doug@terak.UUCP (Doug Pardee) (10/14/85)
> Two other important factors are age and sex, but obviously these can't > be altered. Which brings up a universally overlooked point when "risk factors" are quoted... -- Does altering the risk factor have any effect? For example, both "male pattern baldness" and hemophilia have "male sex" as major risk factors. A male-to-female sex change operation will prevent male pattern baldness but not hemophilia. Determining that "attribute X" and "ailment Y" have some positive or negative correlation is indeed an important finding. But that should be followed up by studies to determine if "intentionally removing attribute X" will change the incidence of "ailment Y". Likewise, there should be further studies to determine if "intentionally introducing attribute X" will change the incidence of "ailment Y". -- ...a new study indicates that breathing has been implicated as a risk factor in a number of respiratory ailments... -- Doug Pardee -- CalComp -- {calcom1,savax,seismo,decvax,ihnp4}!terak!doug
werner@aecom.UUCP (Craig Werner) (10/19/85)
> Which brings up a universally overlooked point when "risk factors" are > quoted... > > -- Does altering the risk factor have any effect? > Doug Pardee -- CalComp Two criteria for true risk factors: 1. The correlation should be statistically significant and of course, reproduceable. 2. It should be dose-dependent. For instance, in american males, heart disease is related to smoking both by the number of years smoked and number of packs (it can be expressed as pack- years.) So quitting smoking does help. As for cholesterol, the higher one's LDL, the higher one's chance of dying early from heart disease. In fact, the dose-relation is not linear - it is more like hyperbolic. And a "normal" cholesterol level of 260 is probably far too high, and "normal" should be revised. (incidentally, the upper limit of normal is defined as 2 std deviations above mean, so by definition, 95% of all people are normal or below. Clearly this is not the case clinically, just statistically.) The recently completed Lipid Research Study has just shown that lowering blood lipid DOES lower risk enough to justify medical intervention. As for high blood pressure, with one small exception, lowering one's pressure and keeping it low, lowers one mortality risk. This is probably the best documented finding in all interventional epidemiology. -- Craig Werner !philabs!aecom!werner "Comedy, like Medicine, was never meant to be practiced by the general public."
doug@terak.UUCP (Doug Pardee) (10/25/85)
[It's hard to argue with someone who knows what he's talking about] > For instance, in american males, heart disease is related to smoking both by > the number of years smoked and number of packs (it can be expressed as pack- > years.) So quitting smoking does help. But, but, but... The conclusion does not *necessarily* follow. I do accept the conclusion, but I don't believe that it has been properly proven. It is the method of "proof" that I'm questioning. As a counter-theory, try this: The kind of person who smokes is often a "Type A" person. The kind of person who has heart disease is often a "Type A" person. If we express the current theory as Type A --> smoking --> heart disease then we can express an alternate theory as Type A --> other factor --> heart disease Type A --> smoking Under this alternate theory, people who quit being Type A quit smoking *and* reduce the risk of heart disease. This gives the effect which has been observed in population studies. But if a person quits smoking in order to avoid heart disease *when he otherwise would continue* and remains a Type A person, there would be no effect. It is *very* difficult to set up a properly controlled experiment when the variable to be introduced is human behavior. It is equally difficult to draw reliable conclusions from population studies when the independent variable is human behavior. Human behavior is too often not (maybe never is?) an "independent" variable. -- Doug Pardee -- CalComp -- {calcom1,savax,seismo,decvax,ihnp4}!terak!doug
werner@aecom.UUCP (Craig Werner) (10/30/85)
> [It's hard to argue with someone who knows what he's talking about] > > But, but, but... > > The conclusion does not *necessarily* follow. I do accept the > conclusion, but I don't believe that it has been properly proven. It > is the method of "proof" that I'm questioning. > > As a counter-theory, try this: > > The kind of person who smokes is often a "Type A" person. > The kind of person who has heart disease is often a "Type A" person. > -- > Doug Pardee -- CalComp -- {calcom1,savax,seismo,decvax,ihnp4}!terak!doug The only part of the article I agree with is the first line. Here is the fatal flaw in the counter-argument: Type A behavior is at best, a lousy risk factor for CHD (Coronary Heart Disease), despite what you read in Reader's Digest. Smoking increases the risk in both type-A and non-type-A people, and is much stronger than type A, and it is Dose-related to the number of pack years. To put it plainly, a Type-A non-smoker is much less likely to have heart disease than a non-Type-A chain-smoker. Counter-argument fails. I repeat: Smoking is the #1 preventable risk factor in heart disease, and is the #1 cause of shortened life expectancy in the United States from all causes. So, if you smoke, see a doctor or other professional to enroll in a quit-smoking program. If you don't wish to quit, cut down. -- Craig Werner !philabs!aecom!werner "I never knew there was anything wrong with me till I met Dr. Hackenbush."