I’m sorry. I’m not trying to be extra difficult, but where is the original source? Is it authentic?
If I understand correctly, the original source is this. I can’t find any way to access it online, but everyone who talks about it says the results mentioned in that blog post and no one says otherwise, so I’m assuming they didn’t just make it up.
As I said, mortality is not the only interesting endpoint. Also the CI upper bound for CVD is not over 1, not matter how hard we want to push it. The other review does support your conclusions. It doesn’t however support increasing dietary saturated fat or changing nutritional guidelines in any other way.
I see where you’re coming from. On the other hand, there is a $40 billion diet industry telling people to eat less fat and causing a spectacular amount of mental anguish around this idea (I don’t know if this is true in Finland; it’s definitely a big deal in America). That eating less fat has literally zero effect on any outcome seems like something we should be trying to make more widely known, whether or not the small effect of changing fat types on events but not mortality holds up.
That statins work equally well in high and low cholesterol populations is to me the most interesting claim that you make. Can you provide a source for it?
Apparently not; I know I’ve heard this but my attempts to Google a study in humans failed. I did find something in rabbits showing effects regardless of cholesterol, but even that wasn’t a direct comparison.
The other drugs you mention have common annoying side effects that mostly reduce compliance, and I wouldn’t be surprised if some of the side effects increased mortality. In Finland they are also usually prescribed by a specialist and are never a first line treatment.
Right, but the studies on niacin and ezetimibe showed that they decreased cholesterol (ie were being used successfully and correctly) but failed to decrease cardiovascular endpoints.
We have guidelines to measure lipid profiles after 48 hours of an ischemic vascular event. Within this time period, the LDL levels plummet, but then they rise again. This might be caused by a ruptured atherosclerotic plague releasing its contents in the blood stream. Since this is a very recent guideline, it might explain the finding you present. Then again, it also might not. If I skimmed correctly there was no mention of the timing of the measurements.
You’re right that the measurements were taken within 24 hours, but I’ve heard this isn’t such a big deal, and according to the full-text version of the Fonarow study (sorry, didn’t find it until this morning) they agree with me. Also, if I’m reading their table right patients having acute coronary events had higher, not lower, cholesterol than those coming in with chronic complaints, and if the effect were really only 5-15% it wouldn’t significantly affect the main finding of the paper by much.
If you can tell me what the other source is, I can look it up. I might have access ;)
It’s the same Framingham paper. I can only reach the abstract (my medical school doesn’t give me electronic access that far back) and I’m relying on reviews and comments to tell me what’s inside.
Atherosclerosis is mostly a nonreversible progressive disease that can start as early as in late adolescence, so it makes sense that hypercholesterolemia before age 50 is most important for its development. All it takes for a plague to rupture after tens of years of accumulation is that the endothelium covering it fails, it doesn’t necessarily have to grow anymore.
Right, this makes sense, I’m just saying it’s the opposite of what Framingham shows. Framingham says that “there is a direct association between falling cholesterol levels over the first 14 years and mortality over the following 18 years (11% overall and 14% CVD death rate increase per 1 mg/dL per year drop in cholesterol levels”
I hope I have provided a POV of how background knowledge can change the way we interpret study findings, and how much easier it makes sceptiscism about them
Yes, I agree with this. But the thesis of Good Calories, Bad Calories is that this allows enough degrees of freedom to be able to back up infinite amounts of confirmation bias. That is, if you see a study that supports your hypothesis, you say “Great, studies have proven we’re right!” And if you see a study that opposes your hypothesis, you say “in light of my background knowledge that my hypothesis is right, we can’t take this study at face value”, then seize on the first flaw you find in the study and use it to throw it all out. Kind of the “one person’s modus ponens is the other person’s modus tollens” thing people here keep talking about.
My new favorite study ever is the Biblically-named Lee, Lord, and Lepper, which asked people to evaluate the methodology of a study on the death penalty. They found that regardless of its actual methodology, if the experimenters wrote the conclusion such that it supported the opinions of the evaluators, the evaluators said it had good methodology. If the experimenters changed the conclusion so that it disagreed with the opinions of the evaluators, the evaluators were—surprise! - able to find a bunch of problems with the methodology and reasons why the study didn’t apply to anything.
I have no idea to what degree that’s happening in medicine these days; I’m really only beginning to seriously engage with the literature beyond a boring student level. I read Good Calories, Bad Calories on the advice of a bunch of other Less Wrongers, it was a really interesting book and has gotten me worried enough that I wanted to vent—as you pointed out, this wasn’t the best place for it and I apologize.
