Incentivizing these kinds of interventions, especially when supporting by a dodgy study, seems like a conflict of interest for doctor. Isn't this what we have blamed pharma for?
I listened to this podcast on a long drive and couldn't believe how quickly the hour passed. Excellent discussion on topics that extend beyond just this study.
You touched on an issue I ran into many times - what to do when you're almost to a goal (LDL, BP, etc) and have to decide whether to intervene further. It seems intuitive that adding a new medication just to squeeze out a few points is counter-productive, adding cost, side effects, compliance issues, etc for what appears to be little benefit. Add to that I know there's a margin of error on the lab instruments and the actual result might actually be at goal already.
Excellent insight from the perspective of the patient. Browbeating is a good term to describe patients grappling with standard guidelines while earnestly seeking best practices for personal health.
It should more and more obvious to everyone that Pharma doesn't know WTF it's doing. Almost every new drug it comes up with works by blocking an enzyme or enzymes. What could go wrong??!!
The intervention group was compared to "standard care," but the latter was not defined. Doesn't "standard care" per guidelines use a risk assessment calculator, especially at older ages? I agree with Matt that a 0.3% and 0.4% absolute risk reduction has questionable clinical meaning. Although you praised them for studying a policy, we don't know to what extent, if any, actual risk reduction strategies were implemented (various medications, lifestyle changes), and of what type and in what proportion. I suppose it doesn't matter if the policy is actually beneficial, but this is not highly beneficial, and what benefit it might have is not universal with all populations. It might be more useful to know to what extent the use of a risk calculator changes treatment and lifestyle. Also, not all risk calculators are equally valid for all populations. I also agree with Matt that, given the median age of patients, all cause mortality should have been an endpoint, since there are now quite a few studies demonstrating an inverse correlation between TC/LDL in those over age 60, perhaps due to the important role of LDL in adaptive immunity, and the increased risk of mortality from infection in the elderly. I would like to see more research into interventions targeted to affecting inflammation in a way that is not harmful to the elderly. Finally, I think if one pre-specifies a non-standard level of probability (i.e., p greater than .05), one should provide a well founded rationale.
Please follow the standard convention and spell out the name of an organization or condition followed by the initials in the first use of the abbreviation, CMS, TIA etc.
I thought about participating in this study but ultimately I decided that loss of time and irritation at filling out the paperwork (aka checking boxes in a government-run website) far outweighed either the remuneration or the benefit to medical science.
Some observations from one who uses the PCE risk estimator daily (and I promise not to bring CAC into the discussion) on why incentivizing providers to lower risk is problematic
1. Age is the primary driver of risk. It gets higher every year. That complicates "lowering" risk.
2. Obesity is the major factor driving DM and HT. Although I do my best at every visit with patients, most of them will not engage in lifestyle changes that improve DM and HT.
3. Socioeconomic status is a big determinant of obesity, dm, ht, compliance, etc. Providers can't control this.
I have a discussion with all my patients on "SPRINT" BP goals-benefits/risks. I think the data is solid on benefits of SPRINT BP goals but it requires a patient who is willing to frequently self-monitor BP and follows up closely and a physician who is willing to frequently look at BP data.
Since p-values are highly problematic and almost meaningless (and so are confidence intervals when used to see if zero is contained in them) why does it matter what threshold is used for them? Why is one interested in a null hypothesis in this situation? Do you really think it’s possible for the treatment to have zero effect to more than 10 decimal places?
I have side comment on the ACC risk calculator. The lifetime ASCVD risk jumps implausibly as a function of total cholesterol in a way which undermines the calculator's credibility. A 50 year old male with 110/70 blood pressure, total cholesterol 179, HDL 50, LDL 120, no other risk factors or statin has a lifetime risk of 5%. If you increase total cholesterol from 179 to 180 (a negligible increase) and keep all other inputs constant, lifetime risk suddenly jumps from 5% to 36% ! I have a hard time trusting a calculator which behaves like that. There's no way a result like that came out of any type of rigorous regression or other careful modeling on a large dataset. I wish they would fix it, because as it is right now, there will be some skeptical patients who will find that calculator and notice that weird behavior and might decide not to make needed lifestyle changes because they don't have confidence in the risk assessment.
If you look at it another way, in a study designed to back up the claims that aggressive treatment and payment can save lives:
1. There was virtually no difference in the proxy endpoints at all. .3%? Are you kidding? Statistical rounding error, and easily manipulated depending upon stuff that is beyond most of us.
2. Why not all-cause mortality? The gold standard for large studies. It is mentioned in the clinical trial but I can’t see it in the abstract and don’t have access to the full study. One summary claims all-cause mortality was “7% lower” in the intervention group, which is a tiny result, again easily manipulated statistically and meaningless.
3. The rest of the story, greater numbers of ER visits and hospital stays, suggests that the intervention group were worse off, not better off.
Bottom line: Treatment that is aggressive and systematic for CVD has virtually no effect on outcomes other than raising the number of times the patients has to endure ER visits and hospital stays.
