Background and Aims Patient education materials (PEMs) support patient understanding and engagement in care. This study evaluated the readability, understandability and actionability of non-oncological urological PEMs produced by the British Association of Urological Surgeons (BAUS) and the European Association of Urology (EAU), and explored agreement between human and AI-assisted PEMAT-P assessment. Methods and Results A cross-sectional descriptive analysis assessed eight matched PEMs relating to urinary tract infection, urinary incontinence, benign prostatic enlargement/lower urinary tract symptoms, and nocturnal polyuria. Readability was assessed using Flesch Reading Ease (FRE), Flesch–Kincaid Grade Level (FKGL), Simple Measure of Gobbledygook (SMOG), and Gunning Fog Index. Understandability and actionability were assessed using the Patient Education Materials Assessment Tool (PEMAT-P). Mean FKGL was 11.85 ± 2.77, SMOG 13.24 ± 2.39, Gunning Fog 14.09 ± 3.22, and FRE 48.5 ± 13.32. All materials exceeded the recommended reading level. Mean PEMAT-P understandability and actionability were 91.0% and 88.3%. Human–AI item-level agreement for understandability and actionability was 83.6% and 81.8%; Cohen's κ was 0.25 and 0.23, and Gwet's AC1 was 0.79 and 0.76. Conclusion The materials exceeded recommended reading levels despite variable PEMAT-P scores. Human vs AI differences remained, requiring further evaluation of PEMs to establish the potential role of AI as an adjunct to human PEMAT-P assessment.
Background Obtaining valid informed consent in emergency orthopaedic trauma surgery is challenging due to pain, distress, time pressure, and limited opportunity for patient reflection. Evidence suggests that verbal explanations alone are often insufficient for patient comprehension of surgical procedures and associated risks. Objective To evaluate whether procedure-specific visual aids improve patient understanding during the consent process for emergency orthopaedic trauma surgery. Methods This quality improvement study using a pre- and post-intervention evaluation was conducted within the Trauma and Orthopaedic Department at Craigavon Area Hospital. Procedure-specific visual aids illustrating anatomy, surgical technique, implant positioning, and categorised risks were introduced during emergency consent discussions. The visual aids were generated using ChatGPT (OpenAI), incorporating its generative image-generation capability, and were subsequently reviewed by the authors to ensure clinical accuracy. Patient understanding, awareness of risks, confidence in decision-making, and anxiety were assessed using a locally designed structured questionnaire. Patients rated their understanding immediately after the initial verbal explanation and again following introduction of the visual aids during the same consent encounter. Statistical analysis was performed using the Wilcoxon signed-rank test. Results Thirty patients undergoing emergency orthopaedic trauma surgery were included. Baseline self-rated understanding of the planned procedure, assessed on a 1 to 10 numerical scale, ranged from 2 to 6 out of 10, with a mean score of approximately 4.1 out of 10. Following introduction of the visual aids during the same consent encounter, understanding scores ranged from 6 to 10 out of 10, with a mean score of approximately 8.2 out of 10. All patients (100%) demonstrated improved understanding of the procedure and reported improved understanding of associated risks. The mean improvement was approximately 4.1 points and was statistically significant (Wilcoxon signed-rank test, p < 0.001). Twenty-four patients (80%) reported reduced anxiety, five (16.7%) reported no change, and one (3.3%) reported increased anxiety despite improved comprehension. Conclusion Procedure-specific visual aids were associated with significant improvements in patient understanding, procedural risk awareness, and confidence during emergency orthopaedic trauma consent discussions. This simple, low-cost intervention may enhance the quality of informed consent in acute trauma settings. However, the absence of a control group, use of a locally designed questionnaire, and single-centre design limit definitive conclusions regarding causality and generalisability.
