Last updated July 2026. Medically reviewed for accuracy. Reading time: approximately 16 minutes.

Category: Optimization — This article compares every validated way to screen for obstructive sleep apnea, from five-question checklists to AI-enabled wearables, so you know which tool actually fits your situation. See also our guide to understanding sleep cycles and the Sleep Apnea Risk Screener.

An estimated 936 million adults worldwide have mild-to-severe obstructive sleep apnea, and 425 million have moderate-to-severe disease — figures nearly ten times higher than the World Health Organization's 2007 estimate , according to a ResMed-led multinational analysis published in The Lancet Respiratory Medicine. Most of those people have never been tested. Polysomnography — the overnight, technician-monitored, electrode-covered gold standard — is expensive, slow to schedule, and simply unavailable to most of the population that needs it. That gap is exactly what OSA screening tools exist to close: quick, low-cost, low-friction ways to flag who is actually at meaningful risk and worth referring for a real diagnostic test.

But "OSA screening tools" is not one thing. It is a whole category — five-minute questionnaires, take-home breathing monitors, and smartwatches that estimate your apnea-hypopnea index (AHI) while you sleep. Each has a different accuracy profile, a different price, and a different appropriate use case, and conflating them is how people end up either falsely reassured or unnecessarily alarmed.

This guide breaks down every major category of OSA screening tool, shows you the actual sensitivity and specificity numbers behind each one (not marketing claims), walks through a worked scoring example, and gives you a decision framework for picking the right tool for your situation — before you ever set foot in a sleep lab.

---

OSA Screening Tools: How the Major Options Actually Compare

The Three-Tier Model of OSA Screening

Every OSA screening tool on the market falls into one of three tiers, and understanding the tier tells you almost everything about what the tool can and can't do for you.

Tier 1 — Self-report questionnaires. Paper or digital checklists scored from your own answers about snoring, tiredness, blood pressure, and body measurements. No hardware, no data collection while you sleep — just pattern-matching against population risk factors. STOP-Bang, the Berlin Questionnaire, the Epworth Sleepiness Scale (ESS), and the NoSAS score all live here.

Tier 2 — Home sleep apnea testing (HSAT). Portable, prescription-based devices you wear for one or more nights at home that directly measure airflow, effort, and oxygen saturation to calculate an actual (if simplified) AHI. This is Type III or Type IV monitoring, not full polysomnography, but it produces objective physiological data rather than a risk score from a questionnaire.

Tier 3 — Consumer wearables and AI-enabled sensors. Smartwatches, rings, and acoustic or radar-based bedside devices that use machine learning to estimate AHI from heart rate variability, movement, oxygen saturation, or sound, without a prescription in most cases. This tier has expanded fastest in the past two years and now includes FDA-cleared clinical-grade options alongside consumer wellness features.

None of these tiers replaces polysomnography (PSG) for a formal diagnosis. What they do is triage: they tell you, cheaply and quickly, whether PSG is worth pursuing at all. If you're unsure whether your daytime exhaustion is even related to breathing disruptions versus general sleep debt, running the numbers through the main Sleep Debt Calculator first can help you separate "I'm not sleeping enough hours" from "something is disrupting the sleep I am getting."

Tier 1: The Validated Screening Questionnaires

Questionnaires are the most widely used OSA screening tools because they cost nothing, take under five minutes, and require no equipment. But their accuracy varies enormously depending on which one you use and which population you belong to.

STOP-Bang: the most-studied option

STOP-Bang is an eight-item yes/no questionnaire covering Snoring, Tiredness, Observed apnea, blood Pressure, BMI, Age, Neck circumference, and Gender. A meta-analysis of 17 studies covering 9,206 patients found sensitivity for detecting any OSA (AHI ≥ 5) reached 90% in sleep clinic populations, climbing to 96% for detecting severe OSA (AHI ≥ 30) , with a corresponding negative predictive value of 90% for severe disease. In other words, a low STOP-Bang score is genuinely useful for ruling severe OSA out.

