What is the value of knowing my AHI numbers?

[Velbor]

Reading posts and chat comments, it strikes me that some discussion of the significance of AHI numbers might be useful. [Some of my comments will be based on specific experience with ResMed data management routines. I assume that other manufacturers use similar conventions.]

From the start, it seems that many of us don't really know (and probably don't really need to know) precisely what the index numbers represent. [Manufacturers sometimes don't make it easy to get their documentation to find out, but that's another issue.] An apnea, for example, to ResMed, is "defined as a greater than 75% decrease in ventilation. The ... algorithm scores an apnea if the 2-second moving average ventilation drops below 25% of the recent time average (time constant 100 seconds) for at least 10 consecutive seconds." Sure, it's obvious exactly what that means!

Now, AHI numbers ARE important. Big numbers are bad, small numbers are good. Numbers getting lower over time are good. Also, how you feel is important. But should you say, "If I feel good, to heck with the numbers." Well, maybe, but only if you assume that daytime sleepiness is the only adverse consequence of sleep apnea.

Focusing on a number as though it completely reflected your condition is a bit simplistic. Let's say I have an AI of 2. Good, or bad? All it tells me is that I have had (on some one night, or as an average over several nights), overall, two "apnea events" (whatever they are) per hour. If I'm on the machine for 8 hours, I've had 16 events. Good, or bad?

How long did each of those events last? The AI doesn't tell. Is the length important? Each event could have been anything from 10 seconds (the minimum to be classified as an apnea) to 60 seconds (each minute is treated as a distinct "data chunk"). So with my AI of 2 I might have been "in apnea" anywhere from 160 seconds (16 events x 10 seconds, or 2m 40s, or 0.56% of the time) to 960 seconds (16 events x 60 seconds, or 16 minutes, or 3.33% of the time).

[By the way, the "one-minute-data-chunk" has a number of other implications: multiple events in the same minute, one event split into two, or an event not recorded at all - but that's an issue is for another topic, as is the way ResMed truncates, rather than rounds, its AutoScan display of percent-time-in-apnea.]

How were these events distributed through the night? The AI doesn't tell. Does it matter whether my AI of 2 means that I had one event every half hour throughout the night, or that I had all 16 events during a single quarter hour period? Which is better or worse? Spread out over the night, interference with sleep may be greater. Getting them all over with at once, I might sleep better, but the likelihood of significant oxygen desaturation is greater. Is sleep deprivation, or oxygen deprivation, the greater concern? Does it matter if most events are early, or late, in the night?

Similar comments may be made regarding "numbers" for leak (and for pressure if it's variable). The meaning of 95th (or 90th) centile does need to be understood carefully. Whether the 95th centile value, or the median value, is more useful or informative or important, can be debated. You certainly can't have a clear picture of what's happening without BOTH statistics.

My bottom line here is to agree with the often repeated view that data access is essential. But "data access" as meaning reading numbers off the machine's LCD screen is only minimally helpful (better than nothing?), and in fact can be misleading (worse than nothing?) if viewed alone and out of context. Knowing what is happening throughout the night, and from one night to another, is the key to understanding and dealing with our sleep apnea. Access (ideally by the user, as well as by medical support personnel) to software which can display graphically what is actually happening throughout the night is much more helpful than simply reading off of individual numbers. [If a picture is worth a thousand words, is a graph worth a thousand numbers?] And even more important is having someone (ideally the user, as well as knowledgeable medical support personnel) who takes the time to LOOK and to UNDERSTAND and to CARE about what the picture is telling us.

Thank you for your patience. The floor is open for comment.

[Wulfman]

Let me try to draw an analogy using being in school.

For the person with a non-data-recording machine, it would be like taking tests every day in school and not knowing how one did on those tests.

For the person with a data-recording machine, getting the nightly summary numbers off of the LCD would be like getting the scores back on the tests, but not knowing which answers were wrong.

