The GB study is not about whether the Hep B vaccine prevents hepatitis B. The authors want to know if giving this shot in addition to BCG and Polio at birth has a higher or lower all cause mortality than delaying it 6 weeks.

Read more

In two weeks, at the AHA meeting in Chicago, investigators will present the full results. The press releases describe decidedly null results for the overall trials.

• ZEUS: For the the IL-6 inhibitor ziltivekimab the primary outcome of CV death, MI, and stroke were almost the same. (HR 0.99; 95% CI 0.88–1.11). Notably, it did its job in lowering markers of inflammation, such as hs-CRP.

• HORIZON: For the Lp(a)-reducer pelacarsen, all we know is that trial failed to meet its primary outcome of CV death, MI, stroke and urgent coronary revascularization. It too reduced levels of Lp(a). We know from phase 2 trials that the drug reduced Lp(a) levels by 80%, though we don’t know what happened in HORIZON.

Both trials enrolled patients in whom you’d think the drugs would work. In ZEUS, patients had to have established vascular disease and chronic kidney disease and evidence of inflammation (hs-CRP ≥2). From the design/rationale paper, the baseline hsCRP was 4.5 mg/L. In HORIZON, patients had established vascular disease and high Lp(a) levels (≥ 70 mg/dl). From the design paper, the baseline Lp(a) level was 108 mg/dl.

The stakes for the makers of these drugs were high. A new inflammation-blocking drug would have been a big deal. Same for Lp(a); previous observational data had strongly correlated the genetically determined Lp(a) with the risk of cardiac events. A novel drug that lowers this particle might have made a lot of money.

The stakes weren’t just financial. Academically, there are many proponents of the inflammatory case for heart disease. While it is clear that inflammation plays some role in vascular disease, the search for a blockbuster anti-inflammatory agent has failed. Colchicine has shown only modest effects on cardiac events. Canakinumab had such a small effect, its makers did not pursue market release.

Many in the academy had high hopes for lowering Lp(a) as well. So the null findings from HORIZON will be a disappointment. Though there are other modulators of Lp(a) still in trials.

Beware of Subgroups

When there are both intellectual and financial interests in a drug, there may be a tendency to search for positive findings in these trials. For instance, the authors may say that in patients with the highest level of Lp(a) or hs-CRP, the drug looked better. One exception may be: HORIZON prespecified patients with Lp(a) ≥90 mg/dL as a specific group to look at, so a positive result there may deserve more weight than a typical subgroup.

Look no further than the drug sacubitril/valsartan in patients with heart failure and preserved ejection fraction (HFpEF). The pivotal trial, PARAGON HF, was null (HR 0.87, 95% CI 0.75-1.01), but the makers of the drug lobbied for approval in women and those with an LVEF less than 57% because those groups seemed to do better.

But these were two of many subgroups tested. I remain unconvinced that these are real. The main trial had a small (13%) reduction in the primary endpoint and it’s possible, even likely, that these results occurred by chance alone.

My simple understanding of subgroups are this: trials are powered for the full enrolled population. When we analyze a subgroup (say, male/female, old/young, diabetes/no diabetes), we compare smaller-sized groups. As such there are fewer events in these groups. Fewer events means a higher chance of detecting noise or missing a true signal. The other risk for detecting noise is multiple looks at the data. At a significant p value of less than 0.05, if you look at 20 subgroups, one would be expected to be positive by chance alone.

I discussed the classic example of detecting noise in subgroups back in 2022 when discussing medical vs N95 masks.

Richard Peto famously showed this in a landmark cardiology trial called ISIS-2, which found a benefit to aspirin after heart attack. Editors wanted to know which group had more or less benefit. Peto refused. But the Lancet forced him.

So, to make his point, he did an analysis of aspirin effects based on astrological sign. His team found that Libra or Gemini patients had no effect from aspirin, but all other signs had massive benefit. Peto explained in this paper:

To demonstrate the unreliability of subgroup analyses, these findings were subdivided into 12 subgroups according to the patients’ medieval astrological birth signs; the results in each were examined to find which of the 12 appeared least promising, and when just those with apparently unpromising results were collected together, it was ‘discovered’ that aspirin appeared totally ineffective for patients born under Libra or Gemini.

Peto described a few scenarios where subgroups may be helpful. Two factors are that both the subgroup and direction of effect should be prespecificed before the trial is done.

One example of a possible and plausible heterogeneous treatment effect comes from the DANISH trial of an implantable cardioverter defibrillator (ICD) or no ICD in patients with heart failure from non-ischemic causes. The main result was null, but there was an interesting subgroup wherein younger patients did better with the ICD and older patients seemed to do worse. This makes sense because an older patient would a) have more competing risks of sudden death and b) a higher risk of complications from the ICD.

While that subgroup finding makes sense (and the p value for interaction was positive), it’s still uncertain. We call such an observation “hypothesis-generating.” Meaning that to answer the question, we would randomize DANISH-like patients who were younger than age 59 to a trial of ICD vs no-ICD.

Another example of a subgroup finding that is likely true came from a study of renal function on statin effects. In this meta-analysis of 28 trials of statins, patients with the worst levels of kidney function seemed to garner no benefit from statins.

We can have confidence in this subgroup analysis because there are two trials of statins in patients with advanced kidney disease (4D and AURORA) and both were null.

Summary

Peto argued that the safest practice would be to apply the main results effect size to any subgroup.

That said, I could find other examples where a subgroup effect might be real. The emphasis is on might, and the proper practice is to use a subgroup finding to plan another trial.

When the null trials are presented in two weeks, be alert to the pitfalls in placing too much weight on subgroups—especially since the main results of the trial were null.

Subscribe now

Patients ask the best questions.

My patient asked me this during an adult ADHD evaluation. He was a pretty typical patient. He had made his own diagnosis. He looked up the diagnostic criteria and realized that he fit ADHD to a tee. He came in to talk with a doctor about whether he really had ADHD. I fumbled through an explanation about the metaphysics of disease, and my patient interrupted me to ask, “So do I need Adderall?”

Patients ask the best questions.

