BACKGROUND--Continuous positive airway pressure (CPAP) therapy is the treatment of choice for the sleep apnoea/hypopnoea syndrome. Compliance with this relatively obtrusive therapy has not been well studied. METHODS--Usage of CPAP was investigated in 54 patients with sleep apnoea/hypopnoea syndrome (median 36 (range 7-129) apnoeas + hypopnoeas/hour slept) over the first 1-3 months after starting CPAP therapy. In all cases CPAP usage was monitored by hidden time clocks that indicated for how long the machines were switched on--that is, the CPAP run time. In 32 patients the time at which the CPAP mask pressure was at the therapeutic level of CPAP pressure set for that patient--that is, the mask time--was also monitored. In all patients objective daytime sleepiness was assessed by multiple sleep latency before and after CPAP therapy. RESULTS--The mean (SE) nightly CPAP run time was 4.7 (0.4) hours. There was no correlation between run time and severity of the sleep apnoea/hypopnoea syndrome as assessed by apnoea + hypopnoea frequency or multiple sleep latency, and no correlation between CPAP usage and improvement in multiple sleep latency. Thirty two patients in whom mask time was recorded had therapeutic CPAP pressures for 89% (3%) of their CPAP run times. Patients who experienced side effects from CPAP used their CPAP machines significantly less than those who did not. CONCLUSIONS--Patients with sleep apnoea/hypopnoea syndrome used CPAP for less than five hours/night on average with no correlation between severity of sleep apnoea/hypopnoea syndrome and CPAP usage. Patients who complained of side effects used their CPAP therapy less. It is recommended that, as a minimum, CPAP run time should be regularly recorded in all patients receiving CPAP therapy.