Friday, June 18, 2010

Why an Average Can Be Misleading and How To Get Number Savvy: Vision Forum Myths About Ectopic Pregnancy as a Prototype

... Why I take issue with the misleading nature of the “1 in 60 million” rate of ectopic pregnancies that result in live births as quoted by Vision Forum and Samaritan Ministries.  

 Start HERE for some prerequisite ideas about how we can easily fall into "lying with statistics."

Start HERE for an overview of Vision Forum's aberrant stance on Ectopic Pregnancy

Briefly backing up…

Vision Forum and Samaritan Ministries cited a physician who stated that the chance of a live birth from an ectopic pregnancy was “60 million to one.”  I suppose that’s not Vision Forum’s fault for republishing this information, but I do fault Samaritan Ministries for disseminating the related information regarding healthcare statistics.  As an entity involved in the procuring and payment of healthcare, they have a higher burden of responsibility to communicate accurate information to their membership.

But how would they know that the doctor actually used a “SWAG” ( what is referred to informally as a “Scientific Wild A_ _ Guess,” a term that is even part of the court record in the State of Texas, for an associate accidentally used it while giving expert testimony!  The Judge stopped him and asked him to define the term.).  The doctor quoted in an interview for broadcast to the general public did not know that his words were going to be used to forge policy and procedure in treatment in ectopic pregnancy.  In a peer reviewed journal within his field, his “guestimate” would have been rejected.  And Samaritan Ministries is not run by healthcare professionals and must have no capable actuarians evaluating their publications, so they would not have been able to evaluate the veracity of any of these health statistics.  Yet, these statistics were used as a community health improvement measure and an educational resource for their membership.

I planned on posting more information here from Darrell Huff’s book, “How to Lie with Statistics,” but I think that I can finish this subject up in one (or two) posts through the example of the psychology of a statement like “one in a million.”  What does that mean?  As our national debt grows, for example, we are forced to comprehend bigger numbers in our everyday lives.  I hope that this little walk thorough simple statistical concepts will give you some tools to help you not fall into traps set by manipulators or even by well meaning people who are just ignorant of the significance of  how some of these things work, particularly in healthcare.

In my field of nursing and in medicine and disciplines like psychology, before anyone can claim that one event has an effect or even a relationship to another event, we are all held to a standard of first trying to disprove our data.  Statistics give us tools that help us sort out mere chance from true relationship by putting data to mathematical tests.  A doctor might see an unusual number of people with a certain disorder, and he might think that his patient sample constitutes all people.

By applying statistics, we can use math as a tool to help us pull away our own blinders or can give us validation if our suspicions are correct.  In psychology in particular, math helps us take something very subjective in nature and makes it as objective as possible by developing tools which are validated first, and if they prove that they yield stable data that is both precise and accurate (Does it really measure what we want to learn about?), that data can then be used to help tell us important and truthful facts about a particular problem, The same rules and principles are those applied in all research, from whether a new wound product speeds healing to whether the lab instrument in my husband’s office is operating properly by accurately detecting chemicals as evidenced by a clean linear regression curve to whether drugs get approved for use by the FDA,

When dealing with populations of people, it should always be noted that generalizations, to some degree, are never true for an individual.  (Ha! Though there are those exceptions to the rule, too.  People are unique about everything, God’s “non-normative” creations.)  But, we can also gain much wisdom from understanding groups, much in the way I believe that the Bible talks about seasons.  We know and trust things about the seasons, and wisdom helps us make decisions.  In a way, the application of math helps us discern wise action and can give us keen understanding.

For the sake of the audience, I've cut a few corners, so a statistician will get a bit irritated a couple of oversimplifications I make.  If there are any people out there that do take issue with me, please help me out and write a post that I can publish here!

 In the previous post, we mentioned the mean, mode, and median.  Building on that concept, I want to show you what those numbers usually look like when plotted on a graph through simple algebra.  Most things having to do with human beings end up falling into a “Bell Curve” pattern, and I’ve created one to illustrate all live births from pregnancy.  The math is simple, but plugging the right numbers into the right places and collecting the data itself can be trickier.  But here is what all live births look like on a bell curve.

Now, looking beyond there, lets put our “million to one” or any average (sum of the values of n divided by n) on our bell curve.  When someone says only “a million to one,” that does not mean that your unusual case is one in a million, because any random one in that million is the same.  The statement really doesn’t say much about the unique nature of the more uncommon examples of things.  A baby born from an ectopic pregnancy may be one out of a billion to one, but every other one in that billion is also “one.”  So these statements are misleading.  In healthcare or any research, it is unethical to use these kinds of rates as probability statements in the way Vision Forum has done for the purpose of advising people regarding their health, because the number has not been scrutinized.  I’d actually planned to run a statistical test to show the rarity of ectopic pregnancy births, but I can’t even obtain numbers to be able to make the roughest of estimates in terms of proper statistics.

 Compare the value for the random mean or average and where it falls on the curve for all live births with where live births from ectopic births fall on that same graph.  One in a million is ANY ONE in that million, and when evaluating the rarity of these cases, it is highly unethical to use such a measure to make a meaningful statement about the rare case. We already learned in the previous post that means (averages) are misleading.  If I had turned in something like this (what Samaritan Ministries disseminated) in a paper in Nursing Research as an undergraduate, I would have failed the assignment.  But I believe that Vision Forum and Samaritan Ministries have misapplied statistics to their advantage to give trusting people a false impression about the safety of their recommendations regarding ectopic pregnancy.

Well, it looks like at least one more post is called for after all….

A little more to come.