A little light weekend reading

While I work on winging my way back to the East Coast, I thought I'd leave you with a couple of links that I became aware of but didn't get the chance to post.

First up is the older piece by that tireless debunker of dubious medicine and quackery and fellow skeptic, Prometheus. In a piece entitled Mercurial Laboratories, he dissects in detail why the laboratories that purport to show parents that their autistic children have elevated mercury levels almost certainly do not do any such thing. Indeed, what he has written can apply to almost any set of lab tests, particularly this part:

One of the conundrums in clinical laboratory medicine is this 5% of the "normal" population that - by definition - is outside of the reference range. This is particularly problematic when you are running a battery or "panel" of tests, since each one of them has a reference range that excludes 5% of the "healthy, normal" population. It is a relatively simple exercise in probability to find how many tests you have to do to reach the point where there is a greater than 50% chance that one of the results will be "abnormal" (i.e. outside the reference range) in a person who is "healthy" and "normal" [Answer: 14].

To this, I would add that, if there are 20 or more tests, the probability of at least one abnormal value approaches 100%. This principle can be applied to research as well, and perhaps I'll expound on this in a future post.

The second post is by Kathleen Seidel and is entitled Strategic Disregard. One thing that I and others bring up about Mark and David Geier is how they had their IRB approval to do a study based on the Vaccine Safety Datalink data suspended in 2003 because of inproprieties in which they renamed files and tried to rename datasets in ways that aren't permitted and in such a way that patient confidentiality might be compromised.

What I hadn't been aware of is the Geiers' response to this criticism. Kathleen deconstructs the Geiers's explanation of what happened and demonstrates pretty convincingly that, for all the verbiage designed to distract and produce sympathy, the Geiers in essence admitted that they did what the CDC accused them of doing. She boils it down thusly:

The same sort of political pressure that enabled this legal consulting team to initially gain access to millions of private patient records was brought to bear in the spring and summer of 2004. Vaccine litigant groups and their political allies trivialized concerns about the privacy of patient records, characterizing these concerns as a "false pretext" for the denial of the Geiers' access to VSD data. To my knowledge, no determination was ever made that the technical monitors' allegations of misconduct were "baseless." However, Mark Geier and David Geier's IRB approval was reinstated that August. This reinstatement occurred in spite of their own admissions that they:

  • initiated analyses not authorized in their research protocol;
  • attempted unauthorized merging of datasets; and
  • acquired unauthorized data files, which were named in a manner that, if undetected, could have enabled the removal from the Research Data Center of private information about millions of U.S. citizens.

Sadly, those nefarious forces aligned against the poor Geiers ultimately reinstated their IRB approval. If I violated my IRB protocol that way, I doubt I'd get the same favorable treatment.

More like this

The data belongs to the taxpayers. Anyone should be able to walk in there and do whatever they want with it anytime they feel like it. The only impropriety here is the government placing ridiculous restrictions on the Geiers. Sounds like part of a coverup to me.

Don't feed the trolls.

By Magnus Malmborn (not verified) on 25 Mar 2006 #permalink

Quote:

"The data belongs to the taxpayers. Anyone should be able to walk in there and do whatever they want with it anytime they feel like it. The only impropriety here is the government placing ridiculous restrictions on the Geiers. Sounds like part of a coverup to me."

Let's do this in parts, shall we?

One. Patient confidentiality. There's a reason why it exists. It's there to protect the weak (in this case sick people) from being discriminated against for being sick. As certain unscrupulous organisations in the past have done. Indeed, insurance companies are pressing for genetic testing so that they can increase the premiums of those whose genetic tests show up any hitherto unforeseen "problem". Worse still, some insurance companies are pushing for the right to raid medical records to gather that information without telling the people whose records they're plundering, before jacking up the premiums of these people without a word of explanation. On the basis of the above comment of yours, you would seem to have no objection to this. Well I, and quite a few others (including an awful lot of doctors) do.

While on the subject of patient confidentiality, here's something else to consider. Not so long ago, haemophiliacs were given a rough ride by ignorant bigots when it became public knowledge that the clotting factors administered to them to save their lives had become contaminated with HIV. There have even been instances of children being told that they were no longer welcome at school in case they somehow 'infected' the other children (a view that is utter gibberish to anyone with a minimum of ten functioning brain cells, but sadly the world is full of people who believe gibberish). No doubt if those haemophiliacs had been on a public database, you would, given your comment above, approve of those same ignorant bigots raiding the database so that yet other bigots could target the unfortunate haemophiliacs for who know what. After all, those bigots are taxpayers too ... ?

Two. Proper use of data. If a scientist intends to use data to reinforce a hypothesis, there are numerous treatises written by researchers in statistics over the years detailing the kinds of inferences that may be drawn legitimately from a data set, and the kinds of inferences that are invalid to draw from that data set. The classic example being the improper use of time as a connecting variable. Illustrated eloquently in a lecture on the subject I took during my mathematics course at university - the lecturer wrote the following upon the blackboard:

Since 1700, life expectancy in the West has risen from an average of 45 years to an average of 72 years. During that same period, tobacco consumption has increased 20-fold. Therefore tobacco is connected with increased longevity.

