Monday, January 28, 2008

"Rockin' on the horse-sized pills"

I've posted before on some of the bullshit that gets reported as fact thanks to research in health and the social sciences, but here's one a little closer to home. The general theme is that when people start talking "Studies show..." they generally mean "I firmly believe..." or something similar. Warning - long post.

A while back my doctor prescibed me Lipitor, a statin, which is used for lowering cholesterol. My cholesterol levels are within the normal range for a healthy person, but I'm a diabetic. Diabetics are at greater risk of heart disease, so my "normal" cholesterol was high enough to get me prescribed statins.

Now, I'm the typical sort of dumbass NZ bloke who'll treat almost everything with the utmost skepticism, except when my doctor says "Here, take these pills." Fortunately, my wife is made of sterner stuff. A net outcome is I now know a bit about statins, and can point up some "studies show..." bullshit as a result.


Bullshit 1: "Increased Risk" as a blunt instrument.

My doctor prescribed this over-hyped placebo for me because diabetics are at "increased risk" of heart attack. Well, no arguments there - 65% of diabetics die of heart attacks or strokes. But where does this increased risk come from? Well, diabetics are often grossly obese and suffer from high blood pressure and high cholesterol. But what if they don't suffer from those? I weigh 75kg, am physically fit and have normal blood pressure and cholesterol - am I genuinely at a greater level of risk of heart failure than the healthy general population, or have I just had this "increased risk" label slapped on me by virtue of sharing an illness with a swag of morbidly obese couch potatoes? Frankly, I suspect the odds of me popping my clogs via a heart attack tomorrow are actually no worse than yours. It may be that "studies show" diabetics are at greater risk, but in fact it's too blunt an instrument to be useful in this case. Unfortunately, this kind of blunt instrument gets a lot of use.

Regardless of my personal situation, statins are in fact good medication - if you're already suffering heart problems. But there's a big question mark over whether they're any use whatsoever to the rest of us.

There's an excellent article summing up the pros and cons in Business Week, of all things. I'm going to include some excerpts from that article to illustrate:


Bullshit 2: the promotion of "relative" vs "absolute" risk.

You come across this one all the time: stuff along the lines of, "Chewing your fingernails increases your risk of arse cancer by a shocking 50%!" This is "relative" risk. Researchers love relative risk because it makes the result sound impressive. For a reality check, look at the "absolute" risk - ie, the actual numbers. If it turns out that the 50% increase amounted to 3 deaths per thousand people rather than 2, things are suddenly much less impressive. These kinds of studies always need going over with a fine-toothed comb - what's the affiliation of the researchers, who's paying for their research, do they have a history of promoting the abolition of fingernail-chewing etc. How many people were in their study, how were controls established, did the same group of people participate all the way through, how long did it run for, have all the results been included in the findings etc. Given that the guys writing the report will have pulled every trick they can think of to get the maximum significance from their results, how big is the difference really likely to be? Getting to a picture of how trustworthy the study might be is far more work than most of us have the time for, but thankfully there are some people out there willing to put the hard yards in.

The case of statins is an example of Bullshit 2. Here's Business Week:

The drug business is, after all, a business. Companies are supposed to boost sales and returns to shareholders. The problem they face, though, is that many drugs are most effective in relatively small subgroups of sufferers. With statins, these are the patients who already have heart disease. But that's not a blockbuster market. So companies have every incentive to market their drugs as being essential for wider groups of people, for whom the benefits are, by definition, smaller.

(...)

The printed ad proclaims that "Lipitor reduces the risk of heart attack by 36%...in patients with multiple risk factors for heart disease."

(...)

The dramatic 36% figure has an asterisk. Read the smaller type. It says: "That means in a large clinical study, 3% of patients taking a sugar pill or placebo had a heart attack compared to 2% of patients taking Lipitor."

Now do some simple math. The numbers in that sentence mean that for every 100 people in the trial, which lasted 3 1/3 years, three people on placebos and two people on Lipitor had heart attacks. The difference credited to the drug? One fewer heart attack per 100 people. So to spare one person a heart attack, 100 people had to take Lipitor for more than three years. The other 99 got no measurable benefit. Or to put it in terms of a little-known but useful statistic, the number needed to treat (or NNT) for one person to benefit is 100.

(...)

Plus, there are reasons to believe the overall benefit for many patients is even less than what the NNT score of 100 suggests. That NNT was determined in an industry-sponsored trial using carefully selected patients with multiple risk factors, which include high blood pressure or smoking. In contrast, the only large clinical trial funded by the government, rather than companies, found no statistically significant benefit at all.

(...)

If we knew for sure that a medicine was completely safe and inexpensive, then its widespread use would be a no-brainer, even with a high NNT of 100. But an estimated 10% to 15% of statin users suffer side effects, including muscle pain, cognitive impairments, and sexual dysfunction.

(...)

A 36% decline in heart attacks sounds more dramatic and important than an NNT of 100. (...) But when it comes to side effects, they flip-flop the message, dismissing concerns by saying only 1 in 100 people suffers a side effect, even if that represents a 50% increase.


These highlights are merely a sample of the hair-raising stuff included in the article - it's well worth a read.

Given how long this post's gone on already, I won't go into Bullshit 3: Correlation vs Causation. Suffice to say there's reason to believe the success of statins in reducing heart attacks among existing heart disease sufferers isn't simply down to lowering your cholesterol.

So, by taking these pills, under optimum conditions, if you're someone with a bunch of risk factors for heart attacks, you might be the lucky one in a hundred who gets saved from a heart attack. Far more likely of course, is that you'll be one of the unlucky 10 to 15 in a hundred who gets treated to some seriously dodgy side effects.

I didn't take the pills.

Next week - the bizarre propaganda war over saturated fats.

6 comments:

Adolf Fiinkensein said...

Milt, if you're looking for bullshit, I think you need go no further than this.

"With statins, these are the patients who already have heart disease."

Have you ever heard of familial hypercholesterol? It's not 'heart desease' but statins sure as hell work for people who have it.

Give me a call when you've died and I'll come to your funeral.

Psycho Milt said...

Adolf, given that statins lower cholesterol, I'm sure they do work for people with a rare genetic condition causing them to produce way too much cholesterol. But an overview article for lay readers doesn't really go into that level of detail.

Sean said...

"Well, diabetics are often grossly obese and suffer from high blood pressure and high cholesterol. But what if they don't suffer from those? I weigh 75kg, am physically fit..."

Well you could still be grossly overweight at 75kg if you were a thieving Swedish dwarf.

Cecilia Tankersley said...

"People with high cholesterol live the longest."

See www.westonaprice.org
/moderndiseases/benefits_cholest.html

I get really cross when I read about studies that everyone agrees "show ..." when the actual studies did not actually prove that at all. However, the "conclusion" is repeated and repeated until it becomes accepted as the truth.

Adolf Fiinkensein said...

Milt, I have never before heard familial hypercholestrol referred to as 'rare.' I remember some fifteen or twenty years ago when my doctor first prescribed Zocor, one of the conditions placed upon it by the gummint was that it was fully funded for this condition. Our family has it and I understand so do many others.

Psycho Milt said...

I looked it up and Wikipedia said it's about 1 in 500. That sounded pretty rare to me - but I guess, given there's 4 million people in NZ, it adds up to quite a few. Apologies, I withdraw the term "rare."