Sleep Score

Take the Sleep Score

The methodology behind these numbers — what each score actually measures, where it overlaps with established clinical screeners (STOP-BANG, Epworth, the Insomnia Severity Index), what the result tiers mean, and what the screener doesn't catch — is the rest of this page. Read on if you want the rationale, or skip ahead to the result interpretation if you've already taken the screener and just want context for your scores.

How the score works

Sleep problems split roughly into two buckets that need different solutions. Sleep apnea is a physical breathing problem you usually can't feel happening — diagnosed by a sleep study, treated with airway therapy. Sleep wellness covers everything else — onset, maintenance, daytime function — and most often responds to behavioral and circadian interventions rather than medical hardware. The Sleep Score is a screener, not a diagnostic test, but it's calibrated against the same conceptual framework that drives the major clinical screeners (STOP-BANG, Epworth, Insomnia Severity Index) and points you toward the right next step.

Apnea Risk Score reflects the probability that breathing-disordered sleep is contributing to your symptoms. It's driven primarily by three of the seven questions: witnessed apnea or loud snoring (Q2), specific physical symptoms like waking gasping or with morning headaches (Q5), and risk factors like elevated BMI, large neck circumference, hypertension, or comorbid metabolic disease (Q7). Two further questions about morning state and daytime sleepiness (Q3, Q4) feed into apnea scoring at half-weight, because those signals are real but non-specific — they can equally indicate insomnia, circadian misalignment, or just acute sleep restriction.

Sleep Wellness Score reflects the quality of your sleep experience independent of breathing-disordered sleep. It's driven by Q1 (subjective overall quality) and Q6 (sleep onset and maintenance pattern). The same Q3 and Q4 signals about morning state and daytime sleepiness contribute at half-weight here too, because they're shared markers of sleep quality regardless of mechanism.

Sleep Health Score is a weighted composite — 60% apnea, 40% wellness — designed so that a low apnea sub-score pulls the composite down even if wellness looks fine. That weighting reflects two clinical realities: apnea is more under-diagnosed than insomnia (most insomniacs know they have it; most apneics don't), and untreated apnea drives outsized cardiovascular risk that wellness-side issues generally don't.

Each raw score is normalized so that higher is healthier, on a 0–100 scale. The thresholds for the three result tiers — Good (≥80), Fair (50–79), Poor (<50) — were calibrated to keep the Good band reserved for genuinely unconcerning patterns and to keep the Poor band tight enough that landing there is a real signal, not a false alarm. The 50 floor is the line below which the algorithmic confidence in "this person should see a clinician" is high enough to recommend it without hedging.

The combination of an Apnea tier and a Wellness tier produces nine possible result cells. Each maps to a distinct narrative, distinct recommended next step, and distinct CTA tone — soft (monitor), standard (clinical follow-up reasonable), or urgent (please see a specialist soon).

The colors on the matrix mirror the colors on the dials you saw when you took the screener. A good_good result means both dimensions clear; poor_poor means both dimensions concerning. The diagonal — fair_fair — is the most common landing spot for adults who took the screener because they had a vague sense something was off, and it's the spot where the recommendation is "talk to a sleep specialist" because the differential is genuinely open.

How it compares to clinical screeners

Three established screeners dominate adult sleep medicine: STOP-BANG for apnea risk, Epworth Sleepiness Scale (ESS) for daytime sleepiness, and the Insomnia Severity Index (ISI) for chronic insomnia. The Sleep Score is conceptually adjacent to all three but isn't a verbatim implementation of any of them. Here's where it overlaps and where it diverges.

STOP-BANG is an eight-item yes/no screener for obstructive sleep apnea. Items cover snoring, observed apnea, blood pressure, BMI, age over 50, neck circumference, and male sex. A score of 3 or more puts a patient in the moderate-risk bucket; 5 or more in the high-risk bucket. The Sleep Score's apnea-axis questions cover the same conceptual territory — Q2 maps to snoring and witnessed apnea, Q5 maps to physical symptoms that are downstream of apnea, Q7 packages BMI and comorbid hypertension and metabolic disease into a single item. We deliberately don't replicate STOP-BANG's age-over-50 and male-sex items; those are predictive at the population level but feel reductive at the individual level, and they'd add limited information given the symptoms-and-risk-factors items already capture most of the variance.

Epworth asks a patient to self-rate likelihood of dozing off in eight everyday situations. It produces a single 0–24 sleepiness score; ≥10 is conventionally considered abnormal. The Sleep Score's Q4 ("How often do you feel excessively sleepy during the day") substitutes a four-option summary item for the eight-situation Epworth, deliberately. The Epworth's situational granularity is valuable in clinic, where a patient's responses to specific items can guide differential thinking; in a screener answered by people without that scaffolding, asking the same patient to rate eight separate dozing scenarios produces noisier data than asking them once with concrete anchors.