Whether ADHD is a disease is a great question. This discussion starts with our lack of a consensus definition for disease. The best definition I can come up with is that a disease is a biological explanation for a patient’s symptoms.

In the case of ADHD, we have not identified a precise biological mechanism. There is no diagnostic blood test or MRI finding that confirms the diagnosis. We make the diagnosis with a subjective checklist of items such as losing your car keys or easily getting distracted. Patients usually come to my clinic already having made a self-diagnosis. I am not sure if ADHD is a disease. I am not even sure what a disease is. However, the practical question is whether the label helps us care for a patient.

There are many reasons to evaluate a patient for ADHD. Some patients learn that they do not have ADHD but are not sleeping enough. Others receive a diagnosis of ADHD and feel better just because they have an explanation for their symptoms. Overwhelmingly, the question in my clinic is whether patients need stimulants.

I appreciate the controversy around prescribing stimulants. On one hand, prescribing stimulants can make us seem like we are medicalizing the normal human condition and/or acting as pawns of Big Pharma. On the other hand, not prescribing them is undertreating a stigmatized condition. As always, the truth is probably somewhere in between.

What does the evidence say about stimulants?

A 2025 meta-analysis of over 100 RCTs summarizes the data. It reports a reproducible and clinically meaningful improvement in ADHD symptoms during the initial couple of months of treatment. Another meta-analysis tells us that cardiovascular harms are clinically insignificant increases in blood pressure and heart rate.

There is much less robust evidence for the long-term effects of stimulants. Few RCTs follow patients longer than a couple of months. The longest RCT found that stimulants continue to improve symptoms at 1 year. However, the primary outcome for this study was symptom score at 12 weeks, and over 40% of patients had dropped out at the 1-year mark. The most rigorous long-term safety study found a statistically significant increase in cardiovascular disease, although the odds ratio was <2. This study did not find an association between stimulants and arrhythmias, heart failure, thromboembolism, and ischemic heart disease. As with any observational study, there are likely to be confounding factors. For instance, patients with early cardiovascular side effects may stop the medication before developing clinically significant disease.

I generally tell my patients that the medication guidance is that the medications can provide benefit, especially in the first few months. While we do not have a complete understanding of the long-term effects on cardiovascular health, these risks are likely to be small. If patients meet the subjective diagnostic criteria for ADHD, whether or not to use medications is a perfect time for shared decision-making.

How severe are the symptoms?

Are you forgetting tasks at work that put you at risk of getting fired?

Do you constantly get into fights with your spouse because you aren’t keeping up with the laundry?

Would an “N-of-One Trial” make sense here?

After this sort of discussion, he decided to try a stimulant. When I saw him back at 4 weeks for follow-up, he was feeling a little more productive and was not having any side effects. We decided to continue the medication at the same dose. We will keep checking in to see if the medication continues to help.

At the end of our visit, I asked him if he thought ADHD was a disease. He wasn’t sure.

Ken Noguchi, MD, PhD, is a family medicine doctor and an assistant professor in the Department of Family Medicine and Community Health at the University of Wisconsin School of Medicine and Public Health.

After a few weeks, though, maybe there is something to add: a discussion of the data we have, where it falls short, and the data that we need.

The study that would prove that the Galleri test is useful is the same study that we should demand for every screening test. People who are at risk for the disease (or diseases) in question are randomized to screening or no screening. The patients are then followed for a prespecified number of years, and then overall mortality for the two groups is compared. If the screened patients die less — enough less that the lives saved are worth the cost — then we adopt the screening test. If the mortality rates are the same in the two groups, we don’t. For Galleri, a “multicancer screening test”, Mandrola and I would not even need to argue about overall vs. cancer-specific mortality.

Although the study we need is simple, there are at least five reasons that it is hard to run.

1. When we are looking at a population-based cancer screening test. This means that the incidence of cancer will be low, so the study will need to be somewhere between large and enormous.

2. We will need to design this study as an “intention to screen” analysis because not everyone to whom we offer the test will follow through with screening. This means the study will need to be closer enormous than to large.

3. Because the lag time between a test finding an early cancer (or precancerous lesion) and a person dying from that disease is long, the large study must run for years.

4. We will be enrolling healthy people in our study so the risk that we harm more people than we help will be high. We will need to delineate the sensitivity and specificity of the test, and quantify the harms of overtesting, overdiagnosis, and overtreatment.

5. Lastly, because our treatment for cancer is always improving, and screening is less important for treatable cancers, our study will have to be occasionally re-run because evidence should come with an expiration date.

This study design is key because we often forget that the goal of screening is not to find more disease or to find it earlier; it is to decrease morbidity and/or mortality from the screened disease. The design described above avoids the biases associated with finding disease early without improving outcomes: lead time bias, length time bias, and overdiagnosis.

Lead time bias occurs because people with disease detected by screening will have longer survival “with the disease” than those whose disease came to attention when they developed symptoms. This is true even when screening does not decrease mortality. Finding a disease earlier but not changing the outcome is harmful.

Length time bias occurs because screening tests detect more indolent subtypes of cancer, cancer that is present in a population for long periods of time. Thus, people whose cancer is detected by screening do better/live longer, on average, than people whose cancer is detected when they become symptomatic. People don’t benefit from this detecting the “less bad” subsets of disease.

Lastly, screening tests risk overdiagnosis. Screening seems effective because we detect tumors that would never have become clinically apparent. This Bleyer and Welch NEJM article from 2012 does an excellent job explaining overdiagnosis as it pertains to breast cancer screening.

Let’s get back to Galleri.

Grail’s Galleri test is a single blood test that screens for multiple cancers by assaying for cell-free DNA, or other circulating substances, shed by tumors. Early studies focused on the test's characteristics and were disappointing. When specificity was set near 100% (to minimize false positives), sensitivity only exceeded 50% for later stage (stage 3) cancers. (See figure 3 here). This means the test is mostly good at detecting cancer when it is clinically detectable and/or when we are not very good at treating it.

