For decades, a sleep apnea diagnosis has hinged on a single night of data. A patient spends one night in a lab, or wears a home monitor for one night, and the number of breathing interruptions per hour — the Apnea-Hypopnea Index, or AHI — determines whether they're told they have mild, moderate, or severe obstructive sleep apnea (OSA). That number then drives whether they get treated at all.
A new study from Flinders University in Adelaide, Australia, published in npj Digital Medicine, suggests that single snapshot is often unreliable — and that it can lead doctors to misclassify patients who would look meaningfully different if measured for longer.
Comparing One Night to Three Months
Researchers led by Bastien Lechat, Ph.D., recruited 100 people who had been referred for suspected obstructive sleep apnea. Each participant underwent the standard diagnostic test — one night of in-lab polysomnography — and then took home a Withings Sleep Analyzer, an under-the-mattress sensor that unobtrusively tracked their breathing and sleep for up to three months.
The comparison let researchers see, for the first time in this population, how a single lab night stacked up against months of real-world sleep data collected with essentially zero effort from the patient.
The answer: not well, for a meaningful share of patients. Sleep apnea severity fluctuated substantially from night to night, and patients whose single lab test landed near a diagnostic threshold were the most likely to be misclassified — labeled with a different severity category, or missed altogether, compared with what months of home data revealed about their typical breathing patterns.
The Lab Itself May Be the Problem
One of the study's more striking observations was that misclassified patients tended to sleep worse in the lab than they did at home. Unfamiliar surroundings, wires, and the awareness of being monitored — a phenomenon sleep researchers sometimes call the "first-night effect" — appear to distort the very test meant to capture a person's typical breathing during sleep.
"Our findings show that sleep apnea can vary considerably from one night to the next, so if you only measure it once, you may not capture the true severity of the condition," Lechat said.
That distortion cuts both ways. Some patients tested as more severe in the lab than they actually are on a typical night at home; others tested as milder, potentially leaving real, clinically significant apnea undiagnosed and untreated.
Why the Stakes Are High
Misclassification isn't a minor technicality. A patient labeled "mild" may be told to monitor their symptoms rather than start treatment, while a patient labeled "moderate" or "severe" may be prescribed CPAP therapy, a device they'll be asked to use nightly for years. Getting that initial category wrong can mean either withholding treatment from someone who needs it or prescribing a burdensome therapy to someone who doesn't.
It also complicates research and drug trials that rely on a single AHI measurement to define who qualifies as having OSA in the first place — a measurement this study suggests may not reflect a person's underlying condition at all.
The Flinders team's findings echo a broader body of research showing that OSA severity is inherently variable rather than fixed. A related paper from the same group, published earlier this year, found that people whose apnea severity swings widely from night to night face independently higher cardiovascular risk — even when their average severity looks unremarkable. Taken together, the two studies suggest that both diagnosis and risk assessment built on one night of data may be missing something real.
Toward Low-Burden, Long-Term Monitoring
The researchers argue that the solution isn't necessarily more nights in a lab — which is expensive and impractical to scale — but low-burden, unobtrusive technology that can quietly monitor sleep at home over weeks or months. Devices like under-mattress sensors require no setup beyond an initial placement and no behavior change from the patient, making extended monitoring feasible in a way that repeated lab visits never could be.
The team, part of the FHMRI Sleep Health center at Flinders, described their approach as "zero burden" monitoring — a description meant to contrast with the disruption and cost of multiple lab-based sleep studies.
What This Means for Patients
If you were diagnosed with sleep apnea — or told you don't have it — based on a single night of testing, that result may be less definitive than it seemed. This is especially true if your test was borderline between severity categories, or if you know you tend to sleep poorly in unfamiliar settings.
Patients with symptoms that don't match their diagnosis — someone labeled "mild" who still feels persistently exhausted, or someone labeled apnea-free who still snores heavily and wakes gasping — may want to discuss extended or repeat monitoring with their doctor. As low-burden home devices become more widely available, asking whether a diagnosis was based on one night or many is likely to become a more routine part of sleep apnea care.