A separate systematic review focused on the general population and commercial drivers found that a STOP-Bang score of 3 or higher reliably helped detect and rule out clinically significant OSA, with prevalence estimates in the studied population reaching 21.3% for moderate-to-severe disease across roughly 8,585 general-population participants and 185 commercial drivers.

The tradeoff: STOP-Bang trades specificity for sensitivity. Its screen-everyone-in approach means a large share of low-risk people also screen positive.

Berlin Questionnaire

The Berlin Questionnaire scores three symptom categories (snoring/apnea, daytime sleepiness, and blood pressure/BMI) into a high- or low-risk classification. Berlin performs comparably to STOP-Bang on sensitivity in most populations — one comparison across six assessment tools in hypertensive patients found Berlin's sensitivity and negative predictive value were the highest of any tool tested at every AHI cutoff studied , reaching 98.8% sensitivity and 95.7% negative predictive value at the AHI ≥ 30 threshold. Like STOP-Bang, it sacrifices specificity for that sensitivity.

Epworth Sleepiness Scale (ESS)

The ESS is often confused with an OSA screening tool, but it measures something different: subjective daytime sleepiness, not sleep-disordered breathing directly. That distinction matters. A meta-analysis of 34 studies found the ESS had a sensitivity of just 0.48 against a specificity of 0.73 , while STOP-Bang in the same analysis reached 0.89 sensitivity but only 0.40 specificity — almost the mirror image. The ESS is the most specific tool of the three but will miss roughly half of true OSA cases used alone. It is best paired with another tool, not used standalone. If persistent daytime sleepiness is your primary concern regardless of cause, our Why Am I Tired tool can help separate sleepiness from other causes of fatigue before you attribute it to OSA.

NoSAS score

The NoSAS score (Neck circumference, Obesity, Snoring, Age, Sex) is a newer, simpler alternative built to require less clinical interpretation. In a comparison of six OSA screening tools in patients with hypertension, NoSAS was evaluated alongside Berlin, STOP, STOP-Bang, and ESS, with researchers noting Berlin's rank correlation with AHI outperformed the alternatives in that specific population even though NoSAS remains a viable lower-burden option in general screening settings. In patients with comorbid insomnia specifically, a 1,266-participant study found STOP-Bang had the highest sensitivity at 93.2% while NoSAS had the highest specificity of the tools compared — useful context if insomnia symptoms are muddying your picture.

Head-to-head questionnaire comparison

Tool	Typical Sensitivity	Typical Specificity	Best For	Weakness
STOP-Bang	89–97%	26–40%	Ruling out OSA (high NPV)	High false-positive rate
Berlin Questionnaire	65–95%	16–61%	General population triage	Specificity varies widely by study
Epworth Sleepiness Scale	34–54%	58–85%	Confirming sleepiness is present	Misses ~half of true OSA cases alone
NoSAS Score	Comparable to Berlin	Often highest specificity	Populations with comorbid insomnia	Less studied than STOP-Bang

Ranges reflect variation across the cited meta-analyses and population-specific validation studies referenced throughout this article; actual performance shifts with age, sex, BMI distribution, and clinical setting.

---

A Worked Example: Scoring Your Own STOP-Bang

Because STOP-Bang is the most-validated and most widely deployed OSA screening tool, it's worth walking through exactly how the scoring works — this is also embedded as a quick self-check you can run right now.