But I do think that further investigation beyond the level of just agreeing we can use background knowledge to interpret away findings is necessary at this point. At the very least, you have to admit the Cochrane review showing restricting dietary fat had no effect on anything means that something has gone atrociously wrong somewhere between what doctors say to their patients and reality (actually, I don’t know how that works in Finland; in the US dietary fat is a pretty big deal).
If I understand correctly, the original source is this. I can’t find any way to access it online, but everyone who talks about it says the results mentioned in that blog post and no one says otherwise, so I’m assuming they didn’t just make it up.
I’m impressed. I took a look at various sources, and it seems like you can’t even buy a used copy of that book anymore—it only exists in various academic libraries, and in many of them, it’s stored only as microfiche/microfilm.
If I understand correctly, the original source is this. I can’t find any way to access it online, but everyone who talks about it says the results mentioned in that blog post and no one says otherwise, so I’m assuming they didn’t just make it up.
Perhaps I don’t trust people enough on the internet. Especially when they’re trying to sell books of their own. See down below. Perhaps most of the people who talk about it haven’t seen it themselves, but are happy to spread the good message anyway. Perhaps the guy who sold it forged it. There are many points where someone could be lying, so you don’t have to blame a guy you like. There are some very dishonest people in my personal life so maybe this is a bias. It’s also pretty amazing to me that google gives only 563 results for “Framingham Diet Study”, because I guess this should be kind of a big deal. I’d really like to get my hands on it.
Right, but the studies on niacin and ezetimibe showed that they decreased cholesterol (ie were being used successfully and correctly) but failed to decrease cardiovascular endpoints.
I agree it’s very weird, but again the decrease in cholesterol not helping is hardly the only explanation for the results. Confounding factors are not fiction. Unfortunately I don’t have time to scrutinize them, as I’m not going to prescribe these drugs. (EDITED)
I see where you’re coming from. On the other hand, there is a $40 billion diet industry telling people to eat less fat and causing a spectacular amount of mental anguish around this idea (I don’t know if this is true in Finland; it’s definitely a big deal in America).
Living in Finland might indeed be the source of my confusion. Since most of the studies come from the US, I should update that this industry affects our nutrition more than I’d currently like to think. Nutrition here is a frequent topic, but excluding a couple of cases I’ve never come across the emotional turmoil that it seems to cause in americans. Isn’t there an industry pushing back too, and how are these people not just trying to make money?
You’re right that the measurements were taken within 24 hours, but I’ve heard this isn’t such a big deal, and according to the full-text version of the Fonarow study (sorry, didn’t find it until this morning) they agree with me. Also, if I’m reading their table right patients having acute coronary events had higher, not lower, cholesterol than those coming in with chronic complaints, and if the effect were really only 5-15% it wouldn’t significantly affect the main finding of the paper by much.
Thank you, definately saving them. Edit: The first study says “Although fasting blood samples were not mandated on the admission baseline test, subsequent samples collected on days 2 and 4 were in the fasting state. ” This means the baseline probably seems higher than it should. See the chapter “Limitations” in the Fonarow study. I can’t see how they agree with you (See Edit3 below, though).
The chronic cases receive statins after the disease has mostly developed so I guess that should explain the results. 30 % of the chronics received lipid lowering medication. 20 % of the admitted patients overall received lipid lowering medication. In the fifth table, lipid lowering medication is the strongest predictor of lower LDL. Edit2 This might also better explain why admitted patients overall have lower LDL than the general population, and can also be combined with the acute effect observed. It just means that patients at risk are properly recognized and treated, although probably too late.Edit3 Read more thoroughly and did some unit conversions from mg/dl to mmol/l, it doesn’t. However what is considered normal varies from risk group to risk group, and can be as low as 70 mg/dl. The authors of this study are supporting lowering the limits of what is considered normal, and I suppose you wouldn’t agree with this.
Framingham says that “there is a direct association between falling cholesterol levels over the first 14 years and mortality over the following 18 years (11% overall and 14% CVD death rate increase per 1 mg/dL per year drop in cholesterol levels”
In people over 50 years, if I understand correctly. Might be an english issue or then the issue is really with my understanding. So if people have more diseases (or they have progressed further) after 50 years that cause starvation (or some other mechanism that lowers cholesterol you can agree with) ie. IBD, depression, dementia, heart failure AND CVD mortality (edit), it makes sense that mortality follows falling cholesterol levels in that age group. Combine this with what I said before. As they speculate, “After age 50 years the association of mortality with cholesterol values is confounded by people whose cholesterol levels are falling—perhaps due to diseases predisposing to death.” They only ruled out CVD and cancer after all. I’d like to find the full copy for free because I have slight difficulty parsing the abstract. Of course it’s the study itself versus their speculation based on some “background knowledge” so I have to admit that your position is the stronger one here.
