Nasal continuous positive airway pressure (CPAP) is presently considered as the "treatment of choice" for obstructive sleep apnea (OSA). Though some OSA patients adhere to treatment recommendations and ultimately respond quite well to CPAP therapy, there is a substantial subgroup for which compliance is a particularly difficult issue. Despite receiving recommendations to the contrary and for reasons that are incompletely understood, the majority of OSA patients settle for a partial compliance pattern. Whether a partial compliance schedule is beneficial or harmful is virtually unexamined: Unlike other medical treatments, there are few data concerning the "dose-response relationship" of CPAP to its desired effects. We argue that CPAP "dose" is a function not only of CPAP pressure but of time-on-CPAP as well. Critical questions that remain unanswered are what "dose" of CPAP is needed to effect an appropriate treatment outcome and which treatment outcomes should form the basis of our recommendations. Recent placebo-controlled studies comparing CPAP to suboptimal CPAP pressures may be informative in this regard. Directions for future research are suggested.