A machine learning-based clinical decision support system mobile application for pain management in children with Sickle cell disease

Abstract

Introduction: Sickle Cell Disease is a genetic blood disorder characterized by recurrent, severe pain episodes that severely impact the quality of life in children. Despite medical advancements, effective home-based pain management remains a major challenge, particularly in resource-limited settings with inadequate access to specialized care. This study developed a patient-centered Clinical Decision Support System mobile application, powered by Machine Learning and Natural Language Processing, to support caregivers in managing SCD-related pain in children. Objective: To design, develop, and evaluate a Machine Learning-based CDSS mobile application that assists caregivers in making timely, informed decisions for managing SCD pain in children. Methods: A user-centered, mixed-methods approach was used in four iterative phases: (1) contextual analysis through interviews and observations; (2) definition of functional and non-functional requirements; (3) design and development of a mobile CDSS integrating SVM for pain severity classification and NLP for interpreting free-text symptom input; (4) evaluation through real-world testing with 5 caregivers and 2 healthcare professionals. Results: The CDSS achieved 100% task completion during user testing, with an average input time of 2.5 to 5 minutes. NLP accuracy reached 90%, and caregivers reported high satisfaction, especially with Luganda support and clear recommendations (e.g., half-tablets, spoons). Healthcare professionals confirmed clinical alignment with national SCD guidelines. Conclusion: Therefore, based on the results the CDSS mobile app significantly enhances caregiver confidence, supports safe home-based decisions, and bridges the gap between household care and clinical guidance. Its culturally appropriate design and ML/NLP integration make it suitable for use in low-resource settings. Recommendation: Scale-up should include multilingual expansion, integration with national health systems, and live escalation features. The CDSS also holds promise for adaptation to other children chronic conditions requiring home-based decision support.