PURPOSE: A higher prevalence of sleep apnea may be present in an inpatient population due to co-morbidities and anesthesia, putting patients at greater risk of poor outcome due to acute cardiovascular, pulmonary and neurologic complications. To address this issue a clinical pathway was implemented to identify, assess and treat patients with known or presumed sleep apnea.
METHODS: A new clinical pathway, PRAIS (Protocol for Recognition of Apnea in Sleep), was implemented in 2004 at a large tertiary care center. A three item nurse-administered questionnaire was considered positive and triggered a sleep medicine assessment, if at least one of the following were true: significant snoring, witnessed apnea in sleep, or prior CPAP use. During the hospitalization if witnessed apnea occurred or clinical history suggested sleep apnea, sleep medicine assessment took place, which included overnight oximetry recording when appropriate. Patients were excluded if their health status was inappropriate for testing or treatment e.g. DNR orders present, sinus surgery, etc., they refused testing or treatment, or they were discharged or expired before assessment could be conducted.
RESULTS: 969 inpatients were evaluated between January, 2004 and March, 2006. Age range was from 17 to 92. Of 365 receiving oximetry testing 321 (87.9%) were positive for significant desaturation episodes. 355 (36.6%) had past sleep apnea and 546 (56.3%) had a new presumptive diagnosis of sleep apnea, 54 (9.9%) of whom had new positive airway pressure treatment initiated. However, the subgroup from the last 6 months of data collection included 111 new sleep apnea diagnoses, 43 (38.7%) of whom had new PAP initiated.
CONCLUSION: An inpatient sleep apnea screening protocol can be successfully implemented to identify and treat patients with presumed sleep apnea.
CLINICAL IMPLICATIONS: Increased risk of complications due to sleep apnea may be reduced by screening and intervening on patients with easily identifiable clinical features.
DISCLOSURE: Debra Repasky, None.