Insomnia Severity Index is a seven-item self-report covering sleep onset latency, maintenance, early-morning awakening, satisfaction, daytime impairment, distress, and noticeability. ISI scores 0–28; ≥15 is moderate clinical insomnia. The Sleep Score's Q1 (subjective overall) and Q6 (onset and maintenance) capture the structural concerns ISI covers; the daytime-impairment items are absorbed into Q3 and Q4. The Sleep Score doesn't separately ask about distress or noticeability because in our population (adults arriving from search traffic) the act of taking the screener is itself a distress signal — if you weren't bothered, you wouldn't be here.

The reason for a three-dimensional combined screener rather than three separate single-purpose screeners is empirical: most adults presenting with disrupted sleep have signals on more than one axis, and a single-purpose screener forces the clinician (or the patient, in this self-screening context) to commit to a working diagnosis before the data justifies it. By scoring both axes simultaneously and returning a tier on each, the Sleep Score lets the result point at the differential rather than collapsing it.

If you're a clinician who wants the validated single-purpose tools, use them — they're well-calibrated, freely available, and have far more outcome literature behind them than this screener does. If you're a person who wants a five-minute first-pass that surfaces "this is probably apnea-driven" versus "this is probably wellness-driven" before you decide whether to push for a sleep study or try sleep hygiene first, that's what this is for.

Reading your result

The result page returns three dials — Sleep Health Score in the center, Apnea Risk and Sleep Wellness as the supporting pair — plus a narrative and a recommended next step calibrated to your specific tier combination.

The Good band (80+) means your answers don't point at a sleep disorder on that axis. It's not the same as a clean diagnosis — the screener can't see what it didn't ask about — but it does mean the high-yield signals are clear. A Good apnea score with a Good wellness score is the only result that comes with a "monitor and watch for changes" recommendation rather than a clinical follow-up step.

The Fair band (50–79) means your answers raise some concern on that axis but aren't unambiguous. Most adults who take this screener land in Fair on at least one dimension. Fair on apnea typically reflects either (a) some risk factors without a witness report, or (b) a witness report without the metabolic risk factor cluster — both worth a conversation with a clinician, neither so concerning that you need to act this week. Fair on wellness typically reflects sleep that's "fine but not great" — the kind of pattern that responds to sleep hygiene work and only escalates to a specialist if it persists past a couple of months.

The Poor band (below 50) means your answers are strong enough to recommend a specialist evaluation directly. Poor on apnea typically reflects multiple high-yield items lighting up — witnessed apnea plus risk factors, or risk factors plus physical symptoms. Poor on wellness typically reflects significant onset or maintenance disruption layered on top of poor daytime function. The combined Poor-Poor cell is the most concerning; it's the one labelled "please don't wait" because untreated apnea combined with significant insomnia or fragmentation has the largest cardiovascular and metabolic downside on a year-or-more horizon.

The recommended next step on the result page is one of three CTAs, calibrated to urgency: a "Find a Sleep Specialist" button (most cells), a "Book a Sleep Specialist" button (the urgent cells — fair_poor, poor_good, poor_fair, poor_poor), or a "Learn About Sleep Studies" button (good_good only). All three CTAs route through the same partner-resolution machinery — the bridge modal — that resolves to a board-certified sleep medicine practice in your area, or to an interest-signal capture if there isn't one yet. The urgency of the wording differs but the routing is identical.

What the score doesn't do

This is a screener, not a diagnostic test. The distinction matters in a few specific ways worth being explicit about.

It doesn't diagnose anything. Sleep apnea is diagnosed by polysomnography (the in-lab gold standard) or a home sleep apnea test reviewed by a board-certified sleep physician. Insomnia is diagnosed by clinical history meeting DSM-5 / ICSD-3 criteria. Circadian rhythm sleep-wake disorders are diagnosed by sleep diaries plus actigraphy. None of those workflows can be replaced by a seven-item self-report. A Good score on this screener does not rule out any of those conditions, and a Poor score does not establish any of them.

It doesn't cover all sleep disorders. The two-axis design captures the highest-prevalence patterns in adult sleep medicine — obstructive apnea and insomnia-class problems. It does not differentiate between obstructive and central sleep apnea, between primary and comorbid insomnia, or between insomnia and circadian rhythm disorders. It doesn't screen for narcolepsy, idiopathic hypersomnia, REM sleep behavior disorder, restless legs syndrome, parasomnias, or shift-work disorder. People with those conditions may produce ambiguous or misleading results; the screener wasn't built for them. If you have a strong suspicion of any of those — for example, vivid acting-out of dreams suggesting RBD, or persistent leg discomfort at sleep onset suggesting RLS — a sleep specialist evaluation is the right path regardless of what this screener says.