The latest study looks at the Galleri test's ability to find cancer and shift the stage of diagnosed cancer, over a short period of time, from high to low. In this 3-year, 140,000-person study, Galleri failed. The primary endpoint was a shift in the incidence of stage III or IV cancers between the screened and non-screened groups. There was no difference: incidence rate ratio, 1.03; 0.92 to 1.14, P=0.63.

So as not to come off as a total buzz-kill, there are some itzy-bitty reasons to be optimistic. If you look at the curves, you could almost imagine that, years into the future, the incidence of stage III and stage IV cancers will fall in the screened group compared to the not-screened group.

BUT! Remember that this has not yet happened in a trial of 140,000 people over 3 years. Any benefit we eventually see, if we see one, is likely to be small and take a long time to accrue. And even if it does, that still does not mean this test saves lives. There are plenty of reasons a screening test that slightly decreases the number of stage 3 and 4 cancers in a trial population does not save lives at a population level, let alone does so at a remotely cost-effective rate.

There will be anecdotes, true anecdotes, of people whose lives might have been saved by Galleri; people whose pancreatic cancer was caught early by the test and survived.

The real question is, does the test save lives and, if it does, what are the costs, financial and otherwise?1 Eventually, we will need to ask these questions unless we want to keep turning over public dollars to private companies and get nothing in return, much like we did with Hoffmann-La Roche/Genentech/Gilead for Tamiflu or with Pfizer for Paxlovid.

Share

The jury rejected claims that Sabatini sexually harassed Dr. Knouse, and ruled that she defamed him— costing him professionally- by falsely claiming he groomed her, or that she was coerced into a sexual relationship. Instead, the jury appears persuaded by Sabatini, who argued she made the complaints to punish him for finding a new partner. These charges led HHMI to fire Dr. Sabatini, and eventually paved the way for his resignation from the Whitehead.

MIT and the Whitehead were previously removed from the litigation, and it is unlikely further action can be taken against them.

What is the lesson?

The lesson is that universities should be careful to avoid feeding into the online pile-ons that happen on X. This time, the university escapes unscathed, but the verdict shows that the judicial system can render a result that is devastating in cases of mass hysteria and pile-ons without careful and thoughtful consideration of the facts. MIT and the Whitehead have escaped this time, but I suspect a future controversial firing may ensnare the university and its principals.

The next case to follow will be Norman Wang— demoted at UPMC because he disagrees with affirmative action.

Catheter ablation for symptomatic atrial fibrillation (PVI-SHAM-AF): a randomised, double-blind, sham-controlled, multicentre trial

• This Week in Cardiology with John M. Mandrola MD

• Cardiology Trials

And a special treat/rescouce we have covered nearly 100 articles at this point on the podcast. Here is an active list of all the articles we have discussed. Take a look, read the articles, and refer to our conversation if you wish!

Share

Read more

Then I remembered that an FDA advisory panel voted 10-0 on safety, 6-4 on efficacy and 7-2 on benefits greater than harms.

More data is being presented this week at the European Society of Medical Oncology meeting.

I searched Sensible Medicine and found that we have not done a formal review. So I went and looked at the September 2026 NEJM publication.

Here is what I learned

NHS Galleri

This was a blinded RCT in 8 English cancer alliance regions. Blood was obtained from adults 50-77 years of age at up to three annual visits. About 71,000 adults were randomly assigned to have the blood tested with the multi-cancer test and another 71,000 had their blood stored.

Trial enrollees had to be cancer-free in the previous 3 years and not be undergoing tests for anything suspicious. Data analysis was conducted independently by Queen Mary University of London and verified by Grail. By design, participants from the most socioeconomically deprived areas were overrepresented.

The primary endpoint was the incidence of stage 3-4 cancer across 12 types of cancer after 3 rounds of tests and at least 12-months of follow-up. A key secondary endpoint was stage 4 cancer in the 12 types. The trial was not powered for mortality. Grail funded the trial and helped draft both papers.

Results:

For the primary endpoint, 706 participants in the intervention group and 688 participants in the control group had received a diagnosis of stage III or IV cancer among the 12 prespecified cancer types, which yielded an incidence rate per 100,000 person-years of 300.5 in the intervention group and 292.5 in the control group. This was not significantly different at an incidence rate ratio, 1.03; 95% CI, 0.92 to 1.14; P=0.63).

For the key secondary endpoint of stage 4 cancers, the incidence rate ratio (0.86 (95% CI 0.74–1.00) favored the intervention arm but had an upper bound of the confidence interval of 1.0. Formal statistical testing was not done because the primary endpoint was not significant.

The authors looked at cancers that did not have a screening test available: 1.05 95% CI 0.96-1.14.

The early-cancer numbers are revealing. The screened group had more early-stage cancers diagnosed: 647 vs. 559 among the 12 prespecified types (RR 1.16, 95% CI 1.03–1.30, a post hoc analysis). Yet there was no reduction in late-stage cancers, the primary endpoint. There are two explanations. One is overdiagnosis: extra cancers found that would never have caused harm. The other is that screening simply moved some diagnoses earlier in time, and the matching cancers in the control group will surface with longer follow-up. Either way, earlier diagnosis has not yet translated into fewer late-stage cancers. The authors say they cannot distinguish these possibilities, though they argue overdiagnosis is biologically unlikely.

About 1% or 1800 adults in the intervention arm had a positive test for cancer. Of these, cancer was found in about half (52%). [I will come back to this.]

A follow-up Nature Medicine paper described the test characteristics.

The test has a strong specificity of about 99.5%, meaning few cancer-free people test positive (≈ 0.45%). But it had a terrible sensitivity of only 30%, meaning it missed a lot of cancers.

A few summary statements from the Nature Medicine paper:

Over three annual visits, the test caught about a quarter of all cancers, and about 45% of the 12 high-mortality ones. But about 1% of cancer-free people had a false positive test.

You may be drawn to the low false positive rate of the test. A specificity of 99+% seems good on the surface.

I like to think of tests with likelihood ratios, which are how much a positive test increases the odds of a disease over baseline and vice-versa, how much a negative test reduces the odds of having the disease.