Yes, I agree with this. But the thesis of Good Calories, Bad Calories is that this allows enough degrees of freedom to be able to back up infinite amounts of confirmation bias.
Now I definately have to read this book! I promise.
At the very least, you have to admit the Cochrane review showing restricting dietary fat had no effect on anything means that something has gone atrociously wrong somewhere between what doctors say to their patients and reality
Our national guidelines already say that the effect of recommended diet (and exercise!) on cholesterol is tiny, but positive. After this discussion I’m even less enthusiastic about giving dietary advise. Concerning diabetes the guidelines might be even more wrong, but lets not open that can of worms this time!
And if you see a study that opposes your hypothesis, you say “in light of my background knowledge that my hypothesis is right, we can’t take this study at face value”, then seize on the first flaw you find in the study and use it to throw it all out.
Yeah, this culture is very prevalent in medicine, but of course varies from person to person. University hospital doctors are very status oriented people. I’ve read Heuristics and Biases, and as I understand knowing about biases doesn’t help much. Then again, I don’t think it’s only (dis)confirmation bias, but caution (a bias in itself), and I think everyone has difficulty differentiating these two. To be honest at this level of my education if I seriously start questioning my elders about everything I’ll be scared shitless in my daily work.
While time allows for additional studies, I feel like I can only concentrate on a few topics to master them. I was previously interested in a few fringes of neurology, mainly MS, but I’m starting to see that this is hardly rational from a utilitarian perspective.
the Cochrane review showing restricting dietary fat had no effect on anything
Restricting dietary fat keeping carbs and proteins constant (decreasing total calorie intake), or restricting dietary fat keeping total calorie intake constant (increasing carbs and/or proteins)? Or something in between?
If I understand correctly, the original source is this. I can’t find any way to access it online, but everyone who talks about it says the results mentioned in that blog post and no one says otherwise, so I’m assuming they didn’t just make it up.
I see where you’re coming from. On the other hand, there is a $40 billion diet industry telling people to eat less fat and causing a spectacular amount of mental anguish around this idea (I don’t know if this is true in Finland; it’s definitely a big deal in America). That eating less fat has literally zero effect on any outcome seems like something we should be trying to make more widely known, whether or not the small effect of changing fat types on events but not mortality holds up.
Apparently not; I know I’ve heard this but my attempts to Google a study in humans failed. I did find something in rabbits showing effects regardless of cholesterol, but even that wasn’t a direct comparison.
Right, but the studies on niacin and ezetimibe showed that they decreased cholesterol (ie were being used successfully and correctly) but failed to decrease cardiovascular endpoints.
You’re right that the measurements were taken within 24 hours, but I’ve heard this isn’t such a big deal, and according to the full-text version of the Fonarow study (sorry, didn’t find it until this morning) they agree with me. Also, if I’m reading their table right patients having acute coronary events had higher, not lower, cholesterol than those coming in with chronic complaints, and if the effect were really only 5-15% it wouldn’t significantly affect the main finding of the paper by much.
It’s the same Framingham paper. I can only reach the abstract (my medical school doesn’t give me electronic access that far back) and I’m relying on reviews and comments to tell me what’s inside.
Right, this makes sense, I’m just saying it’s the opposite of what Framingham shows. Framingham says that “there is a direct association between falling cholesterol levels over the first 14 years and mortality over the following 18 years (11% overall and 14% CVD death rate increase per 1 mg/dL per year drop in cholesterol levels”
Yes, I agree with this. But the thesis of Good Calories, Bad Calories is that this allows enough degrees of freedom to be able to back up infinite amounts of confirmation bias. That is, if you see a study that supports your hypothesis, you say “Great, studies have proven we’re right!” And if you see a study that opposes your hypothesis, you say “in light of my background knowledge that my hypothesis is right, we can’t take this study at face value”, then seize on the first flaw you find in the study and use it to throw it all out. Kind of the “one person’s modus ponens is the other person’s modus tollens” thing people here keep talking about.
My new favorite study ever is the Biblically-named Lee, Lord, and Lepper, which asked people to evaluate the methodology of a study on the death penalty. They found that regardless of its actual methodology, if the experimenters wrote the conclusion such that it supported the opinions of the evaluators, the evaluators said it had good methodology. If the experimenters changed the conclusion so that it disagreed with the opinions of the evaluators, the evaluators were—surprise! - able to find a bunch of problems with the methodology and reasons why the study didn’t apply to anything.
I have no idea to what degree that’s happening in medicine these days; I’m really only beginning to seriously engage with the literature beyond a boring student level. I read Good Calories, Bad Calories on the advice of a bunch of other Less Wrongers, it was a really interesting book and has gotten me worried enough that I wanted to vent—as you pointed out, this wasn’t the best place for it and I apologize.
