A guest post from medical Dr. Ben Knudsen and colleagues describing their peer-reviewed political analysis of policy statements from US professional medical associations
Of course medical organizations are partisan. They are analogous to labor unions and industrial organizations---organized to promote the interests of their members. No way should anyone pay any attention to their policies, recommendations, or guidelines. The people that control them think of things in a collectivistic manner and medical care is one of the most individualistic disciplines. It is a sure bet that even those organized with the most noble missions will soon be dominated by "leaders" and cabals pursuing their own goals.
Haven’t read the study but isn’t Chat-GPT trained on content produced by humans, so why wouldn’t it be similarly biased? I would want to see done type of demonstration that this method produces “neutral” evaluations.
At the end of the day when providing care to patients, physicians should be unbiased. It is not the time for us to “advocate” our political position to the patient when they are in our office seeking medical council. It’s ultimately the patient’s life and their own medical choices for their body. Therefore, medical advice should be neutral and unbiased. Informed consent should include ALL information (positives AND negatives) and not sway towards a physician bias. What we study in science should therefore be unbiased so that we may provide appropriate unbiased medical care. I have been seeing, in my experience, a profoundly liberal agenda in medicine that is pushed by physicians onto their patients, pushed into the scientific topics funded for research, and pushed into medical advocacy and medical boards. This needs to stop to let unbiased medical research and medical care prevail. The truth is not politically biased, and neither should science or medical care.
A policy of being "pro-trans" was probably considered liberal by this study. But it's hardly liberal if it considers gay boys to be trans girls that need a form of conversion therapy known as "gender-affirming" care.
I applaud Sensible Medicine for sharing articles like this. There has been left bias in academic medicine, and it has been risky to mention it. The emperor hasn’t been wearing clothes for some time now.
Very interesting work. The use of Grok as a “falsification” test was a great touch.
One weakness is the 3 human reviewers. I don’t know their biases. If using GPT was to reduce human bias, I would simply have reported the GPT results. If GPT was ambiguous, then consider it ambiguous rather than have 3 humans reclassify. I would be interested in the results strictly by GPT, and what fraction was clearly liberal and conservative and how those compare. And leave the “nuanced” and “neutral” categories unspoiled by human interaction. If it’s nuanced, maybe not great to make a bunch of hay out of that anyway.
Second, each policy statement has multiple parts, and in the end generates multiple recommendations. To take an entire statement and label it in whole in a particular political camp seems like a very broad brush approach. I would have taken each policy recommendation from each statement individually, and tabulate by “how many individual policy recommendations are liberal out of total number of recommendations made in all of the included statements from these 6 organizations”.
Third, the motivation for #2 was this from the paper: “However, given the emphasis on liberal values, such as individual professional judgment and less restrictive government intervention in professional practices, it was determined that overall, the policy statement was probably liberal.”. Individual professional judgement (autonomy) is “probably liberal”?!? What?!? If this recommendation is considered “probably liberal”, this then makes me wonder how the authors would label a policy statement espousing “patient autonomy”. And if THAT is considered “probably liberal”, then that would bring the relevance of this entire project into question, IMO.
In furtherance to that, I wonder if the statements considered to exhibit political bias were biased throughout? Or were they mostly anodyne recommendations with a couple of doozies that tipped them to one direction?
Surprise…surprise…NOT…Physicians are tied at the hip to the green back and that comes from big pharma and the ever ready tax payer plus those politicians passing more dollar programs that have not made us more healthy. The foothold of big pharma has ever expanded into the true earnings of the physicians and their organizations who too have NOT made us healthy as a nation. We must see just how much big pharma pays each physician…but that info is never allowed to be seen…why?? One peds clinic lost millions by allowing parents to decide whether to vac or not vac their kids.
I would add strongly that Big Pharma foothold is actually in the C-Suite with Administration and Lawyers. Both of these set physician pay. I get paid by the hour and take All comers in the ED. I’ll make it easy for you, BigPharma pays me ZERO. No kickback, no trip, no money, no dinner, no nothing-Thank God. Now take out the attorneys and the overpaid MBA and you’ll see healthcare cost come down and someone actually might get healthy in the meantime.
