A minimally invasive portable system for sleep apnea detection

Abstract

Health care is changing the focus from primary and specialty care to prevention and wellness. Therefore, home health care is seen as one of the most relevant wellness services due to high accessibility and low cost of diagnosis. The growth relevance given to the sleep related disorders, due to the high importance of sleep in our lives, is specifically significant in this context encouraging the development of methods capable of non-invasively monitor and detection. One of the most relevant sleep disorders is obstructive sleep apnea, being the focus of the work presented in this paper to develop a minimally invasive portable system to detect this disorder using only oximetry. The system developed in this work is capable of collecting the oxygen saturation and pulse rate signals and send them wirelessly to the processing station where an application records and analyses the data. A graphical user interface guides the patients to start the monitoring session and a report is produced at the end of the analysis. The information is graphically presented to the patients and a resume file is generated to be analysed by the sleep technician. A database with 35 patients recordings was analysed, using a cross validation technique in order to evaluate the performance using a logistic regression model as a classifier. The algorithm achieved an accuracy of 86.6% (sensitivity = 66.9%, specificity = 94.5%, AUC = 90.7).