Incentivizing these kinds of interventions, especially when supporting by a dodgy study, seems like a conflict of interest for doctor. Isn't this what we have blamed pharma for?
I listened to this podcast on a long drive and couldn't believe how quickly the hour passed. Excellent discussion on topics that extend beyond just this study.
You touched on an issue I ran into many times - what to do when you're almost to a goal (LDL, BP, etc) and have to decide whether to intervene further. It seems intuitive that adding a new medication just to squeeze out a few points is counter-productive, adding cost, side effects, compliance issues, etc for what appears to be little benefit. Add to that I know there's a margin of error on the lab instruments and the actual result might actually be at goal already.
Excellent insight from the perspective of the patient. Browbeating is a good term to describe patients grappling with standard guidelines while earnestly seeking best practices for personal health.
It should more and more obvious to everyone that Pharma doesn't know WTF it's doing. Almost every new drug it comes up with works by blocking an enzyme or enzymes. What could go wrong??!!
How much did this study cost?
The intervention group was compared to "standard care," but the latter was not defined. Doesn't "standard care" per guidelines use a risk assessment calculator, especially at older ages? I agree with Matt that a 0.3% and 0.4% absolute risk reduction has questionable clinical meaning. Although you praised them for studying a policy, we don't know to what extent, if any, actual risk reduction strategies were implemented (various medications, lifestyle changes), and of what type and in what proportion. I suppose it doesn't matter if the policy is actually beneficial, but this is not highly beneficial, and what benefit it might have is not universal with all populations. It might be more useful to know to what extent the use of a risk calculator changes treatment and lifestyle. Also, not all risk calculators are equally valid for all populations. I also agree with Matt that, given the median age of patients, all cause mortality should have been an endpoint, since there are now quite a few studies demonstrating an inverse correlation between TC/LDL in those over age 60, perhaps due to the important role of LDL in adaptive immunity, and the increased risk of mortality from infection in the elderly. I would like to see more research into interventions targeted to affecting inflammation in a way that is not harmful to the elderly. Finally, I think if one pre-specifies a non-standard level of probability (i.e., p greater than .05), one should provide a well founded rationale.
https://www.nature.com/articles/s44161-022-00049-1
Please follow the standard convention and spell out the name of an organization or condition followed by the initials in the first use of the abbreviation, CMS, TIA etc.
I thought about participating in this study but ultimately I decided that loss of time and irritation at filling out the paperwork (aka checking boxes in a government-run website) far outweighed either the remuneration or the benefit to medical science.
Some observations from one who uses the PCE risk estimator daily (and I promise not to bring CAC into the discussion) on why incentivizing providers to lower risk is problematic
1. Age is the primary driver of risk. It gets higher every year. That complicates "lowering" risk.
2. Obesity is the major factor driving DM and HT. Although I do my best at every visit with patients, most of them will not engage in lifestyle changes that improve DM and HT.
3. Socioeconomic status is a big determinant of obesity, dm, ht, compliance, etc. Providers can't control this.
I have a discussion with all my patients on "SPRINT" BP goals-benefits/risks. I think the data is solid on benefits of SPRINT BP goals but it requires a patient who is willing to frequently self-monitor BP and follows up closely and a physician who is willing to frequently look at BP data.
Since p-values are highly problematic and almost meaningless (and so are confidence intervals when used to see if zero is contained in them) why does it matter what threshold is used for them? Why is one interested in a null hypothesis in this situation? Do you really think it’s possible for the treatment to have zero effect to more than 10 decimal places?
I have side comment on the ACC risk calculator. The lifetime ASCVD risk jumps implausibly as a function of total cholesterol in a way which undermines the calculator's credibility. A 50 year old male with 110/70 blood pressure, total cholesterol 179, HDL 50, LDL 120, no other risk factors or statin has a lifetime risk of 5%. If you increase total cholesterol from 179 to 180 (a negligible increase) and keep all other inputs constant, lifetime risk suddenly jumps from 5% to 36% ! I have a hard time trusting a calculator which behaves like that. There's no way a result like that came out of any type of rigorous regression or other careful modeling on a large dataset. I wish they would fix it, because as it is right now, there will be some skeptical patients who will find that calculator and notice that weird behavior and might decide not to make needed lifestyle changes because they don't have confidence in the risk assessment.
If you look at it another way, in a study designed to back up the claims that aggressive treatment and payment can save lives:
1. There was virtually no difference in the proxy endpoints at all. .3%? Are you kidding? Statistical rounding error, and easily manipulated depending upon stuff that is beyond most of us.
2. Why not all-cause mortality? The gold standard for large studies. It is mentioned in the clinical trial but I can’t see it in the abstract and don’t have access to the full study. One summary claims all-cause mortality was “7% lower” in the intervention group, which is a tiny result, again easily manipulated statistically and meaningless.
3. The rest of the story, greater numbers of ER visits and hospital stays, suggests that the intervention group were worse off, not better off.
Bottom line: Treatment that is aggressive and systematic for CVD has virtually no effect on outcomes other than raising the number of times the patients has to endure ER visits and hospital stays.