Great read! When the ACOG president came out and claimed "there are no reputable studies linking acetaminophen to autism" and AAP came out with a strong recommendation for Covid vaccines for kids within a day or two of a policy announcement from "conservatives", I couldn't help but wonder if their organization had looked at the evidence or if they just felt pressured to provide an opposite statement. For the Covid vaccine - hardly any other countries vaccinated pediatric populations so I feel the fda and cdc got it right. For tyelenol - there are plenty of observational studies finding a link and OBGyn often relies on observational data in its recommendations, so exactly what did the ACOG president mean when he said "reputable"?
This is an important topic, but the definitions and methodology are not up to the task. Better to take a small list of specific differentiating topics, identify partisan position statements on them that are broadly accepted by their own side (involving researchers from both sides to vet the choices for their side), and then ask the LLMs whether a given medical society policy statement aligns more with the vetted liberal or conservative position on that topic. Perhaps use a scale of 1 to 100, where 50 is completely balanced (e.g., it has some nuances supporting each) or is neutral (no nuances leaning in either direction). If it is a topic for which subsequent evidence has indicated that one of the two positions is much more likely to have been correct, particularly as evidenced by many people on one side changing their minds or one side simply no longer arguing their case and letting the opposing narrative take over, it would be great to see references to that subsequent evidence as well.
I thank Dr. Knudsen for his analysis. It is clear to anyone practicing in the field for any length of time that medical policy is highly politicized and biased towards liberal thinking.
So I have a pertinent question: you say that 33% of the articles demonstrated a bias (Liberal), but what about the other 66%? Drawing broad conclusions from a subset of data would suggest some potential for missing important nuance or other elements in play (or at least overlooked complexities with the conclusion). Wouldn’t it be more accurate to say that of the one-third of literature exhibiting bias, the predominant bias is Liberal; not that the larger body of scientific health articles are Liberal leaning? It is the language that is problematic, and can mislead someone reading the conclusion as simply, “Scientific literature is Liberal biased”. I do not have a metaphorical “dog in the fight”. I am interpreting, for myself, how the conclusion can cause a conservative human to arrive at an understanding that is not what was intended.
I agree and this is great example of cherry picking analysis plan/conclusions after the analysis. There are multiple ways you could analyze the data and make conclusions. And this is exactly why studies pre-register their outcomes and analysis plan!
The measure of “appropriate bias” should not be based solely on definitive evidence in favor of that bias, but rather the preponderance of evidence favoring one viewpoint over another. Positions based on relative strength of evidence are clearly preferred, as in many (most?) medical decisions.
The preponderance of evidence supports a “liberal” viewpoint when addressing public health measures. Societal level interventions have strong supporting evidence across a myriad of dimensions.
The most glaringly obvious unhealthy position by the medical community for years (since the advent of birth control pills/feminist movement) is that sexual promiscuity is an okay lifestyle—it’s actually promoted by doctors, and risks are rarely discussed. Yes, birth control pills help prevent pregnancies, but they do not protect from STD’s and of course we are seeing a surge of them. Adding to that unhealthy behavior is the emotional toll on numerous sexual relationships outside of marriage or thoughtful monogamy—let’s fact it, promiscuity is not healthy for mind, body, or soul. Sexual discernment is seen as prudish and based on religious teaching, therefore, it seems, doctors must be against it, even though they know full well it is an unhealthy lifestyle. Anything goes within a marriage between two virgins—there is no fear of “sleeping with all the people your partner has slept with.” But heaven forbid doctors discuss Godly advice with young people. You may not believe in God, but some advice is not debatable.
I know you mean well. But, I am a God believing physician and nothing in my practice promotes a promiscuous lifestyle. I will talk to any patient at any time on any subject. I speak truth and share Godly advice with everyone every day.
What a great article quantitating and confirming what many of us, which would undoubtedly include John Mandrola’s “neutral Martian”, have observed during the past several years. And it is not just the professional organizations listed but also includes certification bodies such as the ABIM using coercive methods to silence alternative opinion (loss of one’s board certification). And don’t forget prominent medical journals such as the NEJM (even Makary and Bhattacharya have proposed creation of unbiased journals). It is stunning how the malignant liberal bias cancer has metastasized throughout American medicine and created a loss of confidence and trust.
