CARE-ED: A PAEDIATRIC–CAREGIVER AGE-ADAPTIVE EMOTIONAL DESIGN AND BIOPHYSIOLOGICAL MONITORING FRAMEWORK FOR HOME MEDICAL AND LABORATORY DIAGNOSTIC DEVICES

Abstract

Pediatric home medical devices are essential in the management of chronic conditions when patients are outside clinical settings. Nonetheless, such devices often lack the emotional and usability requirements of children patients, as well as users. Current models do not offer age-adjusted or dual-user emotional design, which may result in a reduced satisfaction level, trust, and compliance. This study presents CARE-ED (Child and Relative Emotional Design with Age-Adaptivity) as a new age-adaptive, pediatric-caregiver emotional design framework to improve emotional interactions of and design usability of home medical devices. CARE-ED will both meet the unique emotional needs of children and their caregivers in a multi-tiered, adaptive program. A Multimodal database of physiological measures (heart rate, skin temperature, galvanic skin response), behavior (facial emotive likelihood, pattern of interactions), and reported emotional perceptions was collected from participants of 5 to 12 years of age interacting with home medical equipment of pediatric patients and their caregivers. Baseline architectures took the form of Logistic Regression, Support Vector Machine, k-Nearest Neighbors, Naive Bayes, and Decision Tree, whereas state-of-the-art efforts consisted of Random Forest, Gradient Boosting Machine, and Convolutional Neural Networks or LSTM, and Transformer-based models. The acquired biophysiological signals function as clinically relevant bio signal proxies, enabling standardized remote monitoring and tele-diagnostic interpretation in paediatric home-care settings. This data-driven approach supports diagnostic reliability, interoperability, and consistency with clinical validation workflows used in paediatric healthcare. CARE-ED framework showed better results than baseline models by attaining an ROC-AUC of 0.97, an accuracy of 94.8%, and an F1-score of 94.7% in classifying emotional status. There was a high emotional synchrony between children and the caregivers following the implementation of the CARE-ED (p < 0.05). Such gains were associated with greater device trust, comfort, and compliance in simulated at-home use conditions as examples of the real-world usefulness of age-adaptive emotional design. We show the validation of CARE-ED as the first dual-user, emotional design interface validated by multimodal machine learning training of pediatric home medical devices. Based on the results of the current study, it has been ascertained that the framework could alter the usability of pediatric medical devices with the use of emotional intelligence catered to the requirements of a child and the caregiver assigned to work on them. Futurework will explore how to integrate it into systems of telemedicine and extend the framework to other domains of chronic care devices, in order to achieve the capability to enhance longitudinal emotional support as well as adherence in the reality of healthcare contexts.