An economic model for screening for obstructive sleep apnea in general and occupational populations

Purpose: Using data from previous studies, we created an economic model to preliminarily assess the relative costs of several potential screening modalities for OSA if used in a large-scale preventative screening program for general medical and occupational groups. Methods: Five potential screening modalities for OSA were included in the model, 1) a questionnaire (Q) alone, 2) nocturnal pulse orimetry (nPO) alone, 3) a serial combination of a Q and nPO, 4) the gold standard study, polysomnography (PSG), and 5) current medical practice (CP), i.e., no additional screening. The testing characteristics used for modeling of a Q and nPO came from previously published data looking at sleep center populations. The ability of CP for detecting OSA was estimated based upon data from the Wisconsin Sleep Cohort. The sensitivity and specificity of PSG were both modeled to be 100%. All screen-positive subjects were modeled to require confirmatory PSG, considered diagnostic for OSA with a respiratory disturbance index ≥20 events/hr. Costs used in the base analysis included $1000, $50, and $15 for PSG, nPO, and Q, respectively. The base analysis population prevalence of OSA was set at 5%. All data and modeled assumptions were subjected to sensitivity analysis. Outcomes were computed in terms of cost per OSA diagnosis and cost per subject screened. Results: The base analysis of the model yielded the following economic costs: Screening % Cases Cost per Cost per Modality Detected OSA Dx Screenee PSG all 100 $20,000 $1,000 Q only 86 $10,444 $444 nPO only 86 $3,604 $153 Q±nPO 84 $3,429 $145 CP 10 $1,950 $10 Sensitivity analyses done for data input and modeled assumptions show that the results were sensitive to the prevalence of OSA in the population and the cost of PSG. Results were relatively insensitive to the cost of the questionnaire and the orimetry. Conclusions: A screening modality utilizing a serial combination of a questionnaire and orimetry, or orimetry alone, appears to be a cost-effective method of population screening for OSA compared to PSG and questionnaires alone, which are more costly, and current practice, which is much less effective. Clinical Implications: These more cost-effective modalities need further validation in target populations prior to their wide-spread clinical application.