Of course medical organizations are partisan. They are analogous to labor unions and industrial organizations---organized to promote the interests of their members. No way should anyone pay any attention to their policies, recommendations, or guidelines. The people that control them think of things in a collectivistic manner and medical care is one of the most individualistic disciplines. It is a sure bet that even those organized with the most noble missions will soon be dominated by "leaders" and cabals pursuing their own goals.
Haven’t read the study but isn’t Chat-GPT trained on content produced by humans, so why wouldn’t it be similarly biased? I would want to see done type of demonstration that this method produces “neutral” evaluations.
Why not have humans read the statements instead of predigested AI summaries?
I can’t take this seriously.
At the end of the day when providing care to patients, physicians should be unbiased. It is not the time for us to “advocate” our political position to the patient when they are in our office seeking medical council. It’s ultimately the patient’s life and their own medical choices for their body. Therefore, medical advice should be neutral and unbiased. Informed consent should include ALL information (positives AND negatives) and not sway towards a physician bias. What we study in science should therefore be unbiased so that we may provide appropriate unbiased medical care. I have been seeing, in my experience, a profoundly liberal agenda in medicine that is pushed by physicians onto their patients, pushed into the scientific topics funded for research, and pushed into medical advocacy and medical boards. This needs to stop to let unbiased medical research and medical care prevail. The truth is not politically biased, and neither should science or medical care.
A policy of being "pro-trans" was probably considered liberal by this study. But it's hardly liberal if it considers gay boys to be trans girls that need a form of conversion therapy known as "gender-affirming" care.
I applaud Sensible Medicine for sharing articles like this. There has been left bias in academic medicine, and it has been risky to mention it. The emperor hasn’t been wearing clothes for some time now.
Very interesting work. The use of Grok as a “falsification” test was a great touch.
One weakness is the 3 human reviewers. I don’t know their biases. If using GPT was to reduce human bias, I would simply have reported the GPT results. If GPT was ambiguous, then consider it ambiguous rather than have 3 humans reclassify. I would be interested in the results strictly by GPT, and what fraction was clearly liberal and conservative and how those compare. And leave the “nuanced” and “neutral” categories unspoiled by human interaction. If it’s nuanced, maybe not great to make a bunch of hay out of that anyway.
Second, each policy statement has multiple parts, and in the end generates multiple recommendations. To take an entire statement and label it in whole in a particular political camp seems like a very broad brush approach. I would have taken each policy recommendation from each statement individually, and tabulate by “how many individual policy recommendations are liberal out of total number of recommendations made in all of the included statements from these 6 organizations”.
Third, the motivation for #2 was this from the paper: “However, given the emphasis on liberal values, such as individual professional judgment and less restrictive government intervention in professional practices, it was determined that overall, the policy statement was probably liberal.”. Individual professional judgement (autonomy) is “probably liberal”?!? What?!? If this recommendation is considered “probably liberal”, this then makes me wonder how the authors would label a policy statement espousing “patient autonomy”. And if THAT is considered “probably liberal”, then that would bring the relevance of this entire project into question, IMO.
In furtherance to that, I wonder if the statements considered to exhibit political bias were biased throughout? Or were they mostly anodyne recommendations with a couple of doozies that tipped them to one direction?
Surprise…surprise…NOT…Physicians are tied at the hip to the green back and that comes from big pharma and the ever ready tax payer plus those politicians passing more dollar programs that have not made us more healthy. The foothold of big pharma has ever expanded into the true earnings of the physicians and their organizations who too have NOT made us healthy as a nation. We must see just how much big pharma pays each physician…but that info is never allowed to be seen…why?? One peds clinic lost millions by allowing parents to decide whether to vac or not vac their kids.
I would add strongly that Big Pharma foothold is actually in the C-Suite with Administration and Lawyers. Both of these set physician pay. I get paid by the hour and take All comers in the ED. I’ll make it easy for you, BigPharma pays me ZERO. No kickback, no trip, no money, no dinner, no nothing-Thank God. Now take out the attorneys and the overpaid MBA and you’ll see healthcare cost come down and someone actually might get healthy in the meantime.
Great read! When the ACOG president came out and claimed "there are no reputable studies linking acetaminophen to autism" and AAP came out with a strong recommendation for Covid vaccines for kids within a day or two of a policy announcement from "conservatives", I couldn't help but wonder if their organization had looked at the evidence or if they just felt pressured to provide an opposite statement. For the Covid vaccine - hardly any other countries vaccinated pediatric populations so I feel the fda and cdc got it right. For tyelenol - there are plenty of observational studies finding a link and OBGyn often relies on observational data in its recommendations, so exactly what did the ACOG president mean when he said "reputable"?
