I would say that patients who use CPAP all (or even most) of the time asleep after a years or so are getting some benefit (though perhaps not the optimum benefit that they could be getting). Otherwise, it's hard to imagine that they'd continue use it.
What I would suggest is an alternative model of compliance (in addition to the existing model). For instance fewer hours, but with a low AHI should be considered compliant as well as the existing model.
Also, a different word than "compliance" should be used that does not imply that the patient is at fault when (s)he does not meet the criteria.
