I always enjoy reading Sensible Medicine (and Dr. Prasad's substack), it's consistently enlightening and well written, even if the details are beyond me sometimes.
You said to suggest trials/papers for you to comment on, how about this? I would like some epidemiologist type to look at the PMD 2.5 health risk literature and comment on whether this is real or exaggerated by the EPA. Recall that EPA did a series of experiments where they tried to cause health problems in subjects by putting them in a room and piping it lots of PMD 2.5. No effect, but the whole thing was pretty unethical.
What's the evidence that the EPA is basing regulations on? Any autopsy or necropsy showing PMD 2.5 effects? Probably PMD 2.5 isn't good for you, and probably not harmless, but we get exposed to lots of things in nature (not anthropogenic) that might not be all that great, but some people live a long time and that's been the case forever.
Re:statin/dementia- I think you're spot on about non-experts saying "this drug is BAD" and experts saying "This drug is AWESOME" Of course the former is not driven by that good good good pharma chedda but the latter is. So that's the explanation. The other is pharma-fatigue after C19 debacle. It's laughable now, but prior to C19 vax being mandated for my boys, I used to think FDA/CDC were the good guys. And pharma was advancing humanity. LOL. What a stupid and innocent take that was. There are always silver linings. C19 opened my eyes to the evil that is pharma. Teh corruption that is the gov't. Reducing cholesterol AT ALL COST despite the body needing passes the smell test (inc in dementia), and Sec Rubio can now run for the presidency in 2028. See? Always a silver lining.
Thanks, Vinay, for providing a common sense understanding of ridiculous studies. Med students need to be taught informed skepticism with examples like Zeke Emanuel's ice cream hypothesis. This kind of approach is vitally important. Why did so many pediatricians fall in line when so-called experts like Fauci supported shutting down schools and then reopening with ridiculous levels and kinds of barriers to keep our kids away from each other. That generation has millions of kids who will NEVER recover from the damage imposed on them. Pediatricians were almost entirely silent to their eternal condemnation. Physicians need to be responsible citizens, which is not taught in med school!
Regarding vaccines, there are some interesting findings. I agree an RCT is the clearest test, and a large Finnish trial of Shingrix versus placebo is now recruiting (NCT07502560). Until it reports, how should we weigh the evidence?
Eyting et al.’s Welsh natural experiment (https://doi.org/10.1038/s41586-025-08800-x) used a birthdate eligibility cutoff that produced a sharp rise in vaccine uptake; the design assumes other dementia risk factors vary smoothly across that threshold. Similar quasi-experimental findings have been reported in Australia (doi:10.1001/jama.2025.5013) and Canada (https://doi.org/10.1016/S1474-4422(25)00455-7), while a July 2026 English preprint (https://doi.org/10.64898/2026.07.20.26358345, not yet peer reviewed) found no evidence of a reduction in hospital-coded dementia.
These studies, and the Finnish trial, appear to rely on recorded dementia diagnoses. The Welsh design assumes that diagnosis ascertainment does not itself change abruptly at the eligibility cutoff. Eyting et al. also report a supporting analysis using death certificates, though that is not the same as directly measuring incident dementia. Randomization should help balance differences in care-seeking between trial groups. A diagnosis is still a meaningful clinical outcome, but a registry endpoint measures diagnosed dementia, not biological disease onset.
Where should we place quasi-experimental designs between conventional observational studies and RCTs? How much weight should we give the diagnosis limitation? How far can findings from live Zostavax be extrapolated to recombinant Shingrix?
This has a doozy of a title. I am aware of multiple companies working on building patient databases and calling them “external control arms,” presumably to speed up drug approval.
My sense is this data can and will be used to get drugs to market with poorer data, making control arms easier to manipulate to make an experimental agent look better. I’m not convinced it will help patients, but wonder if you can think of any good role for ECA’s in oncology.
Study I love? Or one I love to hate? How about EBCTCG ‘s meta-analysis published in The Lancet Aug 2025, “Extending the duration of endocrine treatment for early breast cancer…”? Small absolute effects on OS, side effects, plus the large percentage over time of patient non-adherence. My bigger concern is, Why are millions of ER+ breast cancer patients around the world given AIs when the benefit, it seems to a layperson like me, is so small? Your take? Thx.
I think this paper shows that extending therapy past 5 years has v tiny benefits and low compliance. I am actually not sure of global practice, but I would bet few global patients get more than 5 years. As far as papers go, I think it's fine. I disagree with the authors discussion
Love to read and watch all of your content Dr. Prasad. I have learned a tremendous amount from it. However, in my mind, there is a conflict in your statements. You state "When it comes to articles that claim ridiculous benefits of ivermectin or hydroxychloroquine, or fear monger about statins— you often find medical outsiders, people without epidemiology degrees, or folks generally not considered experts in that field."
