Understanding Your Bone Mineral Density Results.
From dAISy, March 2002.
By Andie Hider
I was thinking of calling this article “Statistics For Dummies”, but two things occurred to me; one was that the publishers of a certain series of computer books of similar title might object, the second was that I have yet to meet someone with AIS or any similar condition that is a dummy.
The idea of this quick article is not to be a comprehensive lesson about statistical information and it’s use (or misuse), but rather something brief that will allow people in the support group some better understanding of their Bone Mineral Density (BMD) results.
We will start with the assumption that you have organised to have your BMD checked through your GP or an appropriate specialist. You will (of course!) have arranged to have this done on the same machine in the same clinic you did last time, to ensure that the results have a consistent base line, that is, you are comparing apples with apples. There can be quite a degree of variation of results for the same person if the tests are conducted on different machines. It is not always possible to ensure you have all tests done on the same machine, but where you can it is worth trying to do so. Generally speaking, where testing for BMD is medically indicated (such as in the case of AIS) I would advise having it checked once a year. Some specialists would agree with this, some would say every two years but I personally think two years is too long between tests. A lot can happen in two years if things are not as they should be. For those that have yet to have their BMD tested, the process is painless and just like an X-Ray that takes a bit longer to finish. In the case of AIS, where BMD issues are medically indicated, you are entitled to a Medicare/Private insurance rebate for most of the cost of the procedure.
You (or your GP/specialist) will now have a document that gives you the results of the BMD test and the fun starts. The results are commonly expressed in one of two ways, either as a percentage of the mean or as standard deviations above or below the mean.
The mean is simply the average for a particular group, in the case of BMD it will be somewhere on a scale between the worst BMD result for a particular group tested and the best BMD result for that particular group. It is not necessarily the case that it will be half way between, if there are more results of tests near the top of the scale then the mean (average) will be closer to the top of the scale. Most commonly in the case of BMD, gender and age groups are the groups used to categorise results, as generally speaking there is a consistent age/gender related pattern to the development of peak BMD and BMD decline later in life. Your results will be compared with women of the same age if you are a woman and with men of the same age if you are a man. So far so good.
If your results are expressed as a percentage, then you will get something like 90%, 100%, 110%. Right now you are thinking “she is full of it, you can’t have 110% of anything”. Mostly that is true, but we are talking a percentage of an average here. You can have above average BMD for someone of your age and gender, so in the case of 110% this simply means your BMD is about 10% above average for your age and gender group.
Standard deviations are a little more complex, in fact I don’t know why certain testing laboratories persist in using standard deviations to record BMD results. When we gather information about a particular population the results form a scale or range either side of the mean (average). Some results will be above the mean, some will be the same as the mean and some will be below the mean. If we divide the “above” range and the “below” range into thirds, each third is what is called a standard deviation and of course we have six of them, three in the range above the mean and three in the range below the mean.
Just to confuse the issue there is not a sixth of the results in each standard deviation. Because most people will be near the mean (whether that be average height, weight or BMD), the further you get away from the mean the less people there are that have those results. Generally speaking, 66% of the population will be within one standard deviation below the mean and one standard deviation above the mean (33% below, 33% above). About 98% of the population will be within two standard deviations below the mean and two standard deviations above the mean (49% below, 49% above). All this means is that most people will either be within the first standard deviation above the mean and the first standard deviation below the mean, most of the rest in the second standard deviation above the mean or the second standard deviation below the mean and a very few people will be in the third standard deviations, the last one above and the last one below.
When your doctor tells you that your results are half a standard deviation above the mean, this simply means somewhere above average in the range most others in the population score. Likewise, if you are told you are one and a half standard deviations below the mean, then you are outside the most common range of results, but still within the results of 98 percent of the population as a whole. If your results are anywhere outside the range of two standard deviations above or below the mean, then you are in a very small two percent of people with those results, but still within results recorded for the whole population.
One quick comment about low BMD results (there is more detail later for those that are interested or inclined). Statistical information is gathered in such a way that results expressed as standard deviations can be a bit misleading. If you fall into the category of being two or even three standard deviations below the norm, it is not as low as you think. There is still reason to take steps to try and maintain your BMD though because as the BMD results of the general population get lower, yours will be getting closer to the average for your age/gender group.
All of this should get you by as a rough guide to interpreting your results, for those wanting to further refine their understanding, read on.
When statistical information about any topic is gathered using a reliable technique, there will always be results that don’t seem to make sense because they are so different to most of the other results. This may be because of the way the information has been gathered, it may be because in some cases there is no way to consistently get certain types of information or it may be because the information came as a mistake from a group outside that being studied. Were these very different results to be included, they would make the other results obtained a less than accurate representation of the true picture. It is common practice to take out a percentage of the results that are very different to the rest. This leaves us with a range of results above the mean and a range of results below the mean that should be fairly accurate.
What this does mean though is that the population I mentioned earlier is not a complete record of all samples or answers given, but rather a sample population based on results or answers that fall within what are believed to be the most accurate range of answers. This has the effect of “normalising” the sample population, in practice of course there is no such thing as a “normal” population.
As a case in point, there is much anecdotal evidence about BMD problems in those with AIS, sometimes with BMD results that are within or even below that third standard deviation below the mean. Because statistical information is most often gathered in the way I have described above, scores such as those that might be present in AIS might be discounted because they are considered unreliable. In fact, were it understood that these scores have a reasonable basis, they might be included in the overall population. This would have the effect of moving everybody up the scale, perhaps from one and a half standard deviations below the mean, to one standard deviation below the mean (and hence into that most of the population range) or from the third to the second standard deviation below the mean and into scores that still represent most of the population. Ideally there would be a set of BMD results using a group of people with AIS as a test population so that accurate comparisons could be made, since no such results exist to my knowledge be careful to take this into account when you interpret your results.
As explained earlier, this doesn’t mean you should be complacent about a low BMD result though, exercise and HRT have been shown to benefit BMD and Fosomax is another alternative suitable in some cases. As you get older, if you maintain your BMD at the same level, your BMD will get closer to the average BMD range of the general population. Ideally your BMD should fall within that average range of one standard deviation above or below the norm, but again make sure you remember how these standard deviations are arrived at.
Statistical information is a universally accepted way of recording results, but it does have the potential to be misunderstood. For those that are interested in learning more about statistics, I can arrange to send out “plain language” statistics explanations that are very good.
with
questions or comments about this web site.