Sunday, June 22, 2008

Under pressure to publish? Publish a coffee study!!!

[Click on Post Title for Link to External Article]

Honestly! Trust me. You can never go wrong with coffee. Many people drink it, sell it, swear by it and some also swear at it. This should be enough to confound anyone and anything. And, to publish and see your name in the "scientific" press, you don't have to spend much. Just make a few guzillion pots of coffee!!!

Many moons ago, in 2004 to be precise, I was interning for a blood glucose monitoring startup and it was also the time when "Google Alerts" became the "in-thing". As a result, I had an alert set up for diabetes. And that's where the saga began.

I soon started seeing so many of these coffee studies pop up in relation to diabetes:

1. Coffee causes Type II Diabetes

2. If you drank more than x, y or z cups of coffee a day, you will not get Type II Diabetes

3. Coffee causes Type I Diabetes

4. Coffee does not cause Type I Diabetes (and in fact it doesnt, and there may never have been a study of this sort, but there were so many coffee studies, who can tell anymore?)

5. Coffee "may" cause Diabetes

6. Coffee "may" not cause Diabetes

7. Coffee and Diabetes "may not" cause Diabetes

8. Coffee messes with your testicular production quotas! (We haven't seen one dismissing this yet. But to date, this is my most favorite study ever. They had a bunch of men drink coffee and, er, counted 'em soldiers!? I suppose...)

Coffee studies have become the mainstay of pointless and needless publications. This one takes the "cream" I suppose.

You follow people for years and years and have them fill up a survey. What method could generate better results? What other scientific form of study could be superior to this?

If you spend a few minutes past the cynicism, here are a list of reasons why I have a problem with this "coffee" study:

1. "Coffee consumption has been linked to various beneficial and detrimental health effects, but data on its relation with death were lacking,"

- Yes, because, not many people pop after they drink coffee. Maybe the guy in the photo on Science Daily is going to, one of these days...

2. Coffee consumption was not associated with a higher risk of mortality in middle-aged men and women.

- Dissociation is not causation either. Coffee consumption was not associated with higher mortality, because you could not, based on your own assertion in 1 above.

3. The possibility of a modest benefit of coffee consumption on heart disease, cancer, and other causes of death needs to be further investigated.

- Insert the "Are you kidding me" cliche here! Heart Disease, Cancer and other causes of death? What is this, MOAS (Mother of All Studies)?

I don't understand what the thought process for this was. Cardiovascular risk is associated with any number of problems from congenital defects to a sedentary life-style, diabetes, alzheimer's, side effects from drugs?! and so on.

Cancer causes death through complex biological mechanisms.

And other causes of death? Hmm...let's see kidney failure, influenza, the hit-by-a-bus-while-reading-stupid-coffee-study-reports syndrome....

How is it that they could casually send out a bunch of questionnaires over a 20 year period and expect this to be a study with conclusions of any positive value?

Why is such drivel getting past "peer" review? Maybe the reviewers weren't drinking enough coffee?

And yes, if your head aches, get a little coffee...the caffeine helps get rid of a headache. That much we know. About hot flashes, alien landings and death....I am sorry, we haven't had a good study yet.

But, if you want to publish, you know what to do...whip out those survey monkey questionnaires, now that we have "modern" technology to bungle the same scientific reasoning principles...

Saturday, June 21, 2008

Follow Up: How 23andme is getting some free spin doctorism...who's buying?

[Click on Post Title for Link to External Article]

I also want you to read this one from the same blog:

And my original post here:

Here is my problem:

Most things are actually not your decision. This viewpoint that because it is your body you can do anything with it sounds like something you hear from irascible teenagers and pot smokers.

This is not about you at all!

This is also not about whether or not WebMD is organizing information the way you would appreciate it and how 23andme is better organized at rehashing the same information.

This is about vulnerability.

Of course they are not "telling" you that you should get yourself tested. But heck, they are implying it.

And if the Doctors know less than you about genetics (from the second blog I believe) - however believable that statement might be - it doesn't solve the basic problem.

The practice of medicine is mired by its own puff-able concept "Gold Standards". And even then, genetics hasn't gotten to the stage where insurance will pay enough and Doctors will pay attention. It will get there soon, not just yet.

But, we need to get back to the context.

Here is the problem with 23andme and its Redwood City look-alike:

1. We understand very little about how genetics actually work, even less about why SNPs occur, how they are triggered to start up or shut themselves and so on. Yeah, we know something...we have no clue about most other things.

2. In case this is news to you, people are gullible. The guy who's dad just died from Colon Cancer, the lady who's mom died of breast cancer. 23andme, believe it or not is a lovely, expensive trap for such people.

