Validating the Diagnostic Accuracy of the Sleep Apnea Clinical Score for Use in Primary Care Populations

Abstract

ObjectiveTo provide external validation of the diagnostic accuracy of the Sleep Apnea Clinical Score (SACS) tool in a new setting and patient population. Patients and MethodsWe conducted a prospective cohort study. Potential participants were adult family medicine patients. We excluded patients with a SACS of 0, known obstructive sleep apnea (OSA), negative results of previous testing, or life-limiting conditions. After SACS determination, participants completed overnight oximetry, sleep medicine consultation, and polysomnography. Those interpreting tests were blind to the participant's SACS. We determined likelihood ratios (LRs) for OSA diagnosis and posttest probabilities (PTPs). We calculated OSA prevalence (pretest probability), sensitivity, specificity, and positive and negative predictive values. ResultsOne hundred ninety-one of 312 participants (61%) completed all steps. The prevalence of OSA was similar to that found in the derivation cohort (40% vs 45%; P=.31). With OSA defined as Apnea Hypopnea Index greater than 10, a SACS greater than 15 was 40% sensitive and 90% specific, with a positive predictive value of 73% and a negative predictive value of 69%. A SACS greater than 15 in our cohort produced an LR of 4.03 (95% CI, 3.12-5.22) with 73% PTP for OSA as compared with an LR of 5.17 (95% CI, 2.54-10.51) with 78% PTP found in the derivation cohort. ConclusionThe present study provides external validation of the SACS tool. It reliably predicted OSA for patients in our family medicine practice. Broader implementation in primary care practice is recommended. Further study will examine SACS uptake by clinicians and the resulting impact on utilization and clinical efficiency in primary care practices.