BACKGROUND: Rising requests for obstructive sleep apnea (OSA) diagnostic and treatment services may make traditional in-person evaluation processes unnecessarily expensive and inefficient.
PURPOSE: To assess the comparative effectiveness, harms, and costs of care models for OSA evaluation and treatment.
DATA SOURCES: MEDLINE (Ovid) and CINAHL searched for studies published in English language between January 2000 and May 2016 with hand searching of reference lists of related systematic reviews and included studies.
STUDY SELECTION: Four randomized controlled trials (RCTs) and 4 observational studies evaluated case-finding and care provided by nonsleep specialist practitioners versus sleep specialist physicians (SSP). No reports evaluated electronic versus interactive (eg, in-person or telephone) consultation. Twenty-seven reports (3 RCTs for titration, 22 RCTs for treatment and 2 cohort studies) assessed in-home autotitrating continuous positive airway pressure (APAP) technology versus standard continuous positive airway pressure (CPAP) titrated by in-lab polysomnogram (PSG) as options for PAP titration or treatment.
DATA EXTRACTION: Two investigators independently extracted study data, rated individual study risk of bias, and assessed overall strength of evidence.
DATA SYNTHESIS/RESULTS: No studies assessed the diagnostic accuracy of non-sleep-specialist nurses for case finding and referral. One retrospective study reported good agreement between a primary care pulmonologist and a SSP on what sleep test to order for patients referred by their family physician. Patient-centered outcomes were infrequently and inconsistently reported. When reported there was no significant difference in clinical outcomes between OSA treated by primary care/nurses and SSPs (moderate strength of evidence for quality of life). Sleep symptom scores were similar between groups (moderate strength of evidence). Treatment compliance was similar between patients treated by SSPs and those not (moderate strength of evidence). Strength of evidence for access to care and adverse events was insufficient.
Few studies compared patient-centered outcomes between in-lab CPAP titration and at-home APAP titration. In limited reporting, study groups were generally similar on quality of life (moderate strength of evidence) and cognitive symptoms. Some differences were noted for resource utilization and patient preference. Sleep measures, blood pressure, adverse events, and compliance/adherence were generally similar (moderate strength of evidence for Epworth Sleepiness Scale (ESS) scores and low strength of evidence for compliance).
Twenty-three studies compared treatment with CPAP versus APAP. Few studies reported patient-centered outcomes other than quality of life and patient preference for one treatment approach over another. Quality of life, assessed with the SF-36, was generally similar (moderate strength of evidence). Patient preference was generally similar or favored APAP. Strength of evidence was insufficient for access to care. Post-treatment ESS scores were generally similar for the 2 treatment approaches (moderate strength of evidence). Adverse events were mild and similar for APAP and CPAP (low strength of evidence). Compliance was similar (moderate strength of evidence).
LIMITATIONS: Studies were limited to English language and those published in the United States, Canada, Western Europe, Australia, or New Zealand. Few studies assessed non-sleep specialist case-finding or care and none assessed electronic consultations. Clinical outcomes were infrequently reported.
CONCLUSIONS: Among patients suspected of having OSA, primary care providers and sleep-specialist nurses provide similar outcomes to SSPs, although many outcomes were inconsistently reported. At-home APAP provides similar outcomes to CPAP titrated in the PSG laboratory. No evidence addressed electronic consultation for the management of known or suspected OSA. Future studies are needed to determine which patients derive the most benefit from treatment and should be prioritized for testing and treatment, whether newer models of care with less reliance on SSP time (either through utilization of other types of providers or electronic consultation) result in similar outcomes to traditional models, and, if effective, how such models should be implemented.