Diabetes Management Errors in Australia: A Factorial Randomised Controlled Trial of a Health Workforce Educational Platform (The WDEP.AI RCT) for Rehabilitation Wards

A 6-month Pilot 2x1 Factorial cluster RCT will be conducted in six rehabilitation wards within Sydney metropolitan area. Health professionals in all six wards will be invited to participate in the study. Each Phase will span 3 months. Phase 1: Baseline Data Collection (3 months) 1. All eligible staff members from the first two randomly selected wards will be invited to join the trial. 2. Baseline data on diabetes management knowledge of consenting staff will be assessed using a diabetes management knowledge quiz (Pre-Intervention DMKQ for wards 1 and 2). Phase 2: Intervention 1 – Standard version of the WDEP.AI platform (3 months) 1. Intervention 1 (I1): the first two wards will be randomised in a 1:1 ratio to either receive access to the standard version of the WDEP.AI platform (WDEP.AI-Standard) to complete the online topic on Diabetes and Aged Care over a 10-week period, or, to serve as the Control group, which will continue with standard care without additional education. 2. Following topic completion, staff will be invited to provide feedback through an online topic evaluation (TE) form. 3. 3-month post-intervention DMKQ (I1) will be performed to measure learning outcomes from Intervention 1 (WDEP.AI-Standard) in wards 1 and 2. Phase 3: Post Intervention 1 Analysis, Machine Learning Enhancement and Baseline data collection for Intervention 2 (3 months) 3.1 Post Intervention 1 analysis: 1. After staff data collection is complete, staff from the control ward will have access to WDEP.AI-Standard. 2. Data from the initial intervention with WDEP.AI-Standard will be analysed using AI and Machine Learning (ML) algorithms to enhance the platform's content. Learning outcomes will be measured by comparing baseline DMKQ results pre and post Intervention 1, offering insights into the platform's effectiveness and guiding further improvements. 3. 6-month post-intervention DMKQ (I1) will be performed to measure the retainment of knowledge from Intervention 1 (WDEP.AI-Standard) in wards 1 and 2. 4. Focus group/ one-on-one interviews will be conducted for participants randomised to WDEP.AI Standard during Intervention 1. 3.2 Machine Learning Enhancement and Baseline Data Collection for Intervention 2: 1. All eligible staff members from the remaining four wards (3-6) will be invited to join the trial. 2. Baseline data on diabetes management knowledge of consenting staff will be assessed using a diabetes management knowledge quiz (Pre-Intervention 2 DMKQ for wards 3, 4, 5 and 6). Phase 4: Intervention 2 – Enhanced Version of the WDEP.AI Platform (3 months) 4.1 Intervention 2 (I2): 1. Intervention 2: The remaining four wards will be randomly assigned 1:1 to either receive access to the enhanced version of the WDEP.AI platform (WDEP.AI-Enhanced) to complete the online topic on Diabetes and Aged Care over a 10-week period, or, to serve as the Control group, which will continue with standard care without additional education. 2. Following topic completion, staff will be invited to provide feedback through an online topic evaluation (TE). 3. 3-month post-intervention DMKQ (I2) will be performed to measure learning outcomes from Intervention 2 (WDEP.AI-Enhanced) in wards 3, 4, 5 and 6. Phase 5: Post-intervention 2 Analysis (3 months): 5.1 6-month post-intervention DMKQ (I2): 1. 6-month post-intervention DMKQ (I2) will be performed to measure the retainment of knowledge from Intervention 2 (WDEP.AI-Enhanced) in wards 3, 4, 5 and 6. 5.2 Post-intervention 2 Analysis: 1. Access to WDEP.AI-Enhanced platform to control and randomised groups in wards 1 and 2, and to control group in wards 3, 4, 5 and 6 will be provided, ensuring all participants benefit from the enhanced educational resources. 2. Qualitative data will be gathered through focus groups/one-on-one interviews for wards 3, 4, 5 and 6, exploring the usability, acceptability, and feasibility of the WDEP.AI-Enhanced platform from the perspective of health professionals. 3. Data collected pre and post WDEP.AI-Enhanced will be analysed. 4. Analysis of TE metrics 5. A health economic analyses will be conducted to evaluate the cost-effectiveness of the intervention, providing valuable information on its potential for wider implementation. **Ward inpatient data collection - inpatient data will be collected retrospectively (before during and after the intervention) for the duration of the study for all wards** Phase 6: Wrap-up Phase (3 months): 6.1 Comprehensive analysis: 1. Comprehensive analysis of all trial metrics (Quantitative). 