The positive likelihood ratio is the sensitivity (≈ 0.30) divided by the 1-specificity (.005), which is roughly 69. Anything over 10 is considered strong.

The negative likelihood ratio is 1-sensitivity (0.70) divided by the specificity (.995), which is roughly 0.70.

Now apply these numbers to people age 50-77 who might get the test. A positive result raises the baseline risk of 1.5% of having cancer to 52% (48% would be a false positive). A negative result however does little to the baseline risk of 1.5%, lowering it to only about 1.1%.

This is not good.

It means that if you have a positive test, it’s almost as likely to be a false positive as a true positive. And if you have a negative test, you are not much less likely to have cancer as you were before the test. If a younger person gets the test and their cancer risk at baseline is 0.5% instead of 1.5%, the likelihood ratio only changes it to 26%, so here, there is a 3-fold greater chance that the test is a false positive.

Summary Comments:

I do not understand the FDA advisory committee recommendations. In the randomized controlled trial there was no difference in the primary endpoint of stage 3-4 cancers. The primary endpoint failed to show a benefit. The slightly positive secondary endpoint is not persuasive.

In the Nature Medicine paper, the test characteristics are not great. A positive test leads to as many false positives as true positives, and a negative test is not at all reassuring.

Imagine being a person in the group where your test was positive, but no cancer was found. Now what? How many more tests do you have to get before being reassured?

It would likely be worse in the US. Here, every time a US patient has a test, there is chance of finding an incidental finding, which leads to more tests and more risk of medical harm. Notable is that the trial, as large as it was, did not systematically look at harms once participants were referred for diagnostic workup.

I looked up the Grail test and found that it is available for about a $1000. If FDA approved, insurers (aka you and me), will be asked to pay for it. This is clearly a bad thing.

Let me know if I have anything wrong here.

Subscribe now

In 2009, the Institute of Medicine, now the National Academy of Medicine, defined conflict of interest “as circumstances that create a risk that professional judgments or actions regarding a primary interest will be unduly influenced by a secondary interest.” The IOM focused on financial COI, although since then additional attention has focused on non-financial COI, which is often defined as a particular relationship that might influence a decision. Bero and Grundy have argued that it is important to distinguish between non-financial COI and interest alone, arguing that, for example, personal experiences, religious background, or training are interests but not non-financial COI, because they can never be eliminated. The distinction between interest and conflict of interest is challenging whether it be selecting grant reviewers, guideline panel members, or editorialists. For example, whether anyone who has expressed strongly held, or fixed beliefs should ever serve on a guideline panel seems inconsistent with a fair assessment of evidence.

Over the past few decades, there has been a change in how academic institutions, societies, and journals deal with COI. They have moved toward greater transparency in requiring the reporting of financial COI but have moved away from requiring disclosures of the amount of money (or stock, or other financial compensation) that is involved in various relationships. Many individuals have a standard set of disclosures which they use for all required COI statements, regardless of whether they are true conflicts in each specific instance. This renders such disclosures less meaningful. Two recent examples, one an opinion piece and one a research report, highlight how difficult it is to understand COI statements, with authors in both papers reporting dozens of relationships, making it difficult to determine those that are relevant.

One of the challenges for journals is whether the common approach that is now used – a list of all potential COIs, which can often be long, with no details about why a particular relationship is a COI or the amount of money involved - serves the public good. Speaking and consulting fees can be in the tens of thousands of dollars, and stock options even more, although their value can be very difficult to estimate. We believe that most clinicians and the public would want to know about and would distinguish between a $25 lunch, a $10,000 speaking fee, and $100,000 in stock options. For authors of guidelines, besides ongoing consultant relationships, whether a company or organization has supported or is supporting the research of an author should also be declared. Although this support may not directly impact the salary of that author, any research support is financially beneficial.

In scientific journals, who supports the study is usually reported, but who has control of the data and conducts the analysis is often less clear. We applaud journals that have added a detailed description of many of these issues to their methods section. For example, the role of the study sponsor in design, data collection, analysis, writing the report, and the decision to submit for publication are now required and included by some journals in the methods section of manuscripts. We believe that particularly for RCTs of drugs and devices, which often change practice and fuel the rise in health care costs, other journals should follow this model.

There can be COIs for clinicians who have relationships with drug or device manufacturers, or own, or have consultant arrangements with ambulatory services, like radiology, emergency medicine, and orthopedics. What should they tell patients? For example, if a physician recommends a particular device (or their institution uses a particular device) should that person disclose that they have a financial relationship with that company? If an institution manufactures CAR T-cells, and that is recommended as treatment, should that be disclosed? Virtually no disclosures are ever made in clinical care, with patients trusting that clinicians (and their institutions) will make decisions that are in the best interest of patients. If the institution has safeguards in place – for example, clinicians with a financial relationship with a particular drug or device manufacturer cannot decide which drug or device is used at their institution, or if treatments that are recommended directly benefit the institution, there is an independent review of the treatment -- then we are not sure if any additional disclosure is necessary. That said, it is likely that some patients would want to know about such COI.

One additional area deserves comment. Many journals now have regular podcasts, often with an editor interviewing an author, or a well-known physician, scientist, or policymaker. Yet, we are unaware of any podcast hosted by a journal that reports COI. Given the increasing and ubiquitous nature of podcasts, “published” and distributed by journals that maintain careful COI policies for written material, we believe it is time for them to consider COI policies for podcasts. Given the various possible COIs of both editors and the people they interview, listeners, similar to written articles that are published, hosted, and distributed by journals, should be aware of COIs.

The world of COI will remain murky – it is all around and comes in many sizes and shapes. It is not clear to us that the current approach – full disclosure, but without the type and amount of the relationships – enables public trust. Perhaps it is time to re-introduce financial categories of support (1 -<10K; 2-10-25K; 3->25-50, 4 >50-100K, 5 >100K). At a time of waning trust in medicine and science, a renewed effort to improve meaningful financial disclosure is necessary.

