Cyclic AHI values?

[Guest]

Several posters have commented that thier AHI values bounce all over the place for no apparent reason, as have mine since I started 119 days ago. So I thought I'd graph the daily AND the 7 day average values that Sleepy Head provides to see what I got.

Findings:
During that 119 days logged my daily AHI varied from a high of 1.67 on day 7 to a couple of 0.00s and the line for the 7 day running average had dips at 23, 36, 56, 71, 87, and 104 days. Disregarding the initial 23 days because I was acclimating to the machine, I've got dips on 13, 20, 15, 16 and 17 day intervals or about a 16 day cycle. (if someone could help me I'd post the graph)

I don't know what this means but 119 days isn't a such long time and what I'm seeing could be just a result of the short data base or it's just my bodies personal rhythm. It would be interesting (at least to me) if others with longer data sets would do the same and post their results. Setting up an Excel or Works Spreadsheet to do this is pretty easy, but manually entering each days AHI value going back in time does get a little tedious.

I think it would make an interesting potentailly useful little crowd sourced research project unique to cpapatalk. If periodic cycles are found to be a characteristic of our groups fluctuating daily AHI values it might be beneficial to those who are just getting started with xPAP therapy by helping them better understand what is happening to them.

[JDS74]

Hi Guest.

It would be helpful if you registered and all that stuff so a real conversation could take place on this topic.
Having said that, I think it is a really nice idea to look at the data in that fashion to see if there are any cyclic variations in the data.

I will look at some of my data ( I have about 1,100 days of data stored ) and will see what could be used.
A problem that might come up and make data analysis problematic is the variation in the presssure numbers that some of us have made on their machines trying to get the therapy right.

With that in mind, I probably have two sets of data that might be useful - one of about 500 days and the other about 100 days. The rest of the data is contaminated by pressure setting changes and lack of keeping a single pressure setting for more than 60 to 90 days, a period too short to show up cycles of 15 days or longer very well.

[Jay Aitchsee]

Guest and JDS,

Thought you might find this interesting. It's my last year of data offered without analysis.

Jay

[Black Shampoo]

My AHI is alllll over the place with almost zero change in lifestyle or behavior. I'll literally have a 3.5 one night, a 1.8 the next, a 0.2, a 3.0, etc. without any obvious reason why. There are so many factors that go into it.

[rkuntz]

Hope this works. If it does the first series in Blue are my daily AHI values, the second series in Red is my running 7 day AHI average.

[Jay Aitchsee]

rkuntz, I do like smoothing the data with a rolling average. Here my chart from above with the 7 day rolling average superimposed. Now, it does seem there is some periodicity to the data, but as I said above, I haven't analyzed it. I will say that the spike in AHI in early July is a short night on my travel machine while staying in a hotel.

[rkuntz]

Jay,
Can you pick a periodicity out of the smoothed data? The screen resolution of what you've shown isn't fine enough to read.
Ray

[rkuntz]

Jay,
Incidentally how did you overlay your 7 day average on that chart? That was pretty slick, I did the hard way by manually entering data from SH into Exel.
Ray

[Jay Aitchsee]

Ray,
Well I tried to give us another look, but the images are getting to large to preserve the definition. Anyway, in this graphic I have the original 7 day rolling average (top) broken into the 1st and second half of the year data points. In the 1st half, I could make a case for an approximate 20 day cycle (+ or -). In the second half, not so easily. I'll have to expand it a little and have another look. One thing, comparing your results and mine, AHI does appear cyclical, but the cycle probably varies considerably.
In any event, I think the results do give credence to the advice that people need to let conditions stabilize before judging the effects of any changes made to influence AHI. Like a change in pressure, for example.

Incidentally how did you overlay your 7 day average on that chart? That was pretty slick, I did the hard way by manually entering data from SH into Exel.

To produce the graphs, I exported my data from ResScan as a CSV file into an Open Office spreadsheet (think free Excel) and then used the charting wizard to create the graphs. To produce the 7 day data, I created another column next to my AHI data column and, as the first entry in that column (seven rows down), I used the "average" function to average the first 7 days of AHI. Then I just used the "fill down" function to complete the column. Once the column was created I used the charting wizard to add that data range to the chart I had previously made.

here's a little better look at the 1st 171 days

[rkuntz]

Unless it's hiding in plain sight, neither Encore Viewer or Sleepy Head the two pieces of software that I have to read my Respironics data don't have a data Export function.

I set Excel up exactly as you did in Open Office now I just have to add the data to the daily AHI column manually each morning. Back filling my daily AHIs a hundred days when I started took an hour or so.

[Jay Aitchsee]

I don't own a Resperonics or use SleepyHead, but there are a few posts that imply SleepyHead has an export function. Here is one: viewtopic/p872977/viewtopic.php?f=1&t=8 ... rt#p765157 and here the developer mentions a "CSV export" function: viewtopic.php?f=1&t=94442&p=872684#p872684