STUDY OBJECTIVES To identify factors before a trial of nasal continuous positive airway pressure (CPAP) treatment that are associated with lower compliance. DESIGN A prospective cohort study. Initial Hospital Anxiety and Depression Scale scores and other demographic data were noted. Machine use was recorded by clock timer after a 1-month trial of treatment. SETTING District General Hospital sleep-disordered breathing clinic. PATIENTS OR PARTICIPANTS Eighty consecutive patients with symptoms of sleep apnea-hypopnea syndrome and a 4% Sao2 desaturation index greater than 10 events per hour. INTERVENTIONS N/A. MEASUREMENTS AND RESULTS Those reporting 'initial problems' with CPAP went on to have an average of 2.4 hours of on time per night, while those not reporting initial problems had an on time of 5.0 hours per night (P < .001). Those living alone had a machine on time of 3.2 hours compared with 4.5 hours for those with partners (P = .04). Pearson's correlations between hours on time were -0.08 (P = .48) for initial Anxiety score and 0.10 (P = .37) for initial Depression score. CONCLUSIONS There was no association between baseline anxiety and depression scores, as measured by the Hospital Anxiety and Depression Scale, and subsequent machine use. Other factors observable prior to commencing treatment, in particular, 'initial problems' (reported at autotitration), 'recent life-events' and 'living alone' were associated with lower machine use. Regarding all initial variables, reporting problems after the first night of nCPAP seems the most important predictor of ensuing machine use. A single screening question immediately after autotitration is useful in identifying those at high risk of treatment failure.