If you want to see if ACOG is reputable or not, explore their abortion stance.
This is an important topic, but the definitions and methodology are not up to the task. Better to take a small list of specific differentiating topics, identify partisan position statements on them that are broadly accepted by their own side (involving researchers from both sides to vet the choices for their side), and then ask the LLMs whether a given medical society policy statement aligns more with the vetted liberal or conservative position on that topic. Perhaps use a scale of 1 to 100, where 50 is completely balanced (e.g., it has some nuances supporting each) or is neutral (no nuances leaning in either direction). If it is a topic for which subsequent evidence has indicated that one of the two positions is much more likely to have been correct, particularly as evidenced by many people on one side changing their minds or one side simply no longer arguing their case and letting the opposing narrative take over, it would be great to see references to that subsequent evidence as well.
Very well done! Here’s more on the outrageous pro sex slavery bias of the AAP:
https://gaty.substack.com/p/the-ones-who-didnt-walk-away-from
I thank Dr. Knudsen for his analysis. It is clear to anyone practicing in the field for any length of time that medical policy is highly politicized and biased towards liberal thinking.
Right--so says the lib
So I have a pertinent question: you say that 33% of the articles demonstrated a bias (Liberal), but what about the other 66%? Drawing broad conclusions from a subset of data would suggest some potential for missing important nuance or other elements in play (or at least overlooked complexities with the conclusion). Wouldn’t it be more accurate to say that of the one-third of literature exhibiting bias, the predominant bias is Liberal; not that the larger body of scientific health articles are Liberal leaning? It is the language that is problematic, and can mislead someone reading the conclusion as simply, “Scientific literature is Liberal biased”. I do not have a metaphorical “dog in the fight”. I am interpreting, for myself, how the conclusion can cause a conservative human to arrive at an understanding that is not what was intended.
I agree and this is great example of cherry picking analysis plan/conclusions after the analysis. There are multiple ways you could analyze the data and make conclusions. And this is exactly why studies pre-register their outcomes and analysis plan!
The measure of “appropriate bias” should not be based solely on definitive evidence in favor of that bias, but rather the preponderance of evidence favoring one viewpoint over another. Positions based on relative strength of evidence are clearly preferred, as in many (most?) medical decisions.
The preponderance of evidence supports a “liberal” viewpoint when addressing public health measures. Societal level interventions have strong supporting evidence across a myriad of dimensions.
The most glaringly obvious unhealthy position by the medical community for years (since the advent of birth control pills/feminist movement) is that sexual promiscuity is an okay lifestyle—it’s actually promoted by doctors, and risks are rarely discussed. Yes, birth control pills help prevent pregnancies, but they do not protect from STD’s and of course we are seeing a surge of them. Adding to that unhealthy behavior is the emotional toll on numerous sexual relationships outside of marriage or thoughtful monogamy—let’s fact it, promiscuity is not healthy for mind, body, or soul. Sexual discernment is seen as prudish and based on religious teaching, therefore, it seems, doctors must be against it, even though they know full well it is an unhealthy lifestyle. Anything goes within a marriage between two virgins—there is no fear of “sleeping with all the people your partner has slept with.” But heaven forbid doctors discuss Godly advice with young people. You may not believe in God, but some advice is not debatable.
I know you mean well. But, I am a God believing physician and nothing in my practice promotes a promiscuous lifestyle. I will talk to any patient at any time on any subject. I speak truth and share Godly advice with everyone every day.
What a great article quantitating and confirming what many of us, which would undoubtedly include John Mandrola’s “neutral Martian”, have observed during the past several years. And it is not just the professional organizations listed but also includes certification bodies such as the ABIM using coercive methods to silence alternative opinion (loss of one’s board certification). And don’t forget prominent medical journals such as the NEJM (even Makary and Bhattacharya have proposed creation of unbiased journals). It is stunning how the malignant liberal bias cancer has metastasized throughout American medicine and created a loss of confidence and trust.
Agree. Well said. It is a rampant cancer.
Well said.