Objective: To determine the clinical utility of a limited sleep-recording device used unsupervised in the patient's home, compared with in-laboratory, fully supervised polysomnography for the diagnosis of sleep apnea/hypopnea syndrome (SAHS), and to assess its impact on costs.
Design: Prospective case study.
Setting: The sleep-disorders unit of a tertiary referral university hospital.
Patients: Fifty-five patients suspected of having SAHS and living within 30 km of our laboratory.
Methods: Patients were studied first in their homes with the limited sleep-recording device. Polysomnography was performed within 30 days of the first study. Both studies were read by independent investigators blinded to the results of the other study. Diagnoses and therapeutic decisions regarding the use of continuous positive airway pressure obtained from the home and laboratory studies were compared. Agreement between the home and laboratory study recordings was also assessed using receiver operating characteristic (ROC) curves and Bland-Altman analysis. One half of the home studies were randomly assigned to be performed with a sleep technician's set up of the equipment in the patient's home (group 1), and the other half with the patient's own setup of the sleep-recording device (group 2), after an instruction period in the hospital. An economic analysis was performed, considering the cost of repeating studies in cases with faulty or inconclusive home studies (these patients should undergo polysomnography as a second step).
Results: Seven percent of the home studies in group 1, and 33% in group 2 produced no interpretable data because of artifacts (p < 0.05). Sixteen percent of the home study findings were inconclusive. The diagnosis obtained from the limited sleep-recording device and polysomnography agreed in 75% of the interpretable home studies (89%, if inconclusive home studies were excluded). The area under the ROC curve for the home study-derived parameters was between 0.84 and 0.89, compared with polysomnography. There was no bias between home and polysomnography studies in the Bland-Altman plot. The cost per study of home study recordings was less expensive than that of polysomnography (143.86 euros), either with (93.08 euros) or without (129.97 euros) intervention of the technician in the patient's home.
Conclusion: Home sleep studies are a viable form of diagnosing SAHS, and are less expensive than polysomnography. Intervention of a sleep technician in the patient's home was the least expensive strategy, because of the high percentages of faulty studies with the patient's own setup of the equipment, when using the limited sleep-recording device.