STOP-Bang Scoring Worksheet
----------------------------
S — Snoring loud enough to be heard through a closed door?         [ ] Yes = 1
T — Tiredness/fatigue/sleepiness during the daytime, most days?    [ ] Yes = 1
O — Observed apnea: has anyone seen you stop breathing in sleep?   [ ] Yes = 1
P — high blood Pressure, treated or untreated?                     [ ] Yes = 1
B — BMI greater than 35 kg/m²?                                      [ ] Yes = 1
A — Age over 50 years?                                              [ ] Yes = 1
N — Neck circumference greater than 40 cm (about 15.75 in)?         [ ] Yes = 1
G — Gender: male?                                                   [ ] Yes = 1

TOTAL SCORE: ___ / 8

Interpreting the total:

Score	Risk Category	What the Evidence Suggests
0–2	Low risk	Severe OSA unlikely; consider other causes of symptoms
3–4	Intermediate risk	Discuss with a clinician; consider objective testing if symptomatic
5–8	High risk	Meta-analytic data links higher scores to sharply rising probability of severe OSA

That last row matters more than it looks. One meta-analysis found the probability of severe OSA in a sleep clinic population rose step-wise with STOP-Bang score — from 25% at a score of 3 up to 75% at a score of 7 or 8 , with the same stepwise pattern (15% rising to 65%) observed in surgical populations. A single point on this scale, in other words, is not a rounding error — it materially shifts your probability of having disease worth treating.

If your BMI happens to sit below 35 kg/m² — the range where standard STOP-Bang cut-offs were originally validated — a 2025 study identified refined thresholds specifically for that group, finding that age ≥ 40, BMI ≥ 23 kg/m², and neck circumference ≥ 35 cm each carried strong individual predictive value and proposed a modified scoring approach for people under the standard obesity threshold. This matters because a large share of OSA cases occur in people who are not obese by the BMI ≥ 35 standard, and the unmodified questionnaire underweights their risk.

Why this framework is not just theory: untreated OSA carries a relative cardiovascular event risk of 1.96 compared with people without the condition, and an 82% higher adjusted odds of hospitalization when undiagnosed , according to data compiled in a 2025 PeerJ analysis of modified STOP-Bang scoring. Getting the screening step right is not a formality — it's the entry point to preventing those downstream outcomes.

Once you have a working risk category from this worksheet, running your symptoms and score through the Sleep Apnea Risk Screener can help you frame the conversation with a clinician and decide whether Tier 2 testing makes sense next.

---

Tier 2: Home Sleep Apnea Testing (HSAT)

Once a questionnaire flags you as intermediate or high risk, the next step for many patients isn't a lab-based PSG — it's a home sleep apnea test. HSAT devices are prescribed, worn overnight in your own bed, and directly measure physiological signals like airflow, respiratory effort, and blood oxygen saturation rather than inferring risk from self-reported symptoms.

The American Academy of Sleep Medicine and other international bodies now recommend HSAT as a viable alternative to full PSG for patients without severe cardiovascular disease or suspected central sleep apnea, citing evidence of comparable diagnostic accuracy in appropriately selected populations , a shift that accelerated significantly during and after the COVID-19 pandemic as sleep labs faced capacity constraints. That said, HSAT has real limits: it still requires a physician's prescription, carries meaningful cost, lacks EEG data so it cannot assess actual sleep architecture, and depends on skilled scoring , making it impractical to repeat nightly for tracking night-to-night AHI variability.

A real-world example of HSAT deployment: a 2024 quality-improvement project in a psychiatric outpatient department integrated STOP-Bang screening directly into routine visits, then issued WatchPAT home sleep apnea testing devices to every patient who screened high-risk , reporting that the combined screen-then-test workflow proved both feasible and effective at triaging patients who would otherwise have gone unassessed for OSA. This two-step model — cheap questionnaire first, objective home test second — is becoming the standard pathway across primary care and specialty settings alike, precisely because it avoids sending every low-risk patient straight to an expensive PSG slot.

---

Tier 3: Wearables and AI-Enabled OSA Screening

This is the fastest-moving corner of the OSA screening tools landscape, and it's where the line between "consumer gadget" and "clinical-grade instrument" is blurring in real time.