But then later you note how one non-expert (Bill Maher) is more likely to be correct about ice cream than an expert (Zeke Emanuel, an oncologist).
Please help me understand why these statements/opinions are not contradictory. Once again, I appreciate all that you do to inform the public about health and science matters.
I would love Dr. Prasad to review ANY of the articles that talk about HRT causing cancer. Especially estradiol vs. conjugated estrogens causing breast cancer and micronized progesterone vs. medroxy progesterone acetate causing any cancer. Everyone is so divided about HRT it's hard to know what to believe. Also, blood clots. Everyone so scared to take oral estradiol but literally every female takes birth control pills which my understanding is causes more clots.
Clearly, it's from the field of veterinary medicine. I don't expect you to take this on, but I'm hoping you find it an interesting "challenge." This was big news in our field; urethral obstruction is a devastating yet very common condition. It is treatable. That said however, many owners cannot afford the ER care (depending on geography, managing a "UO" case is often several thousand dollars), and not uncommonly this results in humane euthanasia. Any advancements in treatment are another welcome tool in our arsenal. Although I won't claim to be the de-facto expert in UO management or the literature, this study seemed to me about as well-done as we ever get in veterinary medicine. I've prescribed Lorazepam for a few select outpatient cases based on this trial (typically, a UO case would be kept in hospital for ~1-3 days, depending on severity of illness on presentation; outpatient management is extremely discouraged, but sometimes pursued due to finances).
Completely understandable if you say "no thank you" to this!
Hello, Dr Prasad.
Consider to review this recent article:
doi 10.1056/NEJMoa2608397
It's on intermittent dialysis compared to continuous dialysis for ICU patients who have acute renal failure.
Thx
sincerely,
Eric P Cohen
Have a look into this - recent RACER trial in the Lancet comparing robotic arthroplasty vs. conventional instruments
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(26)00986-4/fulltext
Here's a study I would be curious to hear your take on. You've written about AI and I wonder what your thoughts are:
https://www.cambridge.org/core/journals/judgment-and-decision-making/article/bot-or-not-can-people-tell-the-difference-between-stories-written-by-a-human-or-by-an-ai-system/45E6DC0BB90AA648654D5AE243F6C667
Dr. Prasad, would love to hear your take on the HARMONi trial
Maybe too late to post this--
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2789423
This is a highly cited article in gender medicine.
Thanks either way!
I always enjoy reading Sensible Medicine (and Dr. Prasad's substack), it's consistently enlightening and well written, even if the details are beyond me sometimes.
You said to suggest trials/papers for you to comment on, how about this? I would like some epidemiologist type to look at the PMD 2.5 health risk literature and comment on whether this is real or exaggerated by the EPA. Recall that EPA did a series of experiments where they tried to cause health problems in subjects by putting them in a room and piping it lots of PMD 2.5. No effect, but the whole thing was pretty unethical.
What's the evidence that the EPA is basing regulations on? Any autopsy or necropsy showing PMD 2.5 effects? Probably PMD 2.5 isn't good for you, and probably not harmless, but we get exposed to lots of things in nature (not anthropogenic) that might not be all that great, but some people live a long time and that's been the case forever.
Thanks.
Re:statin/dementia- I think you're spot on about non-experts saying "this drug is BAD" and experts saying "This drug is AWESOME" Of course the former is not driven by that good good good pharma chedda but the latter is. So that's the explanation. The other is pharma-fatigue after C19 debacle. It's laughable now, but prior to C19 vax being mandated for my boys, I used to think FDA/CDC were the good guys. And pharma was advancing humanity. LOL. What a stupid and innocent take that was. There are always silver linings. C19 opened my eyes to the evil that is pharma. Teh corruption that is the gov't. Reducing cholesterol AT ALL COST despite the body needing passes the smell test (inc in dementia), and Sec Rubio can now run for the presidency in 2028. See? Always a silver lining.
Thanks, Vinay, for providing a common sense understanding of ridiculous studies. Med students need to be taught informed skepticism with examples like Zeke Emanuel's ice cream hypothesis. This kind of approach is vitally important. Why did so many pediatricians fall in line when so-called experts like Fauci supported shutting down schools and then reopening with ridiculous levels and kinds of barriers to keep our kids away from each other. That generation has millions of kids who will NEVER recover from the damage imposed on them. Pediatricians were almost entirely silent to their eternal condemnation. Physicians need to be responsible citizens, which is not taught in med school!
