Artificial Intelligence-Supported Nursing Program for Sleep Quality in People With an Ostomy

Individuals living with an intestinal or urinary stoma frequently report disturbed sleep.Night‐time pouch filling and leakage, odour and gas concerns, restricted sleepingpositions, peristomal skin problems and the need for night‐time pouch emptying interruptsleep, while anxiety, depressive symptoms, altered body image and social withdrawal addto this burden. Poor sleep in this population has been associated with reduced quality oflife and impaired daytime functioning; however, most of the available evidence iscross‐sectional and based exclusively on self‐reported measures. Nursing interventionsthat specifically target sleep in people with a stoma are few, generally consist of asingle component, and have rarely been tested in randomised designs or alongsideobjective sleep measurement.This trial evaluates an Artificial Intelligence‐Supported Multimodal Nursing Intervention(AI‐MNI) developed to improve sleep quality in adults living with a stoma. Theintervention was designed in line with the Medical Research Council framework for complexinterventions and is delivered over four weeks through the institutional web‐basedlearning management system. It comprises five structured modules: stoma care andnight‐time comfort; sleep hygiene and night routine; psychological wellbeing andrelaxation techniques; body image, self‐esteem and sexual health; and physical activityand daily energy management. Each module combines written material, video, visual guides,checklists and short knowledge checks. A rule‐based chatbot embedded in the platformsupports participants by prioritising content, suggesting review of relevant material andresponding to questions within a pre‐approved educational content set; it does notprovide diagnosis, treatment advice or clinical decision support. All content wasevaluated for content validity by an expert panel before implementation.The study is a two‐arm, parallel‐group randomised controlled trial conducted in thestomatherapy unit of a university hospital. Forty‐two adults (21 per group) with acolostomy, ileostomy or urostomy of at least three months' duration and a PittsburghSleep Quality Index (PSQI) global score greater than 5 will be allocated in a 1:1 ratiousing permuted block randomisation with variable block sizes, stratified by sex and stomatype. Allocation is concealed through a central, password‐protected web‐based system, andstatistical analyses are performed by a statistician blinded to group codes. Theintervention group receives AI‐MNI in addition to standard care; the control groupreceives standard care alone and is offered access to the programme after the finalassessment (delayed‐access design). Assessments are conducted at baseline (T0), at thefirst month (T1) and at the third month (T2). The primary outcome is the change in PSQIglobal score from baseline to the first month, with the third‐month assessment used toexamine whether any improvement is maintained. Secondary outcomes are objective sleepparameters (total sleep time, sleep efficiency, sleep stage indicators and sleep onsetlatency) recorded with a consumer wrist‐worn device on three consecutive nights at eachassessment point, and stoma‐specific quality of life. Module completion, session durationand level of interaction are reported as process outcomes. Analyses follow theintention‐to‐treat principle and are based primarily on linear mixed‐effects models.