An absurdity writ large of course, but that was done simply to highlight the point. There are legitimate inferences that can be extracted from data, and tools to perform that task. Likewise, there are all kinds of pitfalls that even honest researchers can fall into if they don't stop and ask themselves searching questions along the way. This is before we address the possibility of intentional misuse of data to pursue an agenda.

Are you happy with the idea of public data being misused to further an agenda whose ultimate interests lie not in the well-being of our fellow human beings, but the pursuit of filthy lucre through egregious lawsuits that, if successful, will take money away from genuine research intended to benefit Mankind and line the pockets of self-styled 'consultants'? From your above comment, this would seem to be the case. Sorry, I have to stand a LONG way away from you on this one. Additionally, given how egregiously the Geiers have performed despite these 'ridiculous restrictions', it should be a cause of sheer pants-wetting terror to consider how they might behave if freed from such constraints ...

I'll let Orac fill you in on some of the background. He's been watching these people a long time, and knows snake oil merchants when he sees them (or in this case, most likely, smells them ...)

By David Edwards (not verified) on 25 Mar 2006 #permalink

David;
I'd like to see the Geier's have all the names. It would be nice if they could contact all those parents and let them know the truth that the drug companies who control the media with their advertising don't want us to learn. If there is nothing to hide, I should be able to go in there and look at the data myself.
Your tobacco strawman doesn't give your argument credibility.
Lawsuits to compensate children and families who have had their lives destroyed by mercury poisoning are far from frivolous. Let's have you trade places with me for awhile so you can stay on your feet and alert every second that your autistic son is awake and see how you like living that way.

Fore Sam: The data belongs to the taxpayers. Anyone should be able to walk in there and do whatever they want with it anytime they feel like it. The only impropriety here is the government placing ridiculous restrictions on the Geiers. Sounds like part of a coverup to me.

Great, then you wouldn't mind sharing your complete medical records with us right John? Make sure you include all of those STD's you didn't pick up when you weren't in the millitary.

Still waiting on the brain MRI report you promised to share with us. I understand they didn't find anything. Sounds like part of a coverup to me.

Fore Sam:

The data belongs to the taxpayers. Anyone should be able to walk in there and do whatever they want with it anytime they feel like it. The only impropriety here is the government placing ridiculous restrictions on the Geiers. Sounds like part of a coverup to me.

And:

I'd like to see the Geier's have all the names. It would be nice if they could contact all those parents and let them know the truth that the drug companies who control the media with their advertising don't want us to learn. If there is nothing to hide, I should be able to go in there and look at the data myself.

Let's look at the implications of your assertion that sensitive data that "belongs to the taxpayers" should be freely available to investigators. May I assume that you wouldn't mind letting the the pharmaceutical companies have all your medical data as well, so that they can use the information as part of a project to develop a defense for all those lawsuits that will result? Or how about your taxpayer information? Technically that "belongs to the taxpayers" as well. Should that be freely available to financial consultants who want to target their marketing more freely?

Ask yourself: Do you really want the confidential health information about you and your family available that way? Remember, under your apparent proposal, if people you trust or like (like the Geiers) can get access to your personal data, then people you don't trust at all could also get access to it as well, such as insurance companies, pharmaceutical companies, etc.

Orac;
You have a point but, honestly, if it helps all parents learn the truth about what happened to their kids, let's make all the data public except the names. Just put it on the internet for everyone to see. Perhaps some horseplayers will take an interest in it as a challenge. There's nobody better at interpreting data than horseplayers. I just don't want people like Kathleen Seidel looking at it with names included since she'd blab everyone's personal troubles all over the internet. Maybe Andy Beyer would decide to look at it. Then you'd have an unbiased answer. I want the truth publicized and I don't care what it takes to have that happen.

Leaving the names out would not guarantee that data-mining software wouldn't be able to match a fairly high percentage of cases to patient names.

"There's nobody better at interpreting data than horseplayers."

So professional statisticians aren't very good? You want horse gamblers to try to analyze data? Isn't there a reason that there are so many "systems" to predict which horse is faster? Isn't there a reason most of these "horseplayers" end up losing money in the long term. If the horse gamblers are so good, why run the races anyway? Just get a consensus of gamblers together and award the prizes.

LMAO

Fore Sam: if you seriously think that anti-vaccination campaigners are having difficulty getting their point across, you need to get out more. Pharma companies control drug advertising, but they don't control the media: science has no sway in the soundbite press. People like these stats dropouts invariably do.

The UK tabloids pretty much invented the MMR vaccine scare over here and it rode high on the front page until MRSA became a more satisfying target (with the aid of a fantastically unscrupulous lab which can find MRSA anywhere). The studies which refuted any link were absolutely not reported on: the conjectures and wild handwaving for those in the scare camp got reams over coverage. Now that's a coverup.

By Alexander Whiteside (not verified) on 28 Mar 2006 #permalink