Regarding vaccines, there are some interesting findings. I agree an RCT is the clearest test, and a large Finnish trial of Shingrix versus placebo is now recruiting (NCT07502560). Until it reports, how should we weigh the evidence?
Eyting et al.’s Welsh natural experiment (https://doi.org/10.1038/s41586-025-08800-x) used a birthdate eligibility cutoff that produced a sharp rise in vaccine uptake; the design assumes other dementia risk factors vary smoothly across that threshold. Similar quasi-experimental findings have been reported in Australia (doi:10.1001/jama.2025.5013) and Canada (https://doi.org/10.1016/S1474-4422(25)00455-7), while a July 2026 English preprint (https://doi.org/10.64898/2026.07.20.26358345, not yet peer reviewed) found no evidence of a reduction in hospital-coded dementia.
These studies, and the Finnish trial, appear to rely on recorded dementia diagnoses. The Welsh design assumes that diagnosis ascertainment does not itself change abruptly at the eligibility cutoff. Eyting et al. also report a supporting analysis using death certificates, though that is not the same as directly measuring incident dementia. Randomization should help balance differences in care-seeking between trial groups. A diagnosis is still a meaningful clinical outcome, but a registry endpoint measures diagnosed dementia, not biological disease onset.
Where should we place quasi-experimental designs between conventional observational studies and RCTs? How much weight should we give the diagnosis limitation? How far can findings from live Zostavax be extrapolated to recombinant Shingrix?
I just want to change my vote.😉
https://ascopubs.org/doi/10.1200/JCO-26-01081
This has a doozy of a title. I am aware of multiple companies working on building patient databases and calling them “external control arms,” presumably to speed up drug approval.
My sense is this data can and will be used to get drugs to market with poorer data, making control arms easier to manipulate to make an experimental agent look better. I’m not convinced it will help patients, but wonder if you can think of any good role for ECA’s in oncology.
Great insights!
Study I love? Or one I love to hate? How about EBCTCG ‘s meta-analysis published in The Lancet Aug 2025, “Extending the duration of endocrine treatment for early breast cancer…”? Small absolute effects on OS, side effects, plus the large percentage over time of patient non-adherence. My bigger concern is, Why are millions of ER+ breast cancer patients around the world given AIs when the benefit, it seems to a layperson like me, is so small? Your take? Thx.
I think this paper shows that extending therapy past 5 years has v tiny benefits and low compliance. I am actually not sure of global practice, but I would bet few global patients get more than 5 years. As far as papers go, I think it's fine. I disagree with the authors discussion
Love to read and watch all of your content Dr. Prasad. I have learned a tremendous amount from it. However, in my mind, there is a conflict in your statements. You state "When it comes to articles that claim ridiculous benefits of ivermectin or hydroxychloroquine, or fear monger about statins— you often find medical outsiders, people without epidemiology degrees, or folks generally not considered experts in that field."
But then later you note how one non-expert (Bill Maher) is more likely to be correct about ice cream than an expert (Zeke Emanuel, an oncologist).
Please help me understand why these statements/opinions are not contradictory. Once again, I appreciate all that you do to inform the public about health and science matters.
I would love Dr. Prasad to review ANY of the articles that talk about HRT causing cancer. Especially estradiol vs. conjugated estrogens causing breast cancer and micronized progesterone vs. medroxy progesterone acetate causing any cancer. Everyone is so divided about HRT it's hard to know what to believe. Also, blood clots. Everyone so scared to take oral estradiol but literally every female takes birth control pills which my understanding is causes more clots.
This is a complete shot in the dark - wondering if you'd be comfortable doing an appraisal on this study.
https://avmajournals.avma.org/view/journals/javma/264/9/javma.26.01.0045.xml
Clearly, it's from the field of veterinary medicine. I don't expect you to take this on, but I'm hoping you find it an interesting "challenge." This was big news in our field; urethral obstruction is a devastating yet very common condition. It is treatable. That said however, many owners cannot afford the ER care (depending on geography, managing a "UO" case is often several thousand dollars), and not uncommonly this results in humane euthanasia. Any advancements in treatment are another welcome tool in our arsenal. Although I won't claim to be the de-facto expert in UO management or the literature, this study seemed to me about as well-done as we ever get in veterinary medicine. I've prescribed Lorazepam for a few select outpatient cases based on this trial (typically, a UO case would be kept in hospital for ~1-3 days, depending on severity of illness on presentation; outpatient management is extremely discouraged, but sometimes pursued due to finances).
Completely understandable if you say "no thank you" to this!
This study is practice changing for me as well. I would also prescribe lorazepam for cats in this situation as you do.