Share

Declaration of COI:

Bauchner – Consultant to Vivli (Data Sharing Platform, Category 1); Medeloop (Data Repository and AI Agentic, Category 2); Paper-Wizard (AI Review of Manuscripts, Category 2); CBMRT (Null-Hypothesis Champion, Category 2)

Rivara – Funding from NIH; consultant for Career Physicians (Category 1)

Howard Bauchner, MD, is a Professor of Pediatrics and Public Health at the Boston University Chobanian & Avedisian School of Medicine. He is also a Visiting Scholar at the National University of Singapore. Dr. Bauchner is also the former Editor-in-Chief of JAMA and the JAMA Network (2011-2021) and Archives of Disease in Childhood (2004-2011).

Frederick P Rivara, MD, MPH, is Professor of Pediatrics and Adjunct Professor of Epidemiology at the University of Washington. He is also the former Editor-in-Chief of JAMA Pediatrics (2000-2017) and JAMA Network Open (2018-2024). He continues as an active clinician, mentor, and investigator.

Photo by Arnold Antoo on Unsplash

When I last saw RP, he was 87. He suffered from an impressive list of medical problems: diabetes, obstructive sleep apnea, hypertension, atrial fibrillation, and restless leg syndrome. His most troublesome problem, however, was a broken heart. His wife, also a long-term patient of mine, had died of pancreatic cancer after years of debilitating illnesses. He had never really recovered.

RP’s health was good at this final visit, so we spent a few minutes talking about my new car, a diesel VW Jetta.1 He had spent years as a diesel engineer and was excited to hear what I thought.

Five years later, when a similar car rumbled by me, I thought of RP.

I have many patient panels. These days, most exist in the electronic health record (EHR). An automatically generated “My Patients” list contains all patients in my practice. I’ve also created a couple of my own lists. One is “INR,” a list of my patients who take the blood thinner warfarin and need their international normalized ratio checked every few months. There is also “Enigmas,” a list of patients I have seen with another doctor whose presentations confused us and whose evaluations I want to follow.

Then there is a panel that exists not in the EHR but only in my mind. This is the list of patients I am currently worried about. I can only fully enumerate the patients on this list at about 4:00 AM.

And then there is the ghost patient panel.

This panel is made up of people who used to be my patients but no longer are — people who left my practice without telling me. They might have fired me, lost or changed their health insurance, moved, or died without my knowledge. If I knew who was on this panel, I could learn details about their lives, my practice, my skills as a doctor, and even how we care for our loved ones. This list is ghost-like because its members can never really be known.

Some patients join the ghost panel intentionally, some willingly, some unwillingly, and some unwittingly. This panel speaks to our society's mobility, the care of older relatives, and how little feedback doctors get on the care we deliver.

A patient joins the ghost-patient panel when I realize I no longer see them. This can happen while I'm watching my daughter’s track meet, swimming, or even documenting another patient’s visit. After remembering the patient and adding him or her to the panel, I next consider why they might have left. Death, of course, is the most dreaded cause of ghosting (and a cause that makes the title of this reflection a bit unseemly). The oldest people have probably died. Recalling them makes me feel sorrow and guilt — sorrow for the loss; guilt because I often realize their absence months or even years after their passing. I also think that these were strange relationships, usually ones in which I was close to the patient but not to their family, who would have alerted me to their passing.

I expect that RP falls into this group. We had a very nice relationship and usually agreed on how to manage his medical issues. I enjoyed his company, and he seemed to enjoy mine. We had shared the difficult experience of his wife’s decline. He was on Medicare, so insurance changes would not have affected him. I imagine that he either died or moved closer to his children. 

Health insurance instability causes patients to ghost me. I am aware of some insurance-based separations; patients alert me as they try to navigate the healthcare system. Most patients, however, accept that nothing can be done to retain their doctor and just silently move on. Some of these patients un-ghost at age 65, claiming actual bodies (sort of like Voldemort) when they enroll in Medicare.

Some patients in the panel fired me. Said more gently, they leave my practice because they are unsatisfied with my care. It is no surprise these patients disappear. No doctor is right for every patient, and even when I provide the care a patient needs, I often don’t provide the care that he wants. Most people see no reason to tell a doctor why they are leaving. That said, it is disconcerting that I am so often unaware of a patient’s dissatisfaction. I overestimate patients' satisfaction because the feedback I get comes overwhelmingly from patients who have chosen to remain in my practice. The power differential between doctors and patients also keeps (most) patients from telling me what they really think. We physicians are also well compensated, which can make us feel valued — sometimes beyond our true worth.

A personal experience reminded me how difficult it can be for a doctor to get feedback. I recently saw a physician for a minor surgical procedure. The doctor was exceptional, both as a person and as a surgeon. Our preoperative visits were excellent, my surgical experience was perfect, and my convalescence was uneventful. While the doctor examined me during my postoperative visit, he noticed trauma at the surgical site. “You must have injured this,” was how he reported the finding to me. As a writer, I guess I should have respected his use of the active rather than the passive voice. However, “there seems to have been some trauma here” would have sat better with me. I am sure this difference meant nothing to him, and he could not have known how I — an obsessively adherent perfectionist — was offended. I certainly said nothing.

Of all my panels, the ghost-patient panel is the one that causes me the most wonder. Its very nature leads to startling recollections of patients who have vanished. I analyze past visits and relationships when I realize a patient is on the list. I am sure I could learn a lot about both what I do well and why I fail if I better understood the whos and whys of the ghost-patient panel.

Share

A previous version of this essay was published here.

This account probably has access to embargoed journal articles and, with this advantage, is consistently the first to post, gaining a sizable following. Sadly, it offers no insight or critical appraisal.

To many of us, the account is a metaphor for social media. Many want to be heard. Few have something to say. When it comes to the week’s medical papers, it’s easy to roll your eyes, but when it comes to a life— it becomes jarring.

Recently, a professional colleague and fellow physician died. It was untimely (late 40s), and the cause was cancer.

As an oncologist, I am constantly reminded of the unpredictability and fleetingness of life, and I pray it gives me perspective that many challenges I face daily are petty and minor and absurd. I pray it reminds me why not to go to the next conference or weekend ‘retreat.’

