My Health Software, Support Forum » Averaging multiple readings
logo logo

« Back to forum topic index ... Post new topic


Support Forum Topic: Averaging multiple readings
Started by: HPG
Date: August 30, 2010

My apologies if this has been dealt with elsewhere in the forum; I’m new here. I’m just settling in to using MBP to record regular readings, and I’m impressed by the application. When I visit my GP, I’m put on a machine which takes 6 readings and discards the outlier (which is often the first one), so I’ve taken to doing that at home. I’ve set up a small spreadsheet that records a set of readings — I don’t usually do as many as 6, but enough to see the pattern — discards the outlier if there is one and averages the rest (selecting the outlier is done manually; I’m sure I could write something that would do it automatically, but it didn’t seem worth the effort). Does MBP have a facility to do this that I haven’t discovered yet? If not, is it a feature that could be built in at some stage?

bottom


4 Responses to “Averaging multiple readings”

  1. Steve says:

    Hi HPG,

    No, there is no facility to do this, however I would recommend just recording all the readings you take.

    There are tools in My Blood Pressure that will average multiple readings taken within minutes of each other. For example, click on the “View charts…” task, then click on the “Averages” tab, and then change the period to a “30 minute average”. Assuming you enter all readings, this will display a chart similar to what you would construct by averaging your six readings … but with the advantage that you are not throwing away any data points.

    Similarly you can see a list of these averaged readings by clicking on the “Study readings” task, then click on the “Averages History” tab, and then change the period to a “30 minute average”.

    If not, is it a feature that could be built in at some stage?

    I have thought about doing this in the past and it is still something that I may add in a future version. I think that what has stopped me in the past is the concept that it is better to retain all data and then work to make that more accessible, than the alternative of throwing data points out.

    Hope this helps!
    Steve

  2. HPG says:

    Hi Steve

    I see what you’re suggesting; it’s an interesting point (the one about throwing data away). I hadn’t noticed the half-hour averaging feature, and that’s a nice one. However, you don’t get the half-hour averaging when you’re looking at other charts, do you? Or have I missed something?

    I guess it boils down to the question of what are data and what’s information. Data are individual readings, and what we’re interested in is the information they give us about bp. If the variation in readings is genuine — one’s bp is bouncing up and down like a yo-yo — then that’s information that might, perhaps, be useful. If, on the other hand, it’s just measurement error, then it isn’t useful. I don’t know anything like enough about bp to know which is the more appropriate answer, although I’m guessing that the fact that at least some professional machines take half a dozen readings, toss out the outlier if there is one, and record the average of the rest, suggests that someone thinks that reading-to-reading variation over a short space of time is more to do with measurement error, or at least that the variation is less useful to know about than the average.

    Is there some kind of compromise that’s possible without a complete rewrite of the software? For example, is it possible to record all the readings one takes, mark them as members of the same set of readings, and then allow the user to mark obvious outliers to be ignored (my GP, for example, has told me to ignore the first reading of the set, on the grounds that this one is often the outlier and therefore unreliable)? Then give the user the option for the purposes of charting and other analyses either to use all of the data, or just the averages of each data set ignoring outliers? That gives you the best of both worlds; except, of course, for all I know it might be a nightmare to code.

    Cheers
    H

  3. Steve says:

    Hi HPG,

    However, you don’t get the half-hour averaging when you’re looking at other charts, do you?

    No, although if you click on the “Print report…” task while viewing that chart, that is the chart that will be printed in the report.

    If the variation in readings is genuine — one’s bp is bouncing up and down like a yo-yo

    I am not a doctor, but I believe it does bounce around a bit. Mine does Smiley … but maybe that is my monitor!

    For example, is it possible to record all the readings one takes, mark them as members of the same set of readings, and then allow the user to mark obvious outliers to be ignored

    If you enter all the readings, you could then manually delete the ones you consider to be outliers. To do this, click on the “View readings…” task, then select the outlier(s) by clicking on them, and then click on the “Remove reading…” task.

    … either to use all of the data, or just the averages of each data set ignoring outliers?

    This is an interesting idea. Maybe you could select the readings you want averaged (in the “List of readings” view), and then have a popup menu command that would replace the selected readings with their average. I like this idea! It wouldn’t have any effect on existing use, but would be useful for what you are after. I will add it to the todo list for the next version (not sure as yet when that will be).

    Thanks for prompting me on this!
    Steve

  4. HPG says:

    I am not a doctor, but I believe it does bounce around a bit. Mine does  Smiley … but maybe that is my monitor!

    Yep — that’s the measurement error I was speculating about Smiley

    If you enter all the readings, you could then manually delete the ones you consider to be outliers.

    Yes, you could certainly do that; it would be a whole lot easier than trying to come up with a valid algorithm.

    Maybe you could select the readings you want averaged (in the “List of readings” view), and then have a popup menu command that would replace the selected readings with their average.

    Nice idea — that would do it very simply, wouldn’t it, without having to mess around with the database structure.

    Cheers
    H

bottom ForumAdmin