3. Let's go back to your Doctor thing. Let us say that family X did get their genes profiled. You admit yourself that based on your extensive survey and unfathomable depths of prescience, most Doctors have no clue what A,T,C,G and U do other than serving as alphabets. Given this situation, knowing that no insurance company wants to pay for the information, what is family X to do with the junky printouts 23andme sells them.

4. This may also be news to you. There is this group of people called hypochondriacs...

Oh my God! The things people say and imply. California sent those C&Ds at the appropriate time. Before we understand genetics, trying to randomly decode information at a perky price of $1000 or higher is as good as offering people to clone their dogs for them! (Apparently there is a huge market for them, and yeah, yeah, they are your dogs. God, someone call the SPCA - it's after all the dog's own genes...)

(Note: If 23andme were to do it for free at this point, I would have no problems! That would be a true contribution to science. Otherwise its like trying to get 1000 monkeys to come up with Shakespeare's plays, and asking the monkeys to pay for it as well)

And come on! Who are you fooling? If you walk down the street and described 23andme to random individuals on the street, how many would actually realize, admit or agree to your view that this is NOT about healthcare and that the Department of Public Health should just wait and watch.

Yes, 23andme and Navigenics are just that - snake oil merchants with deep pockets!

And if you are mad, maybe 23andme can figure out and help you remove the "mad" gene. I am sure they have good, well organized notes on why people get mad....

Monday, June 16, 2008

My husband had all this money lying around, so I thought I could do something medically useless in the field of can that be

[Click on Post Title for Link to External Article]


To quote: "Requiring that such tests demonstrate clinical "utility" could pose a particular problem if applied to 23andMe: The company has admitted its tests are not medically useful, as they represent preliminary findings, and so are merely for educational purposes."

Well, what have we here! So much money and they did not get their lawyers to do a simple fact check?

Demonstrating clinical utility could be a problem for a genetics test company?

Of course, all that really needs to happen is for Sergei to take a plane to Sacramento and convince people that demanding utilities in products is an absurdity:

"Our family has a tradition of creating half-functional products that we like to then display as the 'next best thing'.

Demanding fully functional or otherwise useful products interferes with our ideals to do business with no meaning, purpose or use and severely inhibits our ability to charge people for it!

If you don't stop making such demands immediately, we will buy California and sell it to someone else."

And so the story goes...

In reality though, if you look past the silliness, I do not have any sympathies for 23andme or Navigenics. Sailing on funds available through randomness or gullible Venture Capitalists, these firms set off to "create genetic profiles" that are "preliminary results only" and what not.

I am also not ga-ga over Navigenics "going through a doctor".

Its not like everyone with a Medical Degree has been sifted through the sieves of ethical practice.

We have barely understood that we have about 30,000 odd (or 40,000 odd was it?) genes in the human body.

We still theorize and struggle to explain why, our genetic scheme is only 1 - 2% different from that of the chimpanzee and that somehow fabulously makes us superior ( a very subjective opinion) as animals.

We have just finished realizing that our mitochondrial DNA - in statistical summary all of ours, meaning the 6 billion of us, almost shared the same mother once - meaning our entire race was nearly destroyed.

All this means one thing: we need to understand a lot more of the nuances of genetic interactions and how they are turned on and off before running around with a CD of our genetic code. Of course, whenever SNPs lead to disease, we can start working on those in the near future, but for most other complex genetic interactions, data provided by companies like 23andme would be useless.

Yes, indeed, it would be a great idea to cull genetic data from thousands of human beings. It would allow for statistical correlations (if we believe in statistics, that is) and eventually solid proof.

However, I have some serious problems with companies that want to charge people for things even they consider to be useless.

It is also an important lesson for entrepreneurs.

Sometimes funding is not enough.

Sometimes funding itself can make you shortsighted and lead you down the wrong path.

What 23andme or Navigenics have failed in doing will be a great stepping stone for the next company that comes in:

1. Research, research, research. Clearly understand what you are selling, understand who is buying and what they want and why the want it. Are they buying something because they want it, need it, or think they need it (like the practically useless ipods and iphones)?

2. Understand national, international and rather more importantly local regulations - especially if "local" happens to be California.

But don't stop there. Just because something is not being regulated right now does not mean you can run away with anything (or precisely, that something). States and Nations often start regulating an industry in the wake of new practices. Of course, most of this is driven by fear, sometimes unfounded, but you need to spend at least a little time thinking about this, especially when you are doing something "groundbreaking".

3. Be wary of where your funding comes from. While it is often harked that you need supportive investors, you also need independent investors who need to judge if your idea makes for an actual sale-able product or service.

4. Don't name something "educational" (refer back to the quote about 23andme) unless it is actually educational. Yes, genetic testing can be educational. But how can it be educational to potentially gullible customers who paid for something useless?