For the person with a data-recording machine and having the software to see the events, that's comparable to getting the daily tests back and knowing what answers were answered incorrectly. If you make some adjustments to the machine, that's like doing some homework on the ones you got wrong.

With the REMstar machines that can use MyEncore or Encore Pro Analyzer, the approximate length of the apneas are recorded.

I do not believe the manufacturers should make ANY machine that doesn't record detailed statistics. And, the users SHOULD know, care and be aware of the capabilities and importance of the information being collected.


Den

[hades161]

In my opinion they need to work harder on getting even MORE data to us and making the data MORE accurate. So that maybe Doctor's will take the information seriously when we go in to them. Why is data important? that's pretty simple to answer.

Without data, how can you tell if treatment is working fully or where the problem is if it's not? Sleeping, is an everyday thing with tons of varibles. If your one of the lucky ones that got an accurate PSG and titration, you might get hooked up on a Xpap and be off as happy as can be, but a lot of the times that's just not so.

Even if it was that way for you but for some reason something changes in your life like lose/gain wieght, start/stop excersize, new meds or any number of other factors, it will throw off your pressure, and your treatment degrades. Then you will be MISTREATING your condition and have no way to tell until you feel like crap again or actually cause your health to be worse.

Diabetics have glucose meters for a reason, and in my mind we need accurate data for the same reason. To PROPERLY treat our condition on a daily basis without having to risk MISTREATMENT. As well as not running to the Doctor or hospital everytime something changes for a blood test or in our case a PSG.

Also it might be very helpfull to the doctor to be able to SEE whats going on over more time, under normal conditions. Then just a one night study in a strange place, on a stange bed, with all the wires being attached, and some stranger watching over you as you sleep in one of the most helpless position a human can be in.

If you have the cash to burn for $2000 PSG tests, 75$-150$ doctor visits, as well as the cost of whatever other tests they would run as a "guess" whenever you feel like it, kudos to you, but most don't have that option.

The only people that like this way of doing things are:

A. The sheeple, people who like sheep, wish to be herded.

B. The medical community for more profit.

C. The equipment suppliers for more profit.

D. People who have little to no understanding of OSA and it's effects on a person's life.

We need MORE data and MORE accurate DATA then less.

[jules]

I see too many people of the forum try to get feedback on the numbers. They don't tell the story. Having the software and looking at the nightly graphs is CRUCIAL. As much as numbers are nice, there is more that can be found from the "pictures".

Until the companies change their focus to serve the end consumers instead of the "professionals", there will be problems getting good software into the hands of the consumer. The programming in Encore Pro sucks. Round off errors are prevalent. The issues with the dropped data on the M series machine is appalling and nothing is being done to correct it from what I understand.

[kteague]

I agree with the value of having data to support treatment, and think nondata machines should be obsolete. With all their imperfections, I think data collecting machines have saved endless healthcare dollars for the consumers, and even saved lives when treatment was making some worse instead of better.

I would like to add one more category of patients. Let me pull up my soapbox. There is a segment of healthcare recipients who simply do not have the wherewithal to be intricately involved in their treatment. Some are too physically ill, emotionally drained, mentally ill, cognitively impaired, or intellectually limited. I've spent years being frustrated with a friend's attempts to do better at this, and she has tried. But I have come to understand that someone enduring the longstanding health crises she has endured should just be able to trust their physicians to just tell them what they need to do next. That's the real tragedy, when those who can't help themselves get piled on more stress by running them back and forth to sleep clinics or leaving them to their own devices. Giving these patients data capable machines that their physicians can monitor would be helpful.

My friend has to make regular followup visits to the oncologist. One round of chemo threw her into heart and kidney failure. She now has something going on in her chest wall - more visits. Three times a week to dialysis, which her body is not tolerating well. And her unsuccesful cpap treatment will require another visit to the sleep doc and maybe another study because all he can do is guess what's wrong. Makes me so mad I could spit. Some things in life are inevitable, but the way sleep medicine works at this point in time is in my books a crime.

Kathy (stepping off of soap box)