concerned about data security risks, staff over-reliance on AI and the loss of human interaction. Conclusion AI-supported digital surgical wound monitoring appeared acceptable to participants within this feasibility study. Further research is needed to evaluate implementation in broader settings. Trial registration number ISRCTN16900119 and NCT06475703.
Clinical relevance
Surgical site infections and wound complications following cardiac surgery represent a significant source of morbidity, mortality, and NHS resource utilization, yet scalable remote monitoring solutions remain underimplemented. This multi-centre qualitative feasibility study addresses a critical translational gap by evaluating the human acceptability of AI-augmented digital wound monitoring platforms among both patients and clinical staff — a prerequisite for real-world deployment. As health systems increasingly pilot AI-driven remote care pathways, understanding stakeholder acceptability is essential before large-scale randomized efficacy trials can be meaningfully designed.
Key findings
1. Both patient participants (n=20 undergoing cardiac surgery) and clinical staff (n=10) expressed broad support for AI-based digital surgical wound monitoring within the WISDOM feasibility trial (ISRCTN16900119 / NCT06475703). 2. Key perceived benefits converged across three domains: improved patient safety, increased access to wound monitoring, and enhanced operational efficiency — though no quantitative effect sizes were reported given the qualitative study design. 3. AI monitoring was perceived by participants to enable earlier detection and treatment of surgical wound complications compared to standard care pathways. 4. Stakeholders identified AI as a facilitator of monitoring at scale, potentially expanding surveillance to all post-cardiac-surgery patients regardless of geographic or logistical barriers. 5. Notable concerns raised included data security risks, the potential for clinician over-reliance on AI-generated assessments, and loss of valued human interaction in wound care. 6. The study was limited to two cardiac surgery centres in England, and findings are based on qualitative interview data without quantitative outcome measures such as sensitivity, specificity, NNT, or complication rates; generalizability beyond this feasibility context requires further investigation.
Practice implications
Surgeons and perioperative teams considering adoption of AI-based wound monitoring platforms should proactively address patient concerns around data governance and ensure transparent communication about how AI outputs are integrated into — rather than substituted for — clinical decision-making, in order to preserve therapeutic rapport. Staff training protocols must explicitly target the risk of automation bias, emphasizing that AI triage alerts should augment rather than replace clinical assessment, particularly for higher-risk post-cardiac-surgery wounds. Prior to broader implementation, multidisciplinary teams should advocate for prospective trials in diverse surgical populations and settings to establish clinical efficacy endpoints (e.g., time-to-diagnosis of surgical site infection, readmission rates) alongside the acceptability framework demonstrated here.