Neural network prediction of obstructive sleep apnea from clinical criteria

Study objectives: Clinical prediction models for the diagnosis of obstructi ve sleep apnea (OSA) have lacked the accuracy necessary to confidently repl ace polysomnography (PSG), Artificial neural networks are computer programs that can be trained to predict outcomes based on experience. This study wa s conducted to test the hypothesis that a generalized regression neural net work (GRNN) could accurately classify patients with OSA from clinical data. Study design: Retrospective review. Setting: Regional sleep referral center, Patients: Randomly selected records of patients referred for possible OSA. Measurements: The neural network was trained using 23 clinical variables fr om 255 patients, and the predictive performance was evaluated using 150 oth er patients. Results: The prevalence of OSA in this series of 405 patients (293 men and 112 women) was 69%. The trained GRNN had an accuracy of 91.3% (95% confiden ce interval [CI], 86.8 to 95.8). The sensitivity was 98.9% for having OSA ( 95% CI, 96.7 to 100), and the specificity was 80% (95% CI, 70 to 90). The p ositive predictive value that the patient would have OSA was 88.1% (95% CI, 81.8 to 94.4), whereas the negative predictive value that the patient woul d not have OSA (if so classified) was 98% (95% CI, 94 to 100). Conclusions: Appropriately trained GRNN has the ability to accurately rule in OSA from clinical data, and GRNN did not misclassify patients with moder ate to severe OSA. In this study, use of the neural network could have redu ced the number of PSG studies pet-formed. Prospective validation of the neu ral network for the diagnosis of OSA is now required.