Draft Research Plan
Obstructive Sleep Apnea in Adults: Screening
September 03, 2014
Recommendations made by the USPSTF are independent of the U.S. government. They should not be construed as an official position of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
This figure is the proposed analytic framework depicting the eight key questions and the research approach that will guide the evidence review outlined in this research plan. In general, the figure illustrates the overarching question (key question 1): whether screening for obstructive sleep apnea (OSA) in adults leads to improved health outcomes. The framework starts on the left with the patient population of interest: asymptomatic adults and persons with unrecognized symptoms of OSA. Moving from left to right, the figure depicts the ability of clinical prediction tools or screening questionnaires to distinguish persons who are more or less likely to have OSA (key question 2) and the accuracy and reliability of diagnostic tests for OSA (key question 3). There are potential harms of screening and diagnostic tools (key question 7). For adults with OSA, treatment may improve intermediate outcomes, such as changes in the apnea-hypopnea index, blood pressure, and daytime somnolence or sleepiness (key question 4), and may improve other health outcomes, such as mortality, quality of life, motor vehicle crashes, cardiovascular events, heart failure, headaches, and cognitive impairment (key question 5). Treatment may result in harms (key question 8). The framework includes assessment of the evidence supporting an association between the apnea-hypopnea index and the aforementioned health outcomes (key question 6).
Abbreviations: AHI = apnea-hypopnea index; KQ = key question; OSA = obstructive sleep apnea.
1a. What is the direct evidence that screening for obstructive sleep apnea (OSA) in adults improves health outcomes?
1b. How does the evidence differ for subgroups defined by age, sex, body mass index (BMI), or severity of OSA?
2. How accurately do currently existing clinical prediction tools or screening questionnaires distinguish persons in the general population who are more or less likely to have OSA?
3a. What is the accuracy and reliability of diagnostic tests for OSA?
3b. How does the accuracy and reliability of diagnostic tests for OSA differ for subgroups defined by age, sex, or BMI?
4a. How much does treatment with continuous positive airway pressure (CPAP), mandibular advancement devices, surgery, or weight loss programs improve intermediate outcomes (apnea-hypopnea index [AHI], blood pressure, or sleepiness) in persons diagnosed with OSA?
4b. How do the benefits of treatment differ for subgroups defined by age, sex, BMI, or severity of OSA?
5a. How does intervention with CPAP, mandibular advancement devices, surgery, or weight loss programs improve health outcomes in persons diagnosed with OSA?
5b. How do the benefits of intervention differ for subgroups defined by age, sex, BMI, or severity of OSA?
6. What is the evidence supporting an association between AHI and health outcomes?
7a. What are the harms associated with screening or diagnostic testing for OSA?
7b. Do the harms differ for subgroups defined by age, sex, or BMI?
8a. What are the harms associated with treatment of OSA?
8b. How do the harms of treatment differ for subgroups defined by age, sex, BMI, or severity of OSA?
Contextual questions will not be systematically reviewed and are not shown in the Analytic Framework.
- What is the rate of adherence to CPAP, mandibular advancement devices, and weight loss interventions in persons with OSA?
- What are the barriers to undergoing diagnostic testing for OSA (e.g., availability of polysomnography, ability to tolerate testing) and how often do those barriers prevent completion of testing?
- What is the evidence supporting an association between a reduction in sleepiness and quality of life, work productivity, motor vehicle crashes, or other health outcomes?
- What is the evidence supporting an association between a reduction in blood pressure and health outcomes?
- What are clinically meaningful changes in AHI, sleepiness (as measured by the Epworth Sleepiness Scale), and blood pressure?
- What is the evidence supporting an association between OSA and incident diabetes?
The Proposed Research Approach identifies the study characteristics and criteria that the Evidence-based Practice Center will use to search for publications and to determine whether identified studies should be included or excluded from the Evidence Review. Criteria are overarching as well as specific to each of the key questions (KQs).
