Development of an IoT-Based Traditional Bone-Setting Appointment Reminder Using Machine Learning Technology

Abstract

Traditional bone-setting (TBS) is still a popular method of managing musculoskeletal injuries in Nigeria because it is culturally accepted, affordable and readily available, as compared to orthodox orthopaedic treatment. Complications, however, often occur because of inadequate follow-up of patients and the lack of an organized schedule and reminder system, leading to malunion, non-union and limb-threatening infections. This paper introduces the appointment reminder system with a Machine Learning prediction engine for the support of TBS practice and referral to orthodox care, which is developed by the concepts of IoT. The system consists of an IoT sensor node, mounted on a cast and worn by the patient, a GSM based SMS module, a cloud-based database, a mobile/USSD interface, and a Random Forest (RF) Classifier which classifies the chances of missing an appointment using attendance history, distance, socio-demographic, and injury-severity data. Multi-channel reminders (SMS, voice call, push notification, wearable buzzer/LED alert) are automatically sent to high risk patients and one reminder is sent to lower risk patients. The design and implementation was done on the Rapid Application Development (RAD) methodology. Results demonstrated a 97.6% reminder delivery success rate, 88.3% no-show prediction accuracy, and a 31.7% reduction in no-shows relative to baseline as simulated. In conclusion, the study suggests that the combination of low cost IoT hardware with machine learning provides a viable, culturally appropriate approach to enhance compliance in TBS follow-up, and future directions include large-scale field deployment and digital health integration at national scale.