What the accuracy data actually shows

A systematic review and meta-analysis of wearable AI performance across 38 qualifying studies found pooled accuracy of 86.9%, sensitivity of 93.8%, and specificity of 75.2% for detecting sleep apnea generally , though accuracy dropped to a pooled 65.1% for correctly identifying OSA severity level and estimating the actual AHI score at 87.7% pooled accuracy. That gap matters: wearables are currently better at telling you "yes, something is wrong" than at telling you exactly how severe it is.

Individual device studies back this up with more specific numbers. A smartphone-and-smartwatch system tested against 350 hospital-based sleep disorder patients found a 0.92 correlation with lab-measured AHI and classification accuracy of 88.1%, 84.5%, and 85.1% for correctly identifying mild, moderate, and severe OSA respectively , with sensitivity ranging from 84.2% to 89.1% across those same severity bands. A separate 2025 real-world validation study comparing an AI-enhanced smartwatch algorithm against AI-scored Level 1 polysomnography in Korean adults concluded the technology showed strong enough concordance to justify prescribing a smartwatch for two to three days of monitoring specifically to prioritize which patients most urgently need formal PSG testing , positioning wearables as a triage layer on top of triage rather than a full PSG replacement.

Acoustic and other non-wearable sensor approaches are advancing in parallel. One acoustic-based device tested against standard Type 1 PSG in 80 patients referred through primary-care suspicion of OSA recorded overnight alongside lab polysomnography in a Spanish hospital sleep center to validate automatic diagnosis performance.

Where wearables currently fall short

Wearable OSA detection is genuinely promising but not yet a diagnostic substitute. The severity-classification accuracy gap noted above is the central limitation — a device that reliably flags "you likely have apnea" but is meaningfully less reliable at telling you whether it's mild or severe leaves an important clinical decision unresolved. Battery life, night-to-night variability in fit and signal quality, and the fact that most consumer devices have not undergone the same peer-reviewed validation rigor as STOP-Bang or Berlin are all practical caveats worth knowing before you treat a smartwatch "sleep apnea risk" notification as a diagnosis.

Comparing wearable performance to questionnaires

Screening Method	Detecting Any OSA (Accuracy Range)	Severity Classification	Cost	Access
STOP-Bang questionnaire	90% sensitivity (sleep clinic)	Indirect (score-based tiers)	Free	Immediate, self-administered
Wearable AI (pooled, general)	86.9% accuracy	65.1% accuracy	$150–$400 (device)	No prescription typically needed
Best-case smartwatch algorithm	0.92 correlation with lab AHI	84–88% by severity band	$200–$400+	Varies by device/market clearance
HSAT (Type III/IV)	Comparable to PSG in selected patients	Direct AHI measurement	$150–$500 (with prescription)	Requires physician order

---

Choosing the Right OSA Screening Tool and Knowing the Limits

A Hierarchy of Interventions Ranked by Evidence Strength

If you're deciding where to start, this ranking reflects the evidence quality and appropriate use-order for each tool discussed above:

1. Start with a validated questionnaire (STOP-Bang or Berlin) — free, takes minutes, and has the deepest evidence base of any option here.
2. If your score lands in the intermediate-to-high range, add a symptom-specific check, such as the Epworth Sleepiness Scale, to confirm whether daytime sleepiness specifically is part of your presentation, since ESS catches a different signal than STOP-Bang.
3. Bring your results to a clinician before buying a device. A prescription-based HSAT unit remains the most cost-effective objective test for most symptomatic adults who don't have complicating cardiovascular or central apnea concerns.
4. Consider a wearable as a bridge or monitoring tool, not a diagnostic one — useful for tracking trends over time or triaging urgency, less useful as a stand-alone basis for treatment decisions given the current severity-classification accuracy gap.
5. Polysomnography remains the reference standard for anyone whose screening results are ambiguous, whose symptoms are severe, or who has comorbidities (cardiovascular disease, suspected central sleep apnea) that make simplified monitoring less reliable.