KQs 1, 2: Asymptomatic adults and persons with unrecognized symptoms of OSA
KQs 3, 7: Asymptomatic adults, persons with unrecognized symptoms of OSA, and referral populations
KQs 4–6, 8: Persons with a confirmed diagnosis of OSA; the population may include asymptomatic and/or symptomatic adults
OSA severity to be defined as mild for AHI (or RDI) ≥5 and <15; moderate for AHI (or RDI) ≥15 and ≤30; and severe for AHI (or RDI) ≥30
Children and adolescents; pregnant women; studies of adults with acute stroke or other acute conditions that can trigger onset of OSA
Studies focused on screening, diagnosis or treatment of OSA among those with a rare condition (e.g., acromegaly)
KQs 4–6, 8: Studies of persons with suspected, but unconfirmed OSA
Studies conducted in countries categorized as “very high” on the Human Development Index, as defined by the United Nations Development Program
KQs 4, 5, 8: For nonsurgical interventions, studies must evaluate use at home rather than in a laboratory or facility (although the testing and outcome assessments may occur in sleep laboratories or other settings)
|KQs 4, 5, 8: For nonsurgical treatments, interventions studied only in laboratories (e.g., studies of CPAP conducted in sleep laboratories)|
Screening with the Epworth Sleepiness Scale, the STOP Questionnaire, the Berlin questionnaire, the Wisconsin Sleep Questionnaire, or the STOP-Bang Questionnaire
Risk stratification tools or clinical prediction rules that include multiple factors (e.g., the Multivariate Apnea Prediction Index)
|Studies assessing single patient characteristics or risk factors|
PSG conducted in a sleep laboratory, reviewed and interpreted by a qualified physician is the reference standard
Portable monitors used for home-based testing (including Type II, Type III, and Type IV monitors)
Home-based testing followed by PSG
|Treatment/ management interventions||
CPAP, mandibular advancement devices, surgery, and weight loss programs
Variations of fixed oral CPAP are eligible, including auto-titrating CPAP, nasal CPAP, bilevel CPAP, and humidification with CPAP
Atrial overdrive pacing, medications, palatal implants, oropharyngeal exercises, tongue-retaining devices, positional alarms, nasal dilator strips, acupuncture, auricular plaster, and other interventions not listed in the included column
Medications to treat sleepiness, sleep quality, or bruxism (rather than used to treat OSA), such as armodafinil, bromocriptine, donepezil, eszopiclone, and modafininl
Nasal steroids for treatment of allergic rhinitis or similar treatments that might secondarily improve OSA by treating another condition
Studies focusing on potential worsening of OSA that might be caused by treatments for other conditions (e.g., use of testosterone for hypogonadism, use of medications that may cause weight gain)
KQ 1: Screened versus nonscreened groups
KQ 2: Overnight PSG conducted in a sleep laboratory; studies may also determine or compare persons at increased, average, or decreased risk, or persons at higher and lower risk of OSA
KQ 3: Studies on accuracy of screening must include a comparison with PSG; studies on reliability of screening must include measures of reproducibility (e.g., test-retest, comparison between different laboratories or readers)
KQs 4, 5, 8: CPAP versus control or sham CPAP; mandibular advancement devices versus no treatment or inactive mandibular advancement devices; surgery versus sham, conservative treatment, or no treatment; and weight loss interventions versus control
KQ 6: Persons with higher and lower AHIs
KQ 7: Screened versus nonscreened groups or groups undergoing screening and/or diagnostic testing versus groups not undergoing screening and/or diagnostic testing
No comparison; nonconcordant historical controls; comparative studies of various interventions (e.g., comparing CPAP with mandibular advancement devices or comparing different types of CPAP)
KQs 2, 3: Studies with verification bias in which only a subgroup had PSG as the comparator
KQs 1, 5, 6: Mortality, quality of life (both disease-specific measures, such as the Functional Outcomes of Sleep Questionnaire, and general measures, such as SF-36), motor vehicle crashes, cardiovascular events, incidence of heart failure, headaches, cognitive impairment
KQ 2: Sensitivity, specificity, discrimination, calibration
KQ 3: Sensitivity and specificity; measures of reproducibility (e.g., test-retest, comparison between different laboratories or readers)
KQ 4: Change in AHI, blood pressure, daytime somnolence or sleepiness (e.g., measured with the Epworth Sleepiness Scale or other validated measures)
KQ 7: False-positive results leading to unnecessary treatment, anxiety, condition-specific distress, or stigma
KQ 8: Rash, irritation, needing additional sleep medications (e.g., to tolerate CPAP), claustrophobia, oral or nasal dryness, epistaxis, pain, excess salivation, tooth damage or loosening, complications of surgery (e.g., perioperative death, hemorrhage, nerve palsies, additional emergency surgery, cardiovascular events, respiratory failure, rehospitalization, speech or voice changes, difficulty swallowing, airway stenosis)
KQ 1: RCTs comparing screened versus nonscreened groups
KQ 2: Prospective cohort studies and cross-sectional studies that developed or evaluated screening questionnaires or clinical prediction tools
Previously published systematic reviews only to help identify existing studies
Clinical prediction tools and screening questionnaires must be externally validated
KQ 3: Good-quality, recent (within 5 years) systematic reviews comparing diagnostic tests with formal attended PSG conducted in a sleep laboratory
Primary studies published after the search cutoff of the most recent systematic review will be included (i.e., bridge searches will be performed to determine what is new since the review and whether it is consistent with the review)
KQs 4, 5: RCTs; previously published systematic reviews will be used only to help identify existing studies
KQ 6: Good-quality, recent (within 5 years) systematic reviews; bridge searches will be performed to determine what is new since the review and whether it is consistent with the review
Prospective cohort studies, following participants for at least 1 year, published after the search cutoff of the most recent systematic review will be included
Treatment studies included in KQs 4 or 5 reporting both change in AHI and change in a health outcome
KQ 7: Studies eligible for KQs 1, 2, or 3 reporting harms of screening or diagnostic tests
KQ 8: RCTs for all interventions; prospective cohort studies with at least 100 participants reporting harms of surgical interventions
All other designs
KQs 2, 3: Questionnaires, tools, and tests not validated in a group of participants separate from the sample used to develop the test
Abbreviations: AHI = apnea-hypopnea index; CPAP = continuous positive airway pressure; KQ = key question; OSA = obstructive sleep apnea; PSG = polysomnography; RCT = randomized, controlled trial; RDI = respiratory disturbance index.