What Is Sleep Apnea Testing? Home Sleep Test vs. In-Lab PSG, Compared Honestly

Since 2016, I've coordinated a sleep diagnostics program for a multi-site clinic group. In nine years, I've made—and documented—23 significant mistakes that totaled roughly $60,000 in wasted testing, rework, and device purchases. Not proud of it. But that's exactly why I now maintain our team's pre-order checklist, and why I can give you a comparison the sales brochures tend to skip.

What is sleep apnea testing, exactly?

Before comparing anything, let's define the term properly. Sleep apnea testing is the objective measurement of breathing, oxygen levels, and sleep-related arousal during sleep. The key word is objective. It is not a questionnaire. It is not a quick office assessment. And it is definitely not a blood test.

That last point trips up more clinicians than you'd expect. A thyroid panel, HbA1c, ferritin, or metabolic screen run on the clinic's clinical chemistry analyzer is a reasonable part of a fatigue workup. I've made the mistake myself: normal labs plus a normal overnight pulse oximetry reading, and I assumed sleep apnea was ruled out. It wasn't. A clinical chemistry analyzer can tell you a lot about metabolism and endocrine function, but it cannot tell you whether the upper airway collapses during sleep. No chemistry analyzer on the market measures apneas followed by brain arousal.

So when someone asks me what sleep apnea testing involves in 2025, the honest answer comes down to two practical routes:

• In-lab polysomnography (PSG)—the full overnight sleep study, attended by a technologist, with EEG, airflow, effort, oxygen, and usually video or audio.
• Home sleep testing (HST)—a limited, unattended device the patient wears in their own bed, typically tracking airflow, respiratory effort, oxygen saturation, and pulse.

I'll compare them across four dimensions: diagnostic depth, real-world completion, true total cost, and what the result means for the treatment you prescribe afterward.

1. Diagnostic depth: what each test can prove

There's a reason in-lab PSG is still called the gold standard. It records brain waves, so the technologist knows when the patient is actually asleep. It can stage sleep, detect leg movements, and identify events that end in arousal rather than a visible oxygen drop. Home sleep tests can't do most of that.

The clinical consequence is bigger than it sounds. Without EEG, an HST doesn't know when you're asleep. It estimates the apnea-hypopnea index (AHI) using monitoring time instead of sleep time. That denominator alone can make a real problem look mild. And hypopneas that cause arousal without significant desaturation? Often missed entirely.

It's tempting to think a home sleep test is just a lighter version of a lab study. It's not. It's a screening test with a different question.

I've seen the fallout more than once. In 2021, a middle-aged patient with loud snoring and witnessed apneas had a home sleep test that came back "negative." I took the result at face value. He didn't feel better, so we eventually did an in-lab PSG. It showed an AHI of 18 per hour, mostly driven by respiratory-event-related arousals that the home device couldn't detect. We lost nearly a year of treatment time, and I lost some credibility with a referring physician.

That said, I'm not trying to bury HST. For a patient with high pre-test probability of moderate-to-severe obstructive sleep apnea, a clearly positive home test is often enough to start CPAP therapy without a lab visit. The limitation shows up in the gray zone: mild results, negative results in symptomatic patients, and anyone whose story suggests something beyond simple OSA.

Conclusion of this dimension: If the question is "is the airway collapsing during sleep?" a good HST can answer yes. If the question is "what exactly is waking this patient up, and how many times per hour?" the PSG is the only reliable option.

2. Real-world completion: does the patient actually produce a usable test?

Theoretically, patients sleep better at home. And many do. I've had patients who couldn't sleep a wink in the lab produce perfectly clean home studies. So the home test isn't automatically worse on patient comfort.

But comfort and diagnostic quality are different things.

In 2023, our center tracked our first 100 home sleep tests. Fourteen came back technically inadequate: disconnected airflow cannulas, signal loss overnight, or less than two hours of usable data. We caught some and repeated them. More concerning were the eight "negative" reports where the signal quality was technically acceptable but the patient clearly slept poorly and barely changed position.

That's the part nobody puts in the brochure. At home, no technologist is watching. If the sensor slips off at 2 a.m., the test becomes expensive noise. In the lab, a trained technologist can fix the sensor, wake the patient to reposition, and ensure the night actually counts for something. The lab isn't perfect—first-night effect is real, and some patients sleep terribly in a strange bed—but a disrupted lab study is still usually scored and interpreted by someone who knows what to do with it.

We now have a screening checklist before we ship any home test. Are the patient's hands steady enough to apply the sensors? Is there someone at home to help? Do they understand they need to keep the cannula on even when they roll onto their side? Since we introduced that checklist eighteen months ago, we've caught 47 potential failures before they became failed tests. Should have done it after the first dozen.

Conclusion of this dimension: HST only works when the patient can produce a high-quality recording unsupervised. PSG has staff on hand to rescue the night. For elderly patients, patients with limited mobility, or anyone who can't manage the equipment, the lab is the more reliable choice—even if it's less comfortable.

3. What each test really costs, including the line nobody quotes

Here's where I've made some expensive mistakes.

Based on U.S. payer fee schedules and average clinic invoices from early 2025, a home sleep test typically runs between $150 and $600. An in-lab PSG usually runs between $1,100 and $3,000. (Verify current rates in your region—these are general reference ranges, not a quote.)

