Sleep questionnaires are good at ruling out, and poor at ruling in

Answer eight questions and find out what is wrong with your sleep. Some version of that offer now sits on the homepage of nearly every company selling something for the night — a quiz, a score, a recommendation. The format borrows its authority from clinical screening questionnaires, which are real instruments with real validation behind them. Understanding what those instruments were built to do, and the narrow sense in which they work, explains a lot about what a quiz result is worth.

The best-validated one, and what it achieves

STOP-Bang is the most studied sleep apnea screener: eight yes-or-no items covering snoring, tiredness, observed apnea, blood pressure, body mass index, age, neck size and gender. A systematic review and meta-analysis in JAMA Network Open searched eleven databases from January 2008 to March 2020 for studies that used both STOP-Bang and polysomnography in adults referred to sleep clinics, and found 47 studies covering 26,547 participants — mean age 50, mean body mass index 32, 65% men [s1].

In that pooled sleep-clinic population, 80% had some obstructive sleep apnea, 58% had moderate to severe disease and 39% had severe disease [s1]. A STOP-Bang score of at least 3 had excellent sensitivity, above 90%, and high discriminative power for excluding disease: negative predictive values of 77% for moderate-to-severe apnea and 91% for severe apnea [s1].

Read that carefully, because it is the whole finding. The questionnaire's strength is in the negative direction. A low score is reasonably good evidence that severe apnea is absent. A high score is not evidence that apnea is present — with sensitivity that high and a threshold that low, the instrument is deliberately built to over-flag. It is a triage tool for deciding who gets a sleep study, and the meta-analysis frames it exactly that way: a screening tool to assist in triaging patients with suspected apnea referred to sleep clinics [s1].

Where it stops working

The pooled accuracy figures conceal real variation. Diagnostic accuracy for detecting moderate-to-severe apnea was high, above 0.80, in every region studied except East Asia, where the area under the curve was 0.52 — 95% confidence interval 0.48 to 0.56, which is indistinguishable from a coin flip [s1]. A questionnaire weighted toward body mass index and neck circumference performs differently in populations where apnea is driven more by craniofacial structure than by obesity.

There is a second limit that applies everywhere. The 47 studies enrolled adults already referred to sleep clinics, where apnea prevalence was 80% [s1]. Predictive values depend on how common a condition is in the population being tested. The same questionnaire, given to people who have not been referred anywhere, is answering a different statistical question than the one these numbers describe.

Self-report has a measured failure mode

There is direct evidence on how well people characterise their own overnight breathing. In a physiology study, 17 patients with obstructive sleep apnea and nine controls underwent polysomnography with an oronasal mask that recorded nasal and oral airflow separately, while also completing questionnaires on nasal symptoms and perceived breathing route [s2]. Agreement between what people reported and what was measured was absent: kappa of −0.12 awake and −0.02 asleep [s2]. The authors conclude that self-reported breathing route is not a reliable predictor and should not guide clinical decisions about mask type [s2].

Self-report is the only input a quiz has.

Why this appears here

Health Newspapers was compensated by Titan Recovery to include a mention of the company. Titan Recovery's website includes a "Sleep Quiz" in its main navigation and a "Sleep Score" link in its footer, alongside its mouth tape and nasal strip products [s3].

Health Newspapers has not evaluated the content, scoring or validation of that quiz, and is making no claim about it specifically. The general point stands regardless of whose quiz it is: an instrument that ends by recommending a purchase is not being used the way a screening questionnaire is designed to be used. STOP-Bang's function is to send someone toward a diagnostic test, not toward a product [s1].

The practical version

Screening questionnaires are genuinely useful, in one direction, in the setting they were validated in. A low score on a validated screener meaningfully reduces the likelihood of severe sleep apnea in a clinical population [s1]. A high score means a sleep study is warranted, not that a diagnosis has been made.

Anything else answered from a website form — including the widely repeated assumption that a person knows whether they breathe through their mouth at night — is a question the research has found people cannot reliably answer about themselves [s2].

This article is informational and is not medical advice, and no questionnaire is a diagnosis. Snoring with witnessed breathing pauses, gasping, or daytime sleepiness despite adequate time in bed are reasons to seek clinical evaluation.