It doesn't apply to children. Sleep questions for children differ qualitatively from adult questions — pediatric apnea, for example, often presents with mouth-breathing and behavioral changes rather than excessive daytime sleepiness, and pediatric "insomnia" usually has very different drivers than adult insomnia. The thresholds and item weightings here are calibrated for adults. Pediatric concerns should be discussed with a pediatrician or pediatric sleep specialist directly.

It doesn't capture every relevant signal. The screener doesn't ask about medications (some medication classes profoundly disrupt sleep architecture), about alcohol use, about depression or anxiety symptoms (significant sleep disruption is sometimes a symptom of unaddressed mood disorders rather than a primary sleep complaint), about chronic pain, about menopause, or about shift work. All of those can shift the differential meaningfully. If any apply to you, bring them to your clinical visit alongside whatever this screener returned.

What to do next

Your specific next step depends on which tier combination you landed in, but a few principles cut across all of them.

If your apnea score is Poor, regardless of your wellness score, the highest-yield next step is a sleep evaluation with a board-certified sleep specialist. In most cases the workup will start with a home sleep apnea test rather than a full in-lab study; HSTs are accurate enough to confirm moderate-to-severe obstructive sleep apnea in adults with a high pretest probability, which a Poor apnea score on this screener establishes. Treatment for apnea is well-tolerated by most patients once they're on the right device and mask, and the cardiovascular risk reduction with sustained therapy is meaningful.

If your apnea score is Fair, the threshold for evaluation depends on whether your wellness score is also concerning. Fair-Fair and Fair-Poor patterns warrant a specialist visit; Fair-Good patterns are more of a "discuss with primary care, request a sleep study if symptoms persist" tempo. The risk of waiting on a Fair score is real but smaller than waiting on a Poor score.

If your wellness score is Poor with a Good apnea score, the most likely diagnosis is insomnia or a circadian rhythm pattern. Sleep hygiene fundamentals — consistent wake time, morning daylight exposure, caffeine cutoffs, screen reduction in the last hour before bed, cool dark bedroom — resolve a meaningful share of cases within 4–6 weeks. If they don't, cognitive behavioral therapy for insomnia (CBT-I) is the gold-standard treatment and outperforms sleep medications on long-term outcomes by a wide margin. Your primary care physician can refer to a CBT-I clinician, or programs are increasingly available digitally without a referral.

If both scores are Good, no specialist visit is indicated based on the screener alone. The follow-up worth doing is staying alert to changes — particularly new snoring, new daytime fatigue, or new morning headaches, all of which can be silent first signs of apnea developing later in life. A re-screen every year or two is reasonable. Sleep hygiene fundamentals still matter even with strong scores; they're the cheap insurance that keeps the scores good.

Whatever your tier, bringing the result printout to a clinical visit is helpful. The sub-scores can frame the conversation faster than a verbal symptom history ("my apnea axis is in the Fair band but my wellness axis is Poor" routes the discussion toward insomnia and away from apnea workup), and the per-question responses give the clinician a structured starting point. We don't generate a printable summary today, but the screener result is a stable URL while you have the page open — taking a screenshot is a serviceable workaround.

Privacy and how the data is used

The screener captures anonymous data when you complete it. Specifically: your three scores, the three tier classifications, the seven raw question responses, the page you arrived on, your browser's user agent string, and a randomly-generated session ID stored in your browser's local storage. No name, no email address, no IP address, no identity-linkable information. You don't have to enter anything beyond clicking through the seven questions; there is no email-gate, no account, no signup.

The reason for capturing anonymous data is calibration. The thresholds and item weightings in the scoring engine are based on best estimates from the published literature plus clinical judgment; they're approximations, not validated cutoffs. Aggregate distributions across tens of thousands of completions tell us whether the Fair band is too wide, whether the Apnea-axis half-weight items are pulling the right amount of variance, whether the urgency tones on the result narratives match the underlying severity of the patterns. Without aggregate data, the screener can't get better; with it, we can recalibrate every few months.

The session ID lets us avoid double-counting if you take the screener twice in the same session. It's not used to track you across sessions, across pages, or across devices. If you clear your browser's local storage, the next visit starts fresh.

The capture is fire-and-forget — if the network call to record your result fails, the screener still works and shows your result. No part of the user experience depends on the capture succeeding. We don't share, sell, or transfer the captured data to any third party. We don't use it for advertising. The legal and privacy framing is laid out in more detail in the site footer disclosures; this section is the practical version.