So HST looks dramatically cheaper. And for a clean, straightforward patient, it is.

But I learned the hard way that the sticker price isn't the final price. In 2017, I built a budget assuming every home test would succeed and produce a billable diagnosis. That assumption was wrong.

A failed home test still costs the device fee, the shipping, and the clinical coordinator's time to reschedule. A negative home test in a patient with high pre-test probability usually leads to a follow-up PSG anyway—which means the HST wasn't a cheaper alternative, it was an additional expense. And some of the cheaper per-test vendor quotes didn't include physician interpretation, repeat sensor kits, or the extra call time when a confused patient can't figure out the device at 10 p.m.

A colleague taught me a phrase I now use in every vendor conversation: ask what's NOT included before you ask what's the price.

The vendor who lists all fees upfront, even when the total looks higher at first, almost always costs less in the end. The low price that appears after rebates and add-ons? That's usually the expensive one.

Conclusion of this dimension: At face value, HST is cheaper. In total cost per completed, actionable diagnosis, it depends entirely on your patient population. For low-risk patients, HST wins. For complex patients, repeated home tests are the most expensive way to postpone a PSG.

4. From test result to treatment: CPAP, BiPAP, and the device pitfalls I've documented

The testing conversation doesn't end at diagnosis. The result determines the treatment pathway, and that's where procurement mistakes happen.

For most uncomplicated obstructive sleep apnea, an auto-adjusting CPAP is the appropriate first-line device. In our network, that usually means the ResMed AirSense 10 or AirSense 11 family. They're reliable, well-supported, and patients know the names.

But some patients aren't candidates for standard CPAP. If a patient needs a relatively high pressure, has difficulty exhaling against pressure, or has OSA with certain respiratory comorbidities, a BiPAP machine may be medically necessary. BiPAP delivers two distinct pressures—higher during inhalation, lower during exhalation. It is not a "premium comfort upgrade" from CPAP, and it shouldn't be sold as one. The choice is clinical, not a price tier.

Once the right device category is chosen, the next most common errors I see are configuration mistakes. Two practical examples from our own documented failures:

ResMed AirSense 11 temperature settings. The AirSense 11 has a heated tube option and adjustable climate settings. If you attach a standard non-heated tube and leave the temperature or humidity settings configured for a heated tube, you get condensation. Not a malfunction—rainout. Water droplets form in the tube and can literally bubble into the mask. One of our patients woke up to water running across their face and stopped using therapy for three weeks. The device wasn't broken. The tube type and the temperature settings simply didn't match. We now confirm which tube is in the box before changing any AirSense 11 temperature settings.

ResMed AirSense 10 error codes. When an AirSense 10 displays a system error, the first response should be simple: tell the patient to unplug the device, wait 30 seconds, and plug it back in. That recovers a surprising number of units. If the same error returns within a week or two, the device needs manufacturer service—not a clinic-side workaround. I've seen a well-meaning biomed tech delay that service call by trying to reset a unit four times, which only stretched a two-week repair into six. Error codes on AirSense 10 devices are signals, not suggestions. Log them, follow the manual, and escalate early.

Conclusion of this dimension: A diagnostic test that doesn't lead to the right therapy is just an expensive piece of paper. Make sure your treatment pathway—CPAP versus BiPAP, heated tube versus standard, home therapy versus closer follow-up—is defined before you order the test, not after.

So which approach should a sleep program actually choose?

If you came here looking for a simple "HST is better" or "PSG is better," I'm going to disappoint you. The answer depends on who's walking through your door.

For a program that serves mostly straightforward, high-pre-test-probability patients—loud snoring, observed apneas, excessive daytime sleepiness, no significant heart or lung disease—a home sleep test-first pathway is logical. You can diagnose fast, start auto-CPAP quickly, and manage patients remotely. It's efficient, patient-friendly, and genuinely lower cost when the population is selected well.

For a program that sees unexplained daytime sleepiness, suspected narcolepsy or periodic limb movement disorder, patients with heart failure or COPD, or anyone whose story doesn't fit cleanly into obstructive sleep apnea, the in-lab PSG is not optional. It's the standard. According to the American Academy of Sleep Medicine (AASM, aasm.org), in-lab PSG remains the standard for diagnosing OSA, while home sleep tests are appropriate for patients with high pre-test probability of moderate-to-severe disease and no significant comorbid conditions.

Most well-run programs eventually land on a stepped protocol rather than one test: HST first for the straightforward group, PSG for patients who fail, who test negative but still fit the clinical picture, or who have complex presentations. If we'd adopted that protocol in 2017 instead of trying to make one type of testing work for everyone, we'd have saved roughly $26,000 of the $60,000 I mentioned at the start.

And when you get to the treatment side, don't let the device be an afterthought. Confirm the machine matches the patient's physiology—CPAP when it's straightforward, BiPAP when the clinical picture requires it. Confirm the settings match the accessories. Know the platform's error codes and temperature controls before you deploy a fleet, not after your first cold call from an angry patient.

Testing and therapy are two halves of the same decision. Choose the test that actually answers the question, and choose the treatment that follows from the answer.