2. Analysis of Focus Group Transcripts (Qualitative) 3. Preparation of the final report. Building on the work completed on the foundation WDEP platform and utilising a specific focus in the aged care population, WDEPAI will focus on improving care, education and reducing the impact of diabetes related complications. The four themes we have focused on revolve around factors affecting aged care, reducing diabetes complications, supporting staff through national and local policies and early recognition of deterioration. Through the use of artificial intelligence (AI) and machine learning (ML) the platform will evaluate learning styles, learner engagement and how by improving diabetes clinical knowledge will lead to improved patient outcomes, reduced medication errors and harm minimisation. Resources have been developed utilising available best practice guidelines, patient information and diabetes advocacy organisations such as the National Diabetes Service Scheme. Diabetes Australia and the Aged Care Quality and Safety Commission frameworks. WDEP.AI is structured so that the learners can work at their own pace in small sessions bite-sized chunks at a time. The advantage is that they can log into the platform and devote whatever time they have available, while saving their progress and continuing the package promptly. Through the use of introductory videos, web links appropriate to the topic and easily accessible resources, learners can build their knowledge and confidence in delivering evidence-based based safe diabetes care across all populations of those living with diabetes. Learners have the opportunity to reattempt incorrect answers, with resources relevant to the incorrect answers provided between each attempt. This approach allows them to improve their knowledge and likelihood of success in each competency. The competencies support policies and patient-centred strategies, such as language matters, medication safety, nutrition, timing of meals, supporting a holistic approach thereby improving quality of life. Adherence will be assessed using analytics on time spent, number of attempts, and abandonment rates for onboarding, each theme, and each competency. Completion metrics will include % logging in, % completing onboarding, % completing all content, % completing each theme, and % completing each competency. Qualitative feedback will cover appearance, navigation, and overall user experience Artificial Intelligence (AI) personalises learning by tailoring educational content to individual learners' styles, paces, and needs, enabling more effective engagement and outcomes. It analyses large datasets in real-time to track learner interactions, predict areas needing extra support, and dynamically adjust content difficulty through adaptive learning. AI also provides personalized feedback, recommends relevant resources, and identifies at-risk learners, allowing for timely interventions to enhance motivation and success. References: -Onesi-Ozigagun, O., Ololade, Y. J., Eyo-Udo, N. L., & Ogundipe, D. O. (2024). Revolutionizing education through AI: a comprehensive review of enhancing learning experiences. International Journal of Applied Research in Social Sciences, 6(4), 589-607. -Wang, S., Christensen, C., Cui, W., Tong, R., Yarnall, L., Shear, L., & Feng, M. (2023). When adaptive learning is effective learning: comparison of an adaptive learning system to teacher-led instruction. Interactive Learning Environments, 31(2), 793-803. -Kump, B. (2010). Evaluating the domain model of adaptive work-integrated learning systems. Sharma, N., Doherty, I., & Dong, C. (2017). Adaptive learning in medical education: the final piece of technology enhanced learning?. The Ulster medical journal, 86(3), 198. -Wade, S. W., Moscova, M., Tedla, N., Moses, D. A., Young, N., Kyaw, M., & Velan, G. M. (2020). Adaptive tutorials versus web-based resources in radiology: a mixed methods analysis in junior doctors of efficacy and engagement. BMC Medical Education, 20, 1-12. Participants who consent to participate in the focus groups will be invited to take part. Each focus group will include approximately 20 participants, made up of endocrinology consultants, nurses, allied health professionals, researchers, patients or their carers, pharmacists, and general practitioners. We are seeking feedback on the ease of use of the website and program, any issues encountered, the impact of WDEP-AI on participants’ diabetes knowledge and practice, the relevance of the questions and resources to their work, desired content areas to strengthen their knowledge, facilitators and barriers to accessing or completing WDEP-AI, reasons for potentially giving up on the module, and participants’ likes, dislikes, and suggestions for improvement. Retrospective Data Collection for the Study: This study involves the use of retrospective inpatient data to evaluate patient outcomes before the implementation of intervention strategies. The data will be sourced from the electronic medical records of patients admitted to the relevant hospital wards participating in the study. Pre-Intervention Retrospective Data: To establish baseline characteristics and outcomes, retrospective data will be collected for the three months preceding the start of each intervention stage. For Intervention Stage 1 (expected to commence in May 2025), retrospective data will cover the period from February 2025 to April 2025. For Intervention Stage 2 (expected to commence in November 2025), retrospective data will cover the period from August 2025 to October 2025. These retrospective data will assist in comparing pre- and post-intervention outcomes. All data will be extracted from patient medical records maintained by hospital clinical information systems. Note: While additional data will be collected during and after each intervention stage for evaluation purposes, only the time periods mentioned above constitute the retrospective component of data collection.