Identifying poor compliance with CPAP in obstructive sleep apnoea: a simple prediction equation using data after a...

It is important to identify those patients with OSA who are likely to benefit from long term CPAP, not only for symptomatic relief, but also potentially to reduce vascular morbidity and mortality, but are unlikely to adhere to treatment. We have validated a model which we developed previously for predicting long term compliance with CPAP using data after a 2 week trial.

The model was applied retrospectively to patients undergoing a trial of CPAP. Predicted outcomes were compared with the actual outcomes.

Prediction equation was applied to 448 patients [77% males, Age 53 ± 11 years, ESS 14 ± 4, AHI 37 ± 24]. Of 407 patients included in the study 333 were issued a CPAP and 74 declined long term CPAP. At one year, 81% patients were using CPAP at least 2 h and 70% > 4 h. A score >50% from the equation was associated with a high probability of CPAP usage at one year. 295 patients had a probability score of >50% and of them 84% were using CPAP satisfactorily at 1 year. The sensitivity in identifying compliers was 91%. Of the 112 patients with a score ≤50%, 38 opted to accept CPAP and 60% of them were still using it at 1 year.

This simple equation has now been validated to be highly sensitive in identifying long term compliers and it also identifies those with worse compliance. This group could be targeted for a more intensive follow up regime with the aim of improving their compliance.