When Is A Cluster a Cluster?

[mars]

Hi All

Sometimes I get what looks like a cluster of events on my waveform report. That is 2 or 3 or 4 events within about a five minute period. They can be a mixture of Ai and Hi.

However, when they have occurred on a night I am using my oximeter, my blood oxygen level does not go below 90%.

So my question is - when we talk about clusters are we just talking about the number of events within a time period, or should we also be taking into account their actual effect on our blood oxygen level.

cheers

Mars

[ozij]

IMO Blood oxygen saturation should not be taken into account when apnea clusters are described by users viewing their software reports. Many users do not have pulse oximeters, and do have groups of breathing distrubances coming fast and close together, and causing a rise in pressure (or not). They should be able to discuss those clusters without referring to oxygenation.

The case is different if a person has a pulse oximeter, since the oximeter lets you see the effect a cluster has on your oxygen saturation.

O.

[schwawi]

Hi All

Sometimes I get what looks like a cluster of events on my waveform report. That is 2 or 3 or 4 events within about a five minute period. They can be a mixture of Ai and Hi.

However, when they have occurred on a night I am using my oximeter, my blood oxygen level does not go below 90%.

So my question is - when we talk about clusters are we just talking about the number of events within a time period, or should we also be taking into account their actual effect on our blood oxygen level.

cheers

Mars

I track my ODI using pulse oximiter too. I guess my feeling is if the "event" doesnt cause meaningful desaturation, then perhaps its not really an "event".... These CPAP machines all use different algorithms to calculate hypopneas and they do so based only on changes in air pressure (which is subject to leak distrurbance). I prefer the oximiter data myself....

[mars]

Hi All

IMO Blood oxygen saturation should not be taken into account when apnea clusters are described by users viewing their software reports. Many users do not have pulse oximeters, and do have groups of breathing distrubances coming fast and close together, and causing a rise in pressure (or not). They should be able to discuss those clusters without referring to oxygenation.

The case is different if a person has a pulse oximeter, since the oximeter lets you see the effect a cluster has on your oxygen saturation. O.

Thank you Ozij, and I agree and will go along with that.

I track my ODI using pulse oximiter too. I guess my feeling is if the "event" doesnt cause meaningful desaturation, then perhaps its not really an "event".... ... I prefer the oximiter data myself....

For personal use that is what I would tend to do, but still need to check - see below - question 2.

----------------------------------------------------------------------------------------------------------------------------------

Hi All

It just goes to show how far I still have to go in my recovery because I certainly was not clear in asking for the clarification that I was after.

So please bear with me and I will try again.

There are 220 occurences of the word cluster when I search the forum. Mostly self explanatory, and pointing to a wide range of opinion as to what comprises a cluster.

As my original post indicates, I am thinking that a cluster is 2 or more AHi in a 5 minute period, others are using a much longer time frame. The Covidien glossary does not help me on this.


When I look at my waveform analysis, say over a 7 hour period on one page (or screen), I can get what looks like clusters. But when I examine them more closely, say one hour to one page (or screen), then sometimes I can see that they are not what I would call a cluster. So identifying a cluster may depend on the ratio between events to time, and some one screen reports covering all night may show what appear to be clusters, but are not.

So the primary question I have is ------ is there an agreed minimum number of events that need to have happened within a specific period of time, to then be known as a cluster?

My other thought is ------- what if I say that a cluster is when one AH event occurs before I have recovered from the previous Ah event. Now comes the oximeter digression; but as I never go below 90% SpO2 on my machine, then it may be that clusters are not that important, unless desaturation from 97% to 90% is significant. Now let me make this clear - I am talking about clusters where desaturation under 90% does not, I repeat not, happen.

So my secondary question is--------------is a drop in SpO2 from 97% to 90% of significance. I gather that the sleep labs do not think so, nor do the makers of oximeter machines.

I hope I have made all this reasonably clear, and if you can help I would be grateful.

cheers

Mars

PS. My thinking says I have been too pedantic in this post, and if I have I apologise. But with my present level of brain power, it is the best I can do. But rest assured - I am getting better

And, I hope, so are you

[ozij]

There are 220 occurences of the word cluster when I search the forum. Mostly self explanatory, and pointing to a wide range of opinion as to what comprises a cluster.

Isn't that what a forum is about? Opinion?

I'm going to start a new subject -- in the hope Muffy at al. help us through. It will be about oxygen desaturations.

O.

[schwawi]

So the primary question I have is ------ is there an agreed minimum number of events that need to have happened within a specific period of time, to then be known as a cluster?

My other thought is ------- what if I say that a cluster is when one AH event occurs before I have recovered from the previous Ah event. Now comes the oximeter digression; but as I never go below 90% SpO2 on my machine, then it may be that clusters are not that important, unless desaturation from 97% to 90% is significant. Now let me make this clear - I am talking about clusters where desaturation under 90% does not, I repeat not, happen.

So my secondary question is--------------is a drop in SpO2 from 97% to 90% of significance. I gather that the sleep labs do not think so, nor do the makers of oximeter machines.

not sure on the definition of a cluster..... On desats, in my view, drop from 97 to 90 is significant, certainly almost as significant as 97 to 89, with 89 being the cutoff imposed by some on when desats really start to matter. In my opinion 97 to 90 qualifies as an event and I would personally view it the same as 97 to 89.

[Velbor]

I have spent much time trying to define a "cluster" and come up with meaningful analysis.

My most recent operational definition (based on ResMed's 1-minute reporting window) was that a "cluster" begins with at least 5 occurrences of an event within 10 minutes, and ends when there have been no instances for 10 minutes. This is quite arbitrary. Nonetheless, I found it essential to have SOME objective standard, particularly when trying to program an automated detection routine.

I was never able to come up with any meaningful analytic tools, other than the confirmation that, yes, I occasionally had clusters of events.

Given my failure to come up with any meaningful conclusions based on ResMed data, I also have not considered how I might construct a working definition based on Respironics' 30-second reporting window. I'm not sure whether I would use the same, or a modified definition. Frankly, in the absence of being able to figure out what to DO with the information, I haven't been pursuing this issue.