Chai-Coetzer et al designed and validated a predictive model for diagnosing obstructive sleep apnea without polysomnography in 157 general primary care patients considered at average risk for obstructive sleep apnea. Berlin/Epworth questionnaires, snoring, waist size, witnessed apneas and age were fit into an algorithm, followed by home oximetry in a sample of mostly high-Berlin-scoring patients. All patients also underwent home PSG as the internal gold/reference standard. Sensitivity was 88% and specificity 82% in the validation cohort. Thorax 2011;66:213-219.