On social media, the RIP posts began.

even though I have only met her briefly at random meetings…

Although I did not have the privilege of knowing <the doctor> personally…

was a dedicated member of our <professional> committee…

I had the opportunity to present at an <professional> event where she was one of the judges…

Physicians posting RIP and tagging the photo of the deceased with the names of their living colleagues— perhaps in an effort to draw attention to the post?

I literally just had a whole conversation about her amazing shoes…

The posts reminded me of a eulogy I attended.

The speakers took turns talking about our colleague’s hirsch-index, and R01 grants. He had given 1000 presentations and published 300 papers. I felt uneasy. Later a student of mine gave me the words to understand what I felt. A eulogy should be about eulogy values, not resume values. How a person made those closest to them feel— not the accomplishments a person had to someone at a distance.

When David Foster Wallace gave the 2005 Kenyon College commencement speech, he used the venue to explore themes that eluded him in his novels. Using the story of the fish swimming in the ocean—being asked ‘how’s the water,’ and asking ‘what the hell is water’—Wallace notes we are oblivious to what is right in front of our eyes, even though that is what matters most.

The irony is not lost on me that a thoughtful speech has itself been turned into a cliche in the the form of a tiny book (with incorrect spacing) by the publishing house.

The obvious realities are the hardest to talk about.

An oncologist sees an incredible— staggering— amount of death. And you move from being a third-year medical student crying alone to being an attending, briskly moving the next patient and the next, without the time to truly process what you witnessed, and yet, perhaps later, at home, thinking about your patient and her kids.

Untimely death wounds us in unique ways. A few years ago, a colleague asked me to treat an 85-year-old patient with severe Alzheimer’s— unable to recognize himself or others— for non-Hodgkin’s lymphoma. The nodes appeared just weeks after he completed antibiotics for pneumonia. She thought the nodes were post-infectious, but biopsy revealed cancer.

“What is he allowed to die of?” I asked my colleague.

She was startled.

“Because he is not allowed to die of pneumonia, and you think he shouldn’t die of NHL… what is he allowed to die of?” I asked.

What is a 47-year-old allowed to die of? Someone otherwise sound of mind and body— someone with kids still in school. Certainly not cancer. It strikes me as cruel and grotesque.

And yet, in the 15 years I have been in oncology, I have seen, over and over, oncologists ignore the water.

How are we really going to improve if we give standing ovations for drugs that cure 0 people, add 6 months to life, and cost 477,000 per year?

How are we going to improve if we are incapable of learning that retrospective observational target trials give false conclusions?

How are we going to improve if we test our new drugs against dilapidated, antiquated standards of care?

How are we going to improve if we don’t run trials with the best post-protocol care?

What the hell is water?

Even while facing her own diagnosis, she established the <Award>, raising >$1.35M….

But as little consolation, we as a community clearly heeded the urgent call and made sure she could feel loved, heard, and leave a lasting legacy behind.

If a tree falls in the forest, is milled into paper, printed on a poster, carried in an overhead bin to a future cancer conference, and no one looks at the poster, did the tree make a sound?

Untimely death is a reminder that so much of our professional lives is misspent time. A weekend with your kids becomes a weekend in a hotel room.

Wallace’s graduation asks us to reconsider our default setting as humans, as “the absolute center of the universe, the realest, most vivid and important person in existence.”

And yet, online RIP can’t seem to avoid that fate.

We remember our colleague who, like us, expanded access to clinical trials.

But don’t think too much about what it means to leave young children behind when they need guidance most to discover who they are meant to be.

The former senator Ben Sasse is dying of pancreatic cancer. He is 54 years old, and will leave behind 3 kids, and a legacy of being a decent and honorable person in politics at a time when that was as common as water on Mars. When he dies, oncologists will remind us that daraxonrasib may have bought him a few months, perhaps the better part of a year. It did so with tremendous rash, a face that wept with blood, and was unable to hold the cancer at bay for as long as we need it to.

What the hell is water?

I am not sure that the gap is closing, but I know that they can be harder to tell apart.

The Charlatan

Healthcare charlatans are polished characters who, despite having no real expertise, pitch themselves as experts. Their ultimate goal is not to help people but to increase their own wealth and fame. The internet has been a boon to the charlatan. Instead of driving covered wagons from town to town, they can reach thousands (millions?) of marks with well-placed Instagram posts, tweets, or TikToks.

The charlatan’s skill is that we don’t know whether they believe in their products. While I assume the traditional snake oil salesman didn’t use his products, some of today’s might be using the peptides, hormones, and full-body scans he (or now she) promotes.

The Expert

Unlike the charlatan, the expert traditionally stalked the halls of the academy, published in journals, and “presented papers” at meetings. He (she’s were rare) worked only for people’s health; there was little else. There wasn’t much money associated with success, and — unless you were Lister, Fleming or Salk — no chance of fame. Of course, before the era of evidence-based medicine, some of the theories and treatments promoted by experts (lobotomy, pyrotherapy, diethylstilbestrol) might have been as worthless (or worse) as those promoted by charlatans.

Why it is now harder to differentiate the expert and the charlatan

Experts can still be wrong, and the modern world can make experts look like charlatans. Evidence-based medicine (EBM; the expert’s tool) integrates clinical experience and expertise with the best available evidence from systematic research. Room for error remains in EBM: clinical experience can mislead; experts can be seduced by bioplausibility; studies can be flawed or wrong. Drug and device makers spend freely to influence experts and encourage experts to influence. The internet offers fame the tweed-vested academic of yesteryear could never have dreamed of. Look at people like Emanuel, Topol, and Hotez, with their books and online/media presence.

Just because the charlatan and the expert look more alike these days does not mean they are. We hope today’s experts still work single-mindedly in the interests of people’s health. Yet with the inherent (and necessary) flaws of evidence-based medicine, and with wealth and fame tempting experts, I worry they are more likely to be wrong. I believe the expert and the charlatan differ in intent. The expert might be wrong, but s/he is certain s/he is doing the right thing. If this is not the case, the expert becomes a charlatan.

