How do doctors decide what is healthy and unhealthy? Do they arbitrarily decide on risk factors to line their pockets - creating false epidemics as Sandy Szwarc at Junkfood Science suggests? Or, is there actually a science, called epidemiology, that is the basis for health recommendations?
As I've said repeatedly, one of the sure signs you're about to hear total BS is if someone suggests there is some conspiracy by scientists or doctors to hide the truth. In an article challenging the use of serum troponin levels to determine whether myocardial infarction (MI) has occurred (a more sensitive method) Sandy suggests this is yet another example of doctors lowering metrics of illness and risk to generate the impression of false epidemics.
There must be a health crisis to bring the greatest funding for research, treatments and education... even if an epidemic has to be created. One of the most common tactics is to change the definition. When diagnostic criteria is broadened, suddenly, with the stroke of a pen, new cases can appear to explode in number.
With heart disease deaths dropping dramatically for the past half century, the world's top four organizations representing heart disease interests have all gotten together to change the definition ... of a heart attack.
The World Heart Federation, American College of Cardiology, American Heart Association and the European Society of Cardiology have been championing the new criteria over recent years, and will officially release it next month in the Journal of the American College of Cardiology and in the AHA journal, Circulation. The new definition will use elevations of troponin levels, rather than the traditional cardiac biomarkers, such as the MB-CK enzyme.
This is a truly bizarre argument. Because measuring troponin will allow us to detect more MIs that have occurred, it must of course be part of a plot to make Americans think they're less healthy. Never mind that more sensitive tests for MI are what is known as a good thing, and that current tests clearly are missing minor heart damage thus underestimating the number of true MIs. Any revision of current standards must be part of a plot! The fact that those scientists got together is a sure sign. We should never let them do that.
While troponins may go up for other reasons, the idea that the test will misdiagnose as heart attacks other disorders is pretty silly. This isn't a test that is going to be used to diagnose MI in the absence of chest pain or abnormal ECG findings - which enhances the specificity of the test - and doctors are aware of confounding diagnoses - it's their job to find them. Further, the idea that the new criteria were designed to somehow justify funding for heart disease (an area of medicine that will never lack for funding) is downright hilarious.
But this isn't the only example of "false epidemics" being created by those greedy doctors trying to convince people that they're ill. Sandy mentions other excellent examples.
First, because this is Sandy, is of course obesity:
"Overweight:"Definition changed from BMI ⥠27 to BMI ⥠25 by the U.S. National Heart Lung and Blood Institute in 1998, instantly increasing by 43% the numbers of Americans, an additional 30.5 million, deemed 'overweight.'
So, did doctors just pull that figure out of nowhere to line their pockets? Or does data exist that justify the decision? This study in NEJM(1) is a prospective study of over a million people evaluating all-cause mortality (as well as a number of independent risk factors) showing the relationship between BMI and mortality. Here is the relevant figure - the dark line is most relevant - showing the relative risk of death versus BMI.
Continued below the fold:
The guidelines for the NHLBI cite similar (although older) studies that are the basis of their rationale. Consistent with this trial, which I think is the largest study of its kind, and studies specifically justifying the overweight recommendations (2) in a cohort of 50-70 year olds, there is a definite advantage in terms of overall mortality for keeping BMI below 25.
For men:
And women:
So, what do you think about the definition starting at 25 now? I realize BMI is an imperfect measurement, but on average it does an OK job. Until new anthropometry is developed and implemented, we should recognize that it is informative about risk within the bounds of common sense (it's clearly off for some individuals like professional athletes and weight-lifters). I should also note that some studies have indicated a slight advantage of being overweight (25-29) in men over 55, this needs more study and may reflect the confounding tendency of the ill to lose weight before they croak.
Not satisfied with the usual obesity nonsense, she also mentions the changes in the definition of dyslipidemia:
"High cholesterol:"Definition changed from a total cholesterol ⥠240 to ⥠200 in 1998 increasing by 86% the numbers of Americans labeled has having high cholesterol, an additional 42.6 million adults.
Now where did we come up with that definition? Is it possible a study showed an increased risk of cardiovascular disease (CVD) with those levels? Well current standards are based on the Third Report of the National Cholesterol Education Program (NCEP) (3) also known as "ATP III". What is their rationale for keeping the total cholesterol below 200?
Now these may seem like smaller effects, but remember, that cardiovascular deaths are such a big killer that even small decreases in risks can amount to many tens of thousands of lives being extended. It is also of note that the risk of cardiovascular events continue to decrease linearly as cholesterol drops below 200, suggesting an advantage for aggressive control especially among high-risk individuals. Statins also may have some anti-inflammatory benefit as trials lowering cholesterol using statins show benefits in excess of what one would expect from their effects on total cholesterol.