What Screening Tools Are Not Designed to Do

It's worth being direct about an important nuance: none of the tools in this article are endorsed for routine screening of adults who have no symptoms at all. The U.S. Preventive Services Task Force reviewed the evidence on screening asymptomatic adults for OSA and concluded, consistent with its earlier 2017 statement, that current evidence remains insufficient to weigh the balance of benefits and harms , a conclusion based in part on seven studies assessing clinical prediction tools including the Berlin Questionnaire, STOP-Bang, and the Multivariable Apnea Prediction tool against facility-based polysomnography. This is explicitly not a recommendation against screening — the USPSTF was careful to clarify its "I statement" means neither for nor against — but it does mean these tools are best used when you have symptoms or known risk factors, not as a blanket annual check for the general population , and it does not apply to people with recognized OSA symptoms or those in higher-risk groups where clinician judgment should still guide testing decisions.

Risk factors worth taking seriously enough to screen even without an official general-population mandate include male sex, age between 40 and 70, postmenopausal status, higher BMI, and craniofacial or upper airway abnormalities, per the same USPSTF evidence review , while evidence on smoking, alcohol or sedative use, and nasal congestion as independent risk factors remains sparse or mixed.

Building a Realistic Screening Plan

A sensible, evidence-aligned plan for most adults with suspected symptoms looks like this: run a STOP-Bang or Berlin Questionnaire this week, cross-check with the Epworth Sleepiness Scale if daytime sleepiness is prominent, and use those combined results — not a single score in isolation — as the basis for a conversation with your primary care provider about whether HSAT or PSG is the appropriate next step. If irregular sleep timing, jet lag, or shift work is complicating your symptom picture, it's also worth ruling that out separately using the Weekly Sleep Planner or Jet Lag Recovery tool before assuming every symptom points to apnea specifically.

---

Frequently Asked Questions

What is the most accurate OSA screening tool available without a prescription?

Among self-administered options, STOP-Bang and the Berlin Questionnaire have the strongest and most consistent sensitivity across large meta-analyses, though both trade away specificity, meaning they flag many people as at-risk who don't ultimately have significant disease. Consumer wearables now approach similar accuracy for detecting the presence of apnea but remain less reliable for classifying its severity. No unprescribed tool replaces a clinical diagnosis — treat a positive screen as a reason to seek testing, not as a diagnosis itself.

Can a smartwatch actually diagnose sleep apnea?

Not on its own. Current wearable AI systems reach strong pooled sensitivity for detecting that sleep apnea is present, but meaningfully lower accuracy for correctly classifying its severity level, which is the information that actually drives treatment decisions. Recent studies support using smartwatches to help prioritize who needs formal testing soonest, not as a substitute for a physician-ordered test.

How is the STOP-Bang questionnaire scored, and what score means I should see a doctor?

STOP-Bang awards one point for each of eight yes/no criteria (snoring, tiredness, observed apnea, blood pressure, BMI, age, neck circumference, and male gender) for a total possible score of 8. A score of 3 or higher is generally treated as a signal worth discussing with a clinician, and scores of 5 and above are associated with sharply rising probability of moderate-to-severe disease in validation studies. Try the worked worksheet earlier in this article, or the Sleep Apnea Risk Screener, to calculate your own score.

Is the Epworth Sleepiness Scale a reliable way to screen for sleep apnea?

Not by itself. The ESS measures subjective daytime sleepiness rather than sleep-disordered breathing directly, and multiple validation studies have found it misses roughly half of true OSA cases when used alone, despite being reasonably specific when it is positive. It's most useful paired with a symptom- and risk-factor-based tool like STOP-Bang or Berlin, not as a stand-alone screen.

What's the difference between a home sleep apnea test and a wearable device?

A prescribed home sleep apnea test (HSAT) directly measures airflow, respiratory effort, and blood oxygen to calculate an actual simplified AHI, and is recommended by sleep medicine bodies as a viable PSG alternative for many patients. A consumer wearable typically estimates AHI indirectly from signals like heart rate variability, movement, or oxygen saturation trends using machine learning, generally without a prescription, and with somewhat lower accuracy for severity classification specifically.

Should healthy adults with no symptoms get screened for OSA anyway?

Current evidence reviewed by the USPSTF is insufficient to recommend for or against routine screening of asymptomatic adults, which is different from a recommendation against it. If you have known risk factors — elevated BMI, a large neck circumference, being male, being over 40, or a family history of OSA — discussing screening with a clinician remains reasonable judgment even without a formal population-wide mandate.

Why do different OSA screening tools give me different risk levels for the same symptoms?

Each tool weighs different inputs and was validated against different populations, so a questionnaire built around snoring and blood pressure (STOP-Bang) can disagree with one weighted more heavily toward subjective sleepiness (Epworth). Studies directly comparing these tools head-to-head consistently show meaningful differences in sensitivity and specificity by population — sleep clinic patients, hypertensive patients, and people with comorbid insomnia all produce different accuracy profiles for the same questionnaire. This is exactly why combining two tools, rather than relying on one, produces a more reliable picture.

Do OSA screening tools work the same way for people who aren't obese?

No — standard cut-offs for tools like STOP-Bang were built around a BMI ≥ 35 kg/m² threshold, which can underweight risk in people below that range. Modified scoring approaches using lower BMI, age, and neck circumference thresholds have been developed specifically for non-obese populations and should be used if your BMI sits well below 35.

---

The Bottom Line

No single OSA screening tool is definitively "the best" — the right choice depends on whether you need a free five-minute risk estimate, an objective at-home physiological measurement, or ongoing trend monitoring. What the evidence does make clear is that questionnaires like STOP-Bang and Berlin remain the most validated starting point, wearables are rapidly closing the accuracy gap for detection (if not yet severity), and none of these tools substitutes for a clinician-guided path to polysomnography when your results warrant it.

A concrete action plan:

1. Complete the STOP-Bang worksheet above, or run your numbers through the Sleep Apnea Risk Screener, this week.
2. If your score is 3 or higher, add the Epworth Sleepiness Scale to check whether daytime sleepiness is a distinct, corroborating symptom.
3. Bring both results to a primary care visit and ask specifically whether HSAT is appropriate for your risk profile before assuming you need a full lab-based sleep study.
4. If you're already using a wearable, treat any apnea-risk alert as a prompt to seek formal testing — not as a diagnosis to self-manage.
5. Recheck your overall sleep debt using the Sleep Debt Calculator so you're not attributing every symptom of fatigue to apnea alone when insufficient sleep duration may also be a factor.

Untreated OSA is not a minor inconvenience — it carries measurably elevated cardiovascular and hospitalization risk, and it affects roughly a billion people worldwide, most of them undiagnosed. Picking the right screening tool, and actually using it, is the single highest-leverage step most people can take this month toward finding out where they actually stand.

---

Tools Referenced in This Article

• Sleep Apnea Risk Screener — score your OSA risk using validated questionnaire criteria
• Sleep Debt Calculator — separate insufficient sleep duration from sleep-disruption symptoms
• Why Am I Tired — diagnose the likely cause of persistent daytime fatigue
• Weekly Sleep Planner — check whether irregular scheduling is contributing to your symptoms
• Jet Lag Recovery — rule out circadian disruption before assuming apnea

---

Related Reading

• What Is Sleep Debt? — Health — the foundational concept behind chronic sleep insufficiency and how it differs from a sleep disorder like OSA.
• Understanding Sleep Cycles — Optimization — how disrupted sleep architecture from conditions like OSA affects deep and REM sleep.
• The Real Cost of Poor Sleep — Productivity — how undiagnosed sleep disorders translate into measurable work and cognitive impact.