Vinay Prasad

Read more

When the tricuspid valve leaks, so called tricuspid regurgitation or TR, the patient feels full in the abdomen and has severe peripheral edema. It’s a difficult problem to treat, mainly because the main cause of the leakage is pressure overload from the left side of the heart. Though sometimes pacemaker leads can tether a leaflet and create TR. One huge problem with TR is that when a patient comes in with severe TR, they are given diuretic therapy, and days later, when the pressure is reduced, the TR has decreased a lot. TR we say is highly dynamic.

The prevalence of TR increases rapidly with age. Millions of people in the US are affected. In days of old, there were only two options: careful titration of diuretic therapy and/or surgery. The problem with surgery when done only on the tricuspid valve is that complications were super high.

Enter percutaneous solutions. The idea here is that if you could snake a tool up to the valve, which is a straight shot from the femoral venous access site, you might be able to clip the valve leaflets to stop the leakage or even replace the valve, much like what is done with transcatheter aortic valve implantation (TAVI).

Two devices for tricuspid valve have passed FDA clearance. One is called a transcatheter-edge-to edge repair (TEER), which is, basically, a clip device. The other is a complete valve called the EVOQUE valve.

I have previously discussed the TRILUMINATE trial, which was one of the most biased trials in all of cardiology, because one group got the procedure, the other group got tablets, and the primary endpoint was driven only by quality of life changes. In TRILUMINATE, there was no reduction in death or hospitalizations for heart failure. The bias was obviously the lack of placebo control; you can’t compare quality of life when one group gets a procedure, and the other gets no procedure.

The TRIC-I-HF trial

TRIC-I-HF was a German investigator-initiated trial at 29 sites. It was also open-label with no placebo procedure. There was a 2:1 randomization of 360 patients. The authors used two primary endpoints: one was a hierarchical composite of death from any cause, hospitalization for heart failure, and quality-of-life improvement at 1 year, assessed by the win ratio. If the between-group difference was significant, a second primary end point would be tested: a composite of death from any cause or hospitalization for heart failure through 3 years.

The patients were quite old (80) and sick: 55% had a HF hospitalization in the prior year, 82% had cardiorenal syndrome, and 92–98% had AF. While multiple types of intervention were allowed, 98% of patients had TEER.

Notably, it was sicker group than TRILUMINATE, with roughly 17.5% 1-year mortality in the control arm versus 5–8% in prior trials.

Results

The trial posted quite positive results. The win ratio for the first primary endpoint was 2.42 (95% CI, 1.76 to 3.33; P<0.001), favoring tricuspid-valve repair.

For the second primary endpoint, freedom from death from any cause or hospitalization for heart failure through 3 years was 52.4% in the tricuspid-repair group and 21.0% in the medical-therapy group (hazard ratio 0.40; 95% CI, 0.29 to 0.55; P<0.001).

Unlike TRILUMINATE there was a marked reduction in hospitalizations for heart failure 19.4% vs 43.9% HR 0.35 (95% CI 0.25-0.50).

Much was made in the discussion about death rates. In the primary analysis at one year, the death rate was lower in the repair arm: 13.1% vs 16.3% (HR 0.74, 95% CI 0.42-1.29). At three years, the difference was 19.0% vs 27.6% with a HR 0.62 and 95%n CI that now stayed below 1.0 at 0.40-0.96. (More on this in the discussion).

Major adverse events within 30 days occurred in 14 patients (6%) in the tricuspid valve intervention arm.

Comments

The authors declared victory. The accompanying editorialist was highly positive. My friends in structural cardiology were clearly happy, as millions of people have TR.

The thing is that there are indeed people with really bad tricuspid valve disease who I have referred for something interventional. Most often, my colleagues say there is not much to do because of anatomic hurdles. But if this procedure worked as reported in the trial, it would be a real win for lots of people.

But I worry about the trial results. While they look positive, there are at least six nuances that I see. Maybe you see more.

• The endpoint that drove the results were open to bias in this open label trial. In the win ratio, the HF hospitalization endpoint did the work: 9,280 vs 2,871 wins. Death rates were close (4,366 vs 3,166), and 74% of pairs were tied on death. Hospitalization, of course, is a clinician decision. Since everyone knows the treatment assignment, it’s easy to give an increase in oral diuretics to a patient with a device while admitting a patient without one.

• The second issue also relates to HF hospitalization. That is the crossover rule allowed a patient in the medical arm to get crossed over only after a hospitalization for heart failure. Tricuspid repair is reimbursed and available in Germany. For a control patient (or their doctor) who wants the valve fixed, a qualifying hospitalization is required. The data are consistent with this concern: 48 of 123 control patients (39%) crossed over. In 28 of those 48 cases (58%), the procedure was done during the index hospitalization itself.

• The third issue is that congestion markers don’t support a 65% reduction in hospitalizations. For instance, if the repair was so great at reducing congestion, you’d expect more improvements in body weight, edema and 6-minute walk distance. None of these measures were that impressive. In fact, the improvement in KCCQ, a questionnaire regarding quality of life was only 6.9 points better in the device arm, which was less than in TRILUMINATE. In the past trials, there was a big quality of life benefit and no effect on hospitalization, but in this trial there was little effect on quality of life and a big effect on hospitalization. This makes you wonder about a bias-driven effect as both should track similarly.

• The fourth issue was that the amount of TR reduction was modest. Only 70% of repair patients had moderate or less TR at 30 days, versus 30% of controls. That is less complete than reported in TRILUMINATE. How do you explain the outsized effect on hospitalization with such a modest reduction in TR? Maybe there is a threshold effect? Maybe not.

• The fifth issue comes in the death numbers at three years. Multiple proponents have emailed me that I should not ignore the lower death rate at 3 years. Death, they rightly argue, is an unbiased endpoint. The problem is that only 20% of patients had completed the three-year follow-up. At one year, when follow-up was complete, there was no statistical difference in death rates.

• The sixth issue was baseline imbalances which favored the device arm. The control arm had slightly more patients with a higher TRI score, AF, CAD, previous MI and stroke.

  Subscribe now

  ---

The bottom-line is that we need trials to guide our advice to patients.