She also challenges changes in the definition of hypertension.
"Hypertension:"Definition changed in 1997 from 160/100 to 140/90, instantly adding 35% more Americans, 13.5 million, to the rosters of hypertensive. A new definition for 'prehypertension' in 2003 increased to 58% the Americans believing they have hypertension.
The Lancet has the best example of why this was a good idea (4).
As you can see the risk increases between 70- 90mmHg diastolic (considered the more important number) and about 120-140mmHg systolic in most age groups by about 100% (note the log-linear scale though - I find this presentation annoying and a natural scale shows the importance of the 90mmHg cut-off just fine). The point of lowering the standards wasn't to increase the number of people treated, but to include people who could benefit from treatment through lower risk of CVD. Remember again how many people die of cardiovascular disease, and cutting a risk in half through blood pressure control means many many lives saved.
Finally diabetes is apparently an engineered epidemic as well:
"Diabetes:" Definition changed from a fasting glucose of ⥠140 to ⥠126 in 1997 by the American Diabetes Association and WHO Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, increasing by 14% and 1.7 million the people diagnosed with diabetes. With the proposal of a new term, 'prediabetes' by the First International Congress on Prediabetes, and promoted by the International Diabetes Federation (sponsored by 12 pharmaceutical companies), 40% of the adult population was added to the rosters believing they have diabetes and are in need of treatment.
Here we have a kernel of truth. This was an arbitrary change that doesn't appear to show a benefit in terms of decreased morbidity or mortality. In the Expert Committee report(5) they admit their inclusion criteria were somewhat arbitrary in terms of health benefits, but did match where they felt subsequent glucose challenge testing would show a failed insulin response - a reasonable assertion.
I don't think this means, as Sandy suggests, that they were influenced into it by the evil drug companies, but rather they didn't properly take into consideration cost-benefit analysis of treating that extra 14% of people. As far as that and "pre-diabetes" one has to remember that diabetes is a progressive disease, and if people are showing signs of impaired glucose tolerance, it probably is wise to intervene to prevent the inevitable progression of the disease which would occur if you do nothing, and damage will accumulate as long as you have impaired glucose tolerance.
So what do you guys think? Do we have a rational for treating diseases? Or is this just more evidence of greedy doctors trying to ensure repeat visits for checking blood pressure, and diabetes and cholesterol blood work? There is a argument to be had about the cost-effectiveness of each of these interventions. These drugs do cost money, and while we may obtain a modest decrease in mortality, we may also strain the finances of the medical system as a whole. However, for Sandy to start talking about cost effectiveness doesn't seem right considering the most cost effective interventions are diet and exercise.
1. Calle, Eugenia E., Thun, Michael J., Petrelli, Jennifer M., Rodriguez, Carmen, Heath, Clark W. Body-Mass Index and Mortality in a Prospective Cohort of U.S. Adults
N Engl J Med 1999 341: 1097-1105
2. Adams, Kenneth F., Schatzkin, Arthur, Harris, Tamara B., Kipnis, Victor, Mouw, Traci, Ballard-Barbash, Rachel, Hollenbeck, Albert, Leitzmann, Michael F. Overweight, Obesity, and Mortality in a Large Prospective Cohort of Persons 50 to 71 Years Old
N Engl J Med 2006 355: 763-778
3. National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). NHLBI, September 2002. (http://www.nhlbi.nih.gov/guidelines/cholesterol/atp3full.pdf)
4.Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies, The LancetVolume 360, Issue 9349, , 14 December 2002, Pages 1903-1913.
5. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus
Diabetes Care 26: 5S-20S.
- Log in to post comments
It ain't either/or. Sometimes medical professionals have good _medical_ reasons for calling a condition 'pathological,' and sometimes their reasons are economic, or political, or religious, or just plain stupid. But by and large, asking clinicians to provide a coherent rationale for what they do is a fool's errand.
I always fail to see how preventative medicine could be a bad thing. Does she really think that the people who were slipping under the radar due to undiagnosed heart disease were better off before? Does she really think that doctors have time to sit around and make up ways to create epidemics? Because they're already not important enough?
And another question: if additional blood pressure check-ups and cholesterol screenings are just a ploy to make money off of us unwary fatties, why then does my campus clinic check my blood pressure 2x a year and my cholesterol 1x for FREE? Why does my husband get the same treatment from his non-campus affiliated doctor included in the price of an anual physical? This isn't a scam, this is PREVENTATIVE MEDICINE. It's what you want your doctor to do. It's what you pay them to do.
And to Sandy's readers who will probably stumble upon this and shout a few random words of outrage: I know you love her, and hang on her every word, and manage to cite her blog as if it were peer-reviewed research, but it's not. I challenge you to actually go and READ some of the articles she posts on and discover how much of them she leaves out of her posts. I know you think she's helping your cause by giving you tools to counter prejudice, but she is NOT. I've never understood the relationship myself, especially since many of you promote HAES, which includes eating nutritious foods and exercising regularly - both of which have been featured on Sandy's blog as being unlrelated to HEALTH. Hello cognitive dissonance, how are you doing today? Fat people ALREADY have poor relationships with doctors, resulting from failed attempts at weightloss and no current good/permanent options available. Why drive a wedge even further between the two? Why send a fat person into the doctor with an already suspicious attitude? It only increases the chance that advice intended to help will be misconstured as hate.
When viewed historically, diabetes is certainly an engineered epidemic...though not engineered by drug companies. Type I, insulin dependent, diabetics used to die early in life, prior to reproducing. With the introduction of insulin in the 1920s (I believe), these people survived, thrived and went on to reproduce, thus increasing the population of Type I diabetics. These people engineered one part of the epidemic themselves.
Type II diabetes, related to various abnormalities in insulin production and glucose utilization, is brought on by obesity, lack of exercise and a sugar-rich diet. Yes, it's an epidemic, but one brought on by lack self-control, and by the hyper-over-promotion of sugary foods in the US.
In each case, the patient, sufferer, or victim (if you choose) bears some responsibility for the disease. The drug companies are responding to medical demand, and to the serendipities of research which may create a new demand. But most obvious in the United States, is the idea perpetuated by educators and news media alike, that the body is too complex for the individual to understand, and the physician is your oracle.
I submit that if health education and biology were required curricula starting from elementary school through high school graduation we would have a healthier population.
I certainly don't dispute a link between excess weight and health risk. However, I confess the graphs you show do raise a question in my mind that wasn't there before.
NHLBI says normal wieght is BMI 18.5 - 24.9. However, looking at your first graph for healthy non-smoking men, the relative risk of death in the 18.5-20.4 BMI group seems higher than in the 26.5 - 27.9 BMI group. If the first group isn't considered underweight based on RRD, why should the second group be considered overweight?
Looking at men who are/were smokers and/or have a history of disease, the questions are even greater. For those men, the data suggest a BMI slightly above the recommended range has a lower RRD than any of the groups within the recommended range.
Qetzal, that is the classic "U" shaped, or "J" shaped curve that's usually observed in these instances and reflects a few things.
Among smokers, the belief is that the higher mortality among the thin is due to lung cancer and cachexia. You notice that when smokers are removed the curve blunts at the low end. Many studies I found in the search did not replicate that part of the tail at all when smokers were eliminated.
It is interesting that smokers benefit from being slightly overweight, I'm not sure of the explanation for that but know an obesity expert that I can ask. One thing it might represent is that weight loss can be a sign of serious illness. There are also some studies that show a benefit for older men for being slightly overweight - but again, this may reflect that weight loss is a sign of impending death. For men it might be worth changing the BMI scales to eliminate the skinny group, but I'll let the NHLBI or NIDDK figure that one out.
These studies aren't perfect for this reason, there are lots of variables that are difficult to control for.
Well, heavier smokers will get a smaller dose/kg of body weight. Maybe we are seeing a threshold effect.
Changing benchmarks may be justified. The problem comes in that many unscrupulous and/or uninformed people may misuse the new benchmarks to come to conclusions that are not justified. This of course must be battled.
A trivial example from recent history concerns the SAT scores. Remember how a decade or so ago the SATs were "renormed", effectively increasing scores by 50 points on each segment of the test? Sure enough a year or two later USA Today was trumpeting "record" scores on the SATs. Of course, these scores were records only compared to the unrenormed scores of the past.
Another example exists with Phoenix's air quality. A number of years ago the EPA and the city reached an agreement whereby the city was to reduce the level of pollutants to a series of steadily more stringent benchmarks. The city has succeeded in reducing pollution levels, although sometimes they have failed to reach the new targets. Whenever this happens, environmentalists and the media will announce that air quality in Phoenix is "getting worse".
MarkH:
Thanks for the additional info.
I did a bit more digging to try to understand how the BMI value of 25 was selected as the cut-off for being overweight. NHLBI has this report, which states:
However, when I go to the citations, they don't seem to really support the BMI = 25 cutoff.
One citation is to Troiano et al., but they say:
Another citation is to a report by WHO, but their section on overweight and mortality just discusses Troiano et al.
A third citation is to Van Itallie, but he defined overweight as BMI >= 27.8 kg/m^2 for men and >= 27.3 for women.
A fourth citation doesn't discuss specific BMI values in the abstract (and I don't have ready access to the full text), while a fifth is to a book chapter (also not accessible for me).
So, I'm still left wondering where is the evidence for increased mortality at BMI > 25.
On the other hand, the evidence for increased morbidity and risk of certain diseases at BMI > 25 seems much clearer. And there certainly is no dispute that mortality is higher once you get to BMI > 30.
Here's my explanation for smokers: What if smoking causes people to lose weight? This would result in roughly what we see, with a rise in deaths of underweight people. Of course, I don't have any good evidence that smoking does do that. I know there's a common myth that it does, but I haven't seen any studies that confirm it and I don't have the time on my hands to sift around for it right now.
Qetzal, I looked at those in preparation for writing this. They were suggestive, but certainly didn't confirm the overweight risk as subsequent studies did as the 95 CI's tended to overlap with 1 in that range while showing the beginnings of an increased mortality.
I suspect part of the reason was they were also relying on unpublished data from these subsequent studies. After all the expert committees tend to consist of people doing research in the field. Further the ideal BMIs seem to be 23-25, and recommendations all stem from a goal of primary prevention - keeping people from increasing beyond these levels.
Ultimately I agree that a definition of overweight should begin at 25 for healthy people. Subsequent research on larger cohorts of individuals has confirmed that that is where the curve starts to ramp after the well in mortality between 22-25. I don't believe the panels of experts defined it where they did to benefit the weight loss industry or because they wanted to define more people as sick.
Oh, neither do I. Sorry if I gave that impression.
Not to worry, I didn't think you had.
Not to pick a nit, but I think you meant 'ECG findings' in the second paragraph after the first quote.
Quite right. Fixed.
There has been a concerted effort to get people to not only perceive various mental disorders as illnesses, but to expand their perceptions of what counts as a mental illness.
Whether this is accurate or not, or a good idea or not, is something we could argue about for weeks. But it's clearly a "conspiracy" in the sense of a large group of people with common goals and beliefs working to accomplish an end.
How much of the increase in autism diagnoses is because 'autism' may be increasing, and how much is because many people have done a lot of work changing public and medical views on autism?
When we agree with this behavior, we call it "raising awareness". I think you can agree that there are a lot of people trying to "raise awareness" on a lot of issues.
Ivan Illich covered this issue (the rise of the health care-industrial complex) comprehensively in "Medical Nemesis" (1972)
We read it,
passed the pipe....
We debated,
and passed the pipe.....
We gnashed our teeth and vented about the oppressive system
and passed the pipe....
Then we got fat, lost our hair, bought BMWs and it seems nothing changed
Over to you Gen X/Y types!
To be fair, though, the problem isn't so much with >25 as overweight, but with that and <18.5 as underweight. at a quick eyeballing, it seems like the limit for underweight should be upped to 20 or so in order to get similar relative risks at both ends of the normal range.
Sorry. I forgot to use escape codes for <.
To be fair, though, the problem isn't so much with >25 as overweight, but with that and <18.5 as underweight. at a quick eyeballing, it seems like the limit for underweight should be upped to 20 or so in order to get similar relative risks at both ends of the normal range.
It looks great!
this blog showed that the non-smokers and smokers history. The mens and womens are used it. They faced to many other disease.
"As far as that and "pre-diabetes" one has to remember that diabetes is a progressive disease, and if people are showing signs of impaired glucose tolerance, it probably is wise to intervene to prevent the inevitable progression of the disease which would occur if you do nothing, and damage will accumulate as long as you have impaired glucose tolerance."
Since the main prescription at this stage would be exercise and eat right, which if followed correctly would probably mean never having to take drugs, you can hardly put the blame on the drug companies. They would probably profit much more if the diagnosis was left till the patients really needed drugs.
I'd say I feel quite ambivalent - 18-25 looks reasonable for women, whereas 20-27 would be more accurate for men. Having the upper threshold at 25 potentially causes millions of men to needlessly worry that they need to lose weight when in fact they do not.
Also, cutting off the lower part of the 'relative risk of death' axis makes things look worse than they really are.
As an actuary,* I would say you put quite a big foot in your mouth. The data you show provide, at best, very weak support for your assertions, and invite a lot of questions. Beside the points already raised by qetzal, BMI is correlated with physical activity, and it is unclear how informative it is about mortality and morbidity prospects after controlling for fitness. A study that doesn't measure indicators such as resting pulse and pulse after climbing a flight of stairs (or something like that - I am not an MD, so those are not necessarily the most relevant measures) is not going to be sufficiently informative about BMI itself to determine what should be considered normal.
Your cholesterol and blood pressure evidence doesn't suffer from that obvious correlation problem, but it also doesn't provide any reason for the thresholds to be what they are; all it shows is that lower is better.
I have never read the blog you criticize here (and I have no interest in starting to read it), and I am ready to accept (at least for the sake of argument) that it is biased and unscientific. But if it is as wrong in its conclusions (rather than just methodologically flawed) as you say it is, I would think you should be able to present more convincing evidence in support of the particular changes in diagnostic criteria.
* With BMI between 23 and 24, just to dispense with any ideas of a personal agenda.
I think Szwarc has a point - when it comes to how the MSM looks at a graph of increasing diagnoses over time and grabs at the word 'epidemic' without considering the changes in the number of people tested and changes in the benchmarks.
Other than that - the 95% confidence interval is around +/- 10% for many of the measurements in the BMI tables (in the NEJM article). If you make the lines that fat, it becomes much less clear how wide the healthy weight range really is.
While it's naive to assume all doctors are good people, or are doctors for mostly altruistic reasons, I really hate the way Szwarc and all the alties constantly slam doctors. I mean, if doctors were really just all about the $$$, they'd prefer (in this country, at least) to see people get sicker and sicker so that they could prescribe more medications and surgeries. Surgeries for bypasses and amputations probably pay a lot more than office visits/preventive care, I'd think.
Adrienne, you have not clue what you're talking about. As physicians, we work our asses off to prevent and treat disease, but people tend not to cooperate. We don't need to do anything underhanded to keep busy...patients are people, and behave irrationally as humans tend to.
I'll never run out of patients, because some diseases progress inevitably, and many patients can't change.
Sandy's blog post on expanding disease definitions is based on an article on The New York Times that was very well researched, along with some comments from me posted to a listserve.
There are several issues being lumped together here that need to be split out.
First, the utility of recognizing disease earlier and in more people needs to be split from hysterical declarations of "mounting epidemics" that appear in the press daily. All too often, the rising incidence of various diseases is based on stricter definitions of normal levels as well as more testing and evaluation. For example, telephone surveys show that twice as many people tell interviewers that they have diabetes than they did two decades ago. Part of this is that the definition of diabetes has changed from spilling sugar in your urine to fasting sugar of 140 to fasting sugar of 126. Some patients are told they are "diabetic" when their fasting sugar is 110. Another part of the apparent diabetes epidemic is more people are being tested. Twenty years ago, only a third of diabetics were aware of their condition, and now that is up to nearly two thirds in U.S.
Thus, you can have a doubling of the rate of self-reported diabetes even though ACTUAL FASTING BLOOD SUGAR LEVELS HAVE BARELY CHANGED. This last fact is the one you will never hear.
Unfortunately, Sandy Swarcz has gone ahead and labeled this a conspiracy. If it is a conspiracy, it is a beneficial one. The root cause is that pharmaceutical firms have developed new and more effective treatments for diabetes. Thus, doctors have a less fatalistic view of diabetes and are treating it more aggressively. The drug companies have financed mass screening for diabetes and educated doctors about the need for testing their patients. As a result, more people are diagnosed, more people have their blood sugar controlled, and the drug companies make more money. This is clearly a win-win.
Exactly the same thing has gone on for blood pressure and cholesterol. It is improved treatment options that have led to increased screening, and falling definitions of "normal". The level of cholesterol and blood pressure required to receive treatment is now below the population mean, so that more than half of middle aged people are defined as abnormal and given treatment. This is not necessarily a problem, because controlled double blind trials prove that treating high blood pressure, cholesterol and sugar save lives.
The other critical point is that the BMI story is fundamentally different from the stories for high blood pressure, cholesterol and sugar. The epidemiological data linking BMI to mortality are exceedingly weak, and are confounded by the strong relationship of low social status and education to high BMI. We have treatments that are very effective in raking in $50 billion a year in the U.S., but have no long term effectiveness in weight loss (except for dangerous surgeries). Another factor that makes BMI different is that we strongly stigmatize and poke fun at fat people which we do not do for people with high blood pressure. The intense interest of the entire population in losing weight drives the weight loss industry, the stigmatization and the over-selling of BMI as a risk factor.