---

References

1. Benjafield AV, Ayas NT, Eastwood PR, et al. Estimation of the global prevalence and burden of obstructive sleep apnoea: a literature-based analysis. Lancet Respir Med. 2019;7(8):687-698. https://www.thelancet.com/journals/lanres/article/PIIS2213-2600(19)30198-5/abstract
2. Hwang M, Nagappa M, Guluzade N, Saripella A, Englesakis M, Chung F. Validation of the STOP-Bang questionnaire as a preoperative screening tool for obstructive sleep apnea: a systematic review and meta-analysis. BMC Anesthesiol. 2022. doi:10.1186/s12871-022-01912-1. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9710034/
3. Chen L, Pivetta B, Nagappa M, et al. Validation of the STOP-Bang questionnaire for screening of obstructive sleep apnea in the general population and commercial drivers: a systematic review and meta-analysis. Sleep Breath. 2021. doi:10.1007/s11325-021-02299-y. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8590671/
4. Nagappa M, et al. Validation of the STOP-Bang Questionnaire as a Screening Tool for Obstructive Sleep Apnea among Different Populations: A Systematic Review and Meta-Analysis. PLOS One. 2015. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4678295/
5. Modified STOP-Bang questionnaire for detecting obstructive sleep apnea in individuals with a body mass index below 35 kg/m². PeerJ. 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12617366/
6. Screening for Obstructive Sleep Apnea using Epworth Sleepiness Score and Berlin Questionnaire: Which is Better? Meta-analysis of 34 studies. https://www.researchgate.net/publication/322828108_Screening_for_obstructive_sleep_apnea_using_epworth_sleepiness_score_and_berlin_questionnaire_Which_is_better
7. Evaluation of five questionnaires for obstructive sleep apnea screening in the elderly. Sci Rep. 2025. https://www.nature.com/articles/s41598-025-86041-8
8. Comparison of six assessment tools to screen for obstructive sleep apnea in patients with hypertension. PMC. 2021. https://pmc.ncbi.nlm.nih.gov/articles/PMC8571550/
9. Performance of Four Screening Tools for Identifying Obstructive Sleep Apnea Among Patients with Insomnia. PMC. 2024. https://pmc.ncbi.nlm.nih.gov/articles/PMC11887493/
10. Better Sleep, Better Care: Streamlining Obstructive Sleep Apnea Screening in Psychiatric Outpatients. PMC. 2024. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12438339/
11. AI-Enhanced Smartwatch AHI Estimation and AI-Scored Polysomnography for Obstructive Sleep Apnea: Real-World Validation. PMC. 2025. doi:10.2147/NSS.S540460. https://pmc.ncbi.nlm.nih.gov/articles/PMC12474664/
12. Automatic monitoring of obstructive sleep apnea based on multi-modal signals by phone and smartwatch. PubMed. https://pubmed.ncbi.nlm.nih.gov/38083356/
13. Aziz S, Ali AM, Aslam H, et al. Detection of Sleep Apnea Using Wearable AI: Systematic Review and Meta-Analysis. J Med Internet Res. 2024;27:e58187. doi:10.2196/65272. https://www.jmir.org/2024/1/e58187
14. Validation of a Wearable Medical Device for Automatic Diagnosis of OSA against Standard PSG. PMC. 2024. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10816319/
15. Recommendation: Obstructive Sleep Apnea in Adults: Screening. U.S. Preventive Services Task Force. 2022. https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/obstructive-sleep-apnea-in-adults-screening
16. Screening for Obstructive Sleep Apnea in Adults. American Academy of Family Physicians. 2023. https://www.aafp.org/pubs/afp/issues/2023/0300/uspstf-obstructive-sleep-apnea.html

---

Disclaimer: This article is for educational and informational purposes only and does not constitute medical advice. Obstructive sleep apnea is a diagnosable medical condition, and screening tools discussed here are not a substitute for evaluation, testing, or treatment by a qualified healthcare professional. If you suspect you may have sleep apnea, consult your physician or a board-certified sleep medicine specialist.