But trials have to minimize bias. Device trials are always going to be harder than drug trials where a placebo does the work of masking.

I don’t suspect any nefarious intent in this trial, but the biases noted above preclude making strong conclusions about the value of this device.

A proper test of tricuspid valve intervention would at minimum include a sham-control arm, where patients would be blinded to the treatment arm. If we can blind coronary stenting, surely we could blind tricuspid valve interventions. Blinding would not fix everything in this case because the clip is obvious on an x-ray, but at least it would blind patients.

We would also want to know about the difference between cardiovascular death and all-cause death, because when you recruit 80-year-olds, there are many non-valvular causes of death.

Before embracing the tricuspid valve repair approach, I feel we need more compelling data.

Let me know what you think.

This essay is on the long side for us, but I think it warrants space, and not just because I am part of its subject. It tackles the question, “What is medicine for?” by considering what conventional and functional/integrative medicine offer. Like the best essays, I left if more confused than I was when I started it.

Adam Cifu

When I was about eight, I saw a magic trick at a birthday party. The magician raised a pitcher in one hand and a bowl in the other, then poured a stream of sparkling blue water from the pitcher into the bowl. But when he showed us the inside of the bowl, it was empty.

I’ve cycled through various theories for how this worked, but I’ll likely never know for sure.

It’s a little like the body: Something happens. We can see that it happened. But often, why and how remain a mystery.

I’ve spent much of my career probing such mysteries as a freelance magazine writer, covering health topics largely through a functional/integrative-medicine lens. I’ve written about everything from hormones to autoimmunity to complex digestive disorders. As I have zero medical training, these articles rely heavily on expert input, largely from functional, integrative, and sometimes naturopathic doctors who I find insightful and trustworthy.

I’m proud of the work I’ve done. But new doubts are starting to haunt me. I’ve been an enthusiastic reader of Sensible Medicine since it launched in 2022, and its rigorous examinations of medical evidence have led me to cast a newly wary eye on the functional/integrative world’s confident mechanistic stories, often built on shaky evidentiary ground.

Much of what the functional world asserts — that prevention can avert the need for later treatment; that diet and lifestyle are potent tools for health and healing — seems intuitively true. But if there’s one thing I’ve learned from reading Sensible Medicine, it’s that medical claims that feel intuitive often prove to be false. Is it true that cutting out gluten can preserve or restore thyroid function in people with Hashimoto’s? Is there real evidence that healing a leaky gut with an elimination diet and gut-repair supplements resolves inflammatory symptoms?

Doctors I’ve interviewed insist they’ve seen exactly these outcomes in their practices. But every kid at that birthday party would’ve sworn that they’d watched a pitcher’s worth of water vanish into thin air. Is it possible these doctors are equally certain of what they’ve seen, and just as wrong about why?

Then again, if their patients really did achieve better function, less pain, or fewer symptoms, does it really matter if the explanatory model isn’t exactly true? Isn’t that largely the story of health and healing throughout human history, even in the supposedly evidence-based medical world? (How many people who recovered from the flu after a dose of Tamiflu believed the pill was pivotal? Ditto Paxlovid for Covid, or surgery for meniscal tears.)

Dr. Wayne Jonas writes in his terrific book How Healing Works,

Researchers at the Society for Interdisciplinary Placebo Studies have shown that three [things] — belief and expectation, meaningful social learning, and reinforcement or conditioning — are the underlying mechanisms of the placebo response, and the likely explanation for the majority of human healing in any system or from any treatment.

Modern medicine has an extraordinary capacity to treat acute illness and save lives. But much of health and healing unfolds in a murkier realm, where biology becomes difficult to disentangle from the psyche, spirit, and environment. In this realm, Jonas asserts that good healthcare is mostly about learning how to ethically harness the mechanisms of the placebo response and heal through meaning. I think this is right.

But it does present a tension between epistemic purity and practical results. Prizing either one to the exclusion of the other seems unwise, for both patients and doctors. In the hope of finding a wise middle ground, I sought out a strong advocate for each side.

I spoke with Dr. Adam Cifu, familiar to Sensible Medicine readers as a general internist at the University of Chicago and outspoken champion of evidence-based medicine. I also spoke with Dr. Rahul Iyengar, a family physician and founder of a concierge primary care clinic in Nashville focused on prevention and health optimization.

Despite their philosophical differences, Cifu and Iyengar agree on quite a lot. Both think time, attention, and careful listening are central to good healthcare — and that the current system makes those things difficult.

“Conventional medicine has sort of invited the rise of alternative medicine because we do such a crappy job at [using] the placebo effect,” Cifu told me. “We’ve outsourced it, saying, ‘I don’t have a pill for this, so go find somebody who will convince you they’re doing something.’” He’d like to see primary care generalists have more time to provide caring attention, which he sees as the main thing that helps when there’s no obvious medical fix. For his part, Iyengar founded his clinic specifically out of a desire to have the time to provide “relationship-based care.”

They also agree that diet, exercise, and basic lifestyle factors matter more than elaborate optimization regimens. And neither has much patience for the wilder reaches of the wellness marketplace. “Big Supplement is arguably much scarier than Big Pharma because there’s zero regulation,” Iyengar says. “It’s all just pretty marketing and packaging.”

The differences start to creep in around how much medicine should do in pursuit of health. Specifically, here are four areas where the philosophies meaningfully diverge:

1. When should medicine intervene?

The first disagreement is on the value of prevention and optimization.

For Iyengar and many others in the functional/integrative space, waiting for disease to fully manifest before intervening is akin to skipping routine auto maintenance that might prevent a major mechanical problem down the road. “Come see your doctor when you feel healthy,” he says, “and it’s really easy and affordable to keep you healthy as opposed to waiting for something to break.”

He advocates for preventive blood work that goes beyond the standard annual panel, including more extensive cardiac, metabolic, thyroid, and nutritional markers. “Let’s find out all these little things and fine-tune them before you feel bad,” he says. He sees particular value in catching a trajectory early enough that diet or other lifestyle changes may alter it before medication becomes necessary.

Cifu says we need more evidence to know if this juice is worