02 Statin Relative Risk Reduction and Absolute Risk Reduction: Who's Right?
ฝัง
- เผยแพร่เมื่อ 23 ก.ย. 2024
- In this video, I discuss the difference between absolute and relative risk reduction and explain how each figure is misused by their proponents. The figures given by both camps may not apply to you.
--------------------------------------
DISCLAIMER: The following presentation and all presentations in this series are for informational purposes only and should not be used to diagnose or treat any medical condition. Please consult with your personal physician when deciding whether to treat or not treat any condition with pharmaceuticals or by any other treatment or therapy. In particular, never change dosage, stop taking medications, or change a therapy because of something presented in this series without discussing your options with your primary care physician or other health professional.
The author is not a medical professional and is presenting his own story, research, and opinions as they apply to himself, and himself alone.
If you think you may be experiencing a medical emergency, immediately call your doctor or dial 911.
-----------------------------------
The 36% Relative Risk Reduction figure I give is from one study; risk reductions vary from study to study.
TH-cam video with Dr. David Diamond, which includes his take on Risk Reduction: • David Diamond - An Upd...
While I credit Dr. Diamond and others for bringing the Relative Risk vs Absolute Risk difference to our attention, I believe the case about absolute risk reduction is overstated.
Related article:
www.researchga...
Article describing the Jupiter Trial:
www.ahajournal...
Note: I will be discussing Number Needed to Treat (NNT) in a later video.
Opening Music:
Music Title: Inspiring Optimistic Upbeat Energetic Guitar Rhythm
/ 62282987
Released by: Oleg Mazur / fm_freemusic
Music promoted by www.chosic.com...
Even with all of the very generous assumptions tilted toward statins, your analysis comes far closer to what in my mind would be truth than any of the analyses coming from the adamant statin proponents or from the dogmatic statin opponents.
What people viewing this video and commenting need to remember is that you point out multiple times in the video that the analysis you're presenting is a thought experiment that relies heavily on assumptions that may or may not be accurate and are not based on specific, quantifiable, verifiable evidence. They are simply to demonstrate an analysis process.
Any statistic that claims to represent future outcomes is based on probabilities, estimates, and assumptions. None of them are based on hard data about what will happen. At best they are based on hard data about what has happened. The past is not an accurate predictor of the future unless the entire future period and everything affecting the outcome is identical to conditions that existed in the past.
Even when evaluating past actual events in trial, researchers often apply assumptions about causes and effects, rather than actual measurements. One example of the way assumptions impact measured results is when a researcher identifies confounding variables that affect the reliability of the results. To adjust for the confounding variables, the researcher must make estimates of the impact it may have had. That estimation is made without having the ability to actually measure the effect of the confounding variable.
As for estimating my own long-term heart event risk, I'm not inclined to put much faith in any of the ASCVD risk calculators. At least not in relation to determining heart event probability due to cholesterol in order to justify prescribing statins. That is because, as you mentioned briefly in the video, the risk calculators seem to be more heavily weighted based on age rather than cholesterol levels. A statin is not going to prevent me from getting older. At my age, mid-70s, the ASCVD calculators give me a probability of between 18% and 22% for a 10 year CVD event, with only about 5% of that related to cholesterol levels. Based on mortality tables for my gender, age, and race, my overall probability of dying from something within 10 years is about 50%. Statins will provide me no benefit for most of those causes of death but have the potential to increase my probability of dying from a couple of them. Yet every time I go to the doctor I end up being told that I will surely die soon if I don't take statins to reduce my slightly high cholesterol levels.
Thanks for your thoughtful analysis. I found out recently from my wife (who shared the same doctor years ago) that even my statin-pushing doctor from the 2000-2010 time frame wouldn't push them on people over 70 for the very reasons you state.
Great presentation, lucid and comprehensible.
Hi Russ, have you ever discussed the fact that drug companies publicize drug benefits in relative risk, but they use absolute risk for adverse effects.
I'm aware of this but not sure I've addressed it directly. The closest I've come on the subject is probably this video (th-cam.com/video/cQbgDeFiezI/w-d-xo.html) where I go over how the info we get appeals to emotions rather than presenting real facts.
Thank you! The numbers are stretching my brain cells - that is a good thing.
Very few people know this. Most doctors don’t.
What I find interesting is the residual events rates from the statin group never get mentioned. They tend to run 4 times to 16 times the ARR. We should not ignore those who took the statin and suffered a cardiac event anyway.
I also wonder: suppose the RRR is 33% (to make the arithmetic easy). Then for every one person who avoids a heart attack, 2 still have one. But how do we know that some people who weren't destined to have an event, ended up having one actually caused by the the statin? Like instead of 1 person saved and 2 not, it was actually 2 saved, 1 not, and 1 person with an extra event? No way to tell, unfortunately, without a working time machine and the ability to run the experiment a second time.
@@mystatinfreelife You bring up an excellent point. And if we did run the experiment again, it may show different results. Which brings up another point. What if we gave both groups the placebo. Odds are one group will have more events than the other. So should the FDA approve placebo B because there were fewer events? Food for thought....and yes we need to pay attention to the residual event rate in the statin group. Unfortunately it gets swept under the rug.
Good, the argument for statins are even weaker if you were to show the pie as the total of all causes of death. The green is not 1/8th then but 1/8 * 36% = 4.5%.
Great point!
Interesting.
Thanks for watching.
The problem is how the risk calculations are made. Those models are useless
I describe it to people this way. The Absolute Risk Reduction is simply the difference in EVENT RATES between the two groups. That's easy. Absolute Risk is defined as "Your risk" and can be converted to Number Needed to Treat (NNT). Now your can calculate your odds of a win just like the lottery.
I think there are probably several mathematically equivalent ways to describe it. We should use whichever description seems to resonate best with whoever we're explaining it to.
@mystatinfreelife Please clarify the pie chart at 11:30. My best guess how you assembled it is as follows: Green = ARR=RRRxP(HA)= ~1/8 outcome take drug no HA
Red = (1-RRR) x P(HA) = 0.2176 ~ 2/8 (to nearest 1/8)
outcome take drug but HA
Yellow = 1-(1/8+2/8)= 5/8 outcome no drug AND no HA
They also lied about the side-effects of statins, claiming only 1-2% muscle-pain/myopathy while UK cardiologist said 29%.
Doctors hate to acknowledge or inform of side-effects of drugs that make patients hesitate to take them.
Isn't applying the RRR "across the board" a dangerous extrapolation?
It IS an extrapolation for sure, and all extrapolations are suspect. So I wouldn't take the conclusion as anything too precise - it is only as good as the assumptions and like you say (and I think I alluded to), it's not a solid assumption. Just something for the sake of argument. If anything it gives too much credit to the meds so it helps us define an upper limit so to speak.
Wake me when it is over. Good job though!
I've been improving (I hope) in getting more succinct. It takes practice though.
Hi Russ, how are you? Look, I really understood the point of your video, but you are missing some basic concepts. Relative risk is not data manipulation. A study will hardly last 30~40 years, so using relative risk is important, as it can capture an effect in a shorter period of time. From the moment a study is started and it is proven that the drug is effective, it is no longer ethical to continue the study, since half of the participants are receiving a placebo. In other words, there will be no controlled study lasting 30 years, since it is not right to let people receive a placebo for 30 years. In your calculations, you are using absolute risk data from 5 years and extrapolating to 30 years, which is wrong. I suggest you do the following: compare the absolute risk of quitting smoking for 5~10 years with the result of using statins for 5~10 years. You will see that quitting smoking for 5~10 years does not result in a high absolute risk, but the benefit of quitting smoking is extremely obvious. Compare this data and then see if you don't change your mind. Cheers!
Thank you for taking the time to comment. I understand what you're saying and for the researchers who understand relative risk vs absolute risk, as long as it is properly identified that's fine. The problem is, it is used in advertising and used to exaggerate the effects for purposes of persuading people to take the drugs, and doesn't give them honest risk assessment. My real point was to show that the people who claim a specific absolute risk are also incorrect. I address a lot of your concerns in the many videos I have. But there is no guarantee that the relative risk reduction discovered during a clinical trial is applicable across all risk levels or over all time periods either. Rest assured my decision to not go near statins is not really based on this data, it is from experience with adverse affects. As far as saying I was taking absolute risk data from five years and extrapolating to 30 years, I was very clear what my assumptions were and how I arrived at them and labelled it a thought experiment. Again, thanks for commenting.
@yuonutre Russ NEVER said it was data manipulation.
You know what is data manipulation? Reporting drug benefits in relative risk, but show adverse effects in absolute risk. That's what drug companies do.
Interesting that the "relative risk reduction" could be constant, while the "absolute risk reduction" is different for everyone. Seems backward.
Yeah, I thought of that while I was going over it. I really do appreciate that the difference is being pointed out by some, but they really miss the mark and the "absolutists" are being as disingenuous as the relativists (if they're doing it on purpose.)
@@mystatinfreelife What is most important is calculating one's own risk of CV event ie if one is older, has diabetes, and very high LDL cholesterol, his absolute risk of MI could be as high as 30%. In that case, using a statin can lower his risk significantly. I would certainly take the statin.
@theojones2593 That would be an absolute risk reduction of over 10%, which I agree is worth considering, as long as the person taking it is fully aware that this is the risk reduction, and not a guarantee, so that they also take other measures such as healthy diet and safe exercise.