To estimate reliability of self-reported compliance with nasal continuous positive airway pressure (CPAP) therapy for obstructive sleep apnea (OSA), we studied 63 OSA patients aged 53.7 +/- 1.2 years (mean +/- SEM) with an apnea hypopnea index (AHI) of 50.8 +/- 2.9 and lowest sleep SaO2 of 65.6 +/- 2.3 percent receiving nasal CPAP for 539 +/- 44 days. During a follow-up polysomnography (PSG) on the pressure prescribed for home therapy (10.3 +/- 0.3 cm H2O), the hours of operation shown on the built-in time counter of the patients' devices were read to determine objective compliance by dividing the run time by the days since initiation of therapy. This parameter was compared with subjective compliance reported in a self-administered questionnaire. Mean measured use time was 4.9 +/- 0.3 h per night, whereas reported daily use time calculated from reported nights a week and hours a night was 6.1 +/- 0.3 h per night. As predominantly patients with poor compliance misestimated daily use time, we conclude that self-reports are unable to distinguish between compliant and noncompliant patients.