But I do think that further investigation beyond the level of just agreeing we can use background knowledge to interpret away findings is necessary at this point. At the very least, you have to admit the Cochrane review showing restricting dietary fat had no effect on anything means that something has gone atrociously wrong somewhere between what doctors say to their patients and reality (actually, I don’t know how that works in Finland; in the US dietary fat is a pretty big deal).
I’m impressed. I took a look at various sources, and it seems like you can’t even buy a used copy of that book anymore—it only exists in various academic libraries, and in many of them, it’s stored only as microfiche/microfilm.
Perhaps I don’t trust people enough on the internet. Especially when they’re trying to sell books of their own. See down below. Perhaps most of the people who talk about it haven’t seen it themselves, but are happy to spread the good message anyway. Perhaps the guy who sold it forged it. There are many points where someone could be lying, so you don’t have to blame a guy you like. There are some very dishonest people in my personal life so maybe this is a bias. It’s also pretty amazing to me that google gives only 563 results for “Framingham Diet Study”, because I guess this should be kind of a big deal. I’d really like to get my hands on it.
I agree it’s very weird, but again the decrease in cholesterol not helping is hardly the only explanation for the results. Confounding factors are not fiction. Unfortunately I don’t have time to scrutinize them, as I’m not going to prescribe these drugs. (EDITED)
Living in Finland might indeed be the source of my confusion. Since most of the studies come from the US, I should update that this industry affects our nutrition more than I’d currently like to think. Nutrition here is a frequent topic, but excluding a couple of cases I’ve never come across the emotional turmoil that it seems to cause in americans. Isn’t there an industry pushing back too, and how are these people not just trying to make money?
Thank you, definately saving them. Edit: The first study says “Although fasting blood samples were not mandated on the admission baseline test, subsequent samples collected on days 2 and 4 were in the fasting state. ” This means the baseline probably seems higher than it should. See the chapter “Limitations” in the Fonarow study. I can’t see how they agree with you (See Edit3 below, though).
The chronic cases receive statins after the disease has mostly developed so I guess that should explain the results. 30 % of the chronics received lipid lowering medication. 20 % of the admitted patients overall received lipid lowering medication. In the fifth table, lipid lowering medication is the strongest predictor of lower LDL. Edit2 This might also better explain why admitted patients overall have lower LDL than the general population, and can also be combined with the acute effect observed. It just means that patients at risk are properly recognized and treated, although probably too late.Edit3 Read more thoroughly and did some unit conversions from mg/dl to mmol/l, it doesn’t. However what is considered normal varies from risk group to risk group, and can be as low as 70 mg/dl. The authors of this study are supporting lowering the limits of what is considered normal, and I suppose you wouldn’t agree with this.
In people over 50 years, if I understand correctly. Might be an english issue or then the issue is really with my understanding. So if people have more diseases (or they have progressed further) after 50 years that cause starvation (or some other mechanism that lowers cholesterol you can agree with) ie. IBD, depression, dementia, heart failure AND CVD mortality (edit), it makes sense that mortality follows falling cholesterol levels in that age group. Combine this with what I said before. As they speculate, “After age 50 years the association of mortality with cholesterol values is confounded by people whose cholesterol levels are falling—perhaps due to diseases predisposing to death.” They only ruled out CVD and cancer after all. I’d like to find the full copy for free because I have slight difficulty parsing the abstract. Of course it’s the study itself versus their speculation based on some “background knowledge” so I have to admit that your position is the stronger one here.
Now I definately have to read this book! I promise.
Our national guidelines already say that the effect of recommended diet (and exercise!) on cholesterol is tiny, but positive. After this discussion I’m even less enthusiastic about giving dietary advise. Concerning diabetes the guidelines might be even more wrong, but lets not open that can of worms this time!
Yeah, this culture is very prevalent in medicine, but of course varies from person to person. University hospital doctors are very status oriented people. I’ve read Heuristics and Biases, and as I understand knowing about biases doesn’t help much. Then again, I don’t think it’s only (dis)confirmation bias, but caution (a bias in itself), and I think everyone has difficulty differentiating these two. To be honest at this level of my education if I seriously start questioning my elders about everything I’ll be scared shitless in my daily work.
While time allows for additional studies, I feel like I can only concentrate on a few topics to master them. I was previously interested in a few fringes of neurology, mainly MS, but I’m starting to see that this is hardly rational from a utilitarian perspective.
Restricting dietary fat keeping carbs and proteins constant (decreasing total calorie intake), or restricting dietary fat keeping total calorie intake constant (increasing carbs and/or proteins)? Or something in between?