How to Read Your CPAP Data Without Driving Yourself Mad
AHI, leak rate, usage hours, event types. What each number on your CPAP app means, what's normal variation, and which numbers warrant a phone call.
Your machine records what happened last night and shows it to you in an app. That’s genuinely useful — most patients in medicine never get to see their own therapy data — and it’s also a fast route to anxiety if nobody explains what the numbers mean.
Here’s how to read them.
Usage hours
The simplest number and the one insurance cares about. Total time you were on therapy.
Medicare’s adherence standard is “use of PAP ≥4 hours per night on 70% of nights during a consecutive thirty (30) day period anytime during the first three (3) months of initial usage”1 , so this is the number that keeps your machine paid for. More on that in CPAP compliance rules.
Clinically, more is simply better. Four hours is an insurance threshold, not a health target — the apnea comes back the moment the mask comes off, so the last three hours of the night are as untreated as they ever were.
AHI
The headline number: estimated apnea-hypopnea events per hour.
Most clinicians want to see it under 5 on therapy, and plenty of well-treated people run between 0.5 and 2. Full explanation of the number itself in what is a good AHI.
Three things to hold in mind:
It’s an estimate. A sleep study measures airflow, chest and abdominal effort, oxygen, and brain activity. Your machine has one sensor — flow through the hose — and infers the rest.
It usually can’t tell if you’re asleep. Lie awake for an hour with the mask on and that hour typically lands in the denominator, quietly lowering your reported AHI.
It varies enormously night to night. Alcohol, back sleeping, a cold, a worn cushion, or a stressful day can double it. One bad night means nothing. A two-week upward trend means something.
Leak
Air escaping that shouldn’t be. Reported in litres per minute, usually as an average and a 95th-percentile figure.
Read this before you read your AHI. High leak makes everything else unreliable — the machine can’t accurately detect events through a wall of escaping air, and it may be delivering less pressure than prescribed. An AHI of 1.5 with terrible leak is not the reassurance it looks like.
Every mask has intentional vent flow built into the number, so the raw figure differs by mask. Most manufacturer apps flag when leak exceeds the threshold for your specific setup — use that flag rather than a number from a forum.
If yours is high: how to fix mask leaks.
Event types
Most machines break events into categories — obstructive apnea, hypopnea, central apnea, and sometimes “unknown” or clear-airway events.
Obstructive and hypopnea are the expected ones: your airway is narrowing or closing.
Central events mean the machine detected no effort to breathe. The NIH describes central sleep apnea as a problem with “the way your brain controls your breathing while you sleep,” associated with heart failure, stroke, and opioid use2 .
Here’s the calibration: a small number of centrals is common and often unremarkable, particularly as you’re falling asleep or during arousals. A consistently high central count, or a central count that has climbed since therapy started, is worth showing your doctor rather than researching further. The classification is also algorithmic and manufacturers differ, so it’s a flag rather than a diagnosis.
Pressure
If you’re on an auto-adjusting machine you’ll see the pressure range it used — usually a median and a 95th percentile.
Useful because it tells you where you’re actually living. If your machine has a range of 6–16 and spends the night at 15.8, it may be hitting the ceiling and wanting more. If it never leaves 7, the top of your range is doing nothing. Either observation is a good thing to bring to your clinician; neither is a reason to change the setting yourself.
The apps
myAir (ResMed) and DreamMapper (Philips) are the consumer apps — friendly, simplified, and score-based. Good for tracking usage and spotting obvious problems.
OSCAR is free desktop software that reads the SD card and shows detailed flow waveforms. SleepHQ does something similar in a browser. Both show far more than the consumer apps, including breath-by-breath flow shapes.
These detailed tools are genuinely powerful and they come with a warning: it is very easy to spend months analysing flow limitation graphs and convince yourself of a problem that isn’t there. If detailed data helps you have better conversations with your clinician, use it. If it’s fuelling 2am anxiety about your own breathing, close the laptop — that’s a real pattern and it makes sleep worse, not better.
When to actually call someone
- Your AHI has trended upward over two or more weeks.
- Leak is consistently flagged and you can’t resolve it.
- Central events have appeared or increased noticeably.
- Your numbers look good and you still feel terrible.
That last one matters most. A good AHI is not a reason for anyone to dismiss how you feel — it means the obvious explanation has been ruled out and something else deserves a look.
Common questions
Why is my CPAP AHI different from my sleep study AHI?
What is a normal CPAP leak rate?
Should I worry about central apneas on my CPAP report?
Is OSCAR worth using?
How often should I check my CPAP data?
References
Every clinical claim above links to one of these. Where the evidence is mixed or thin, we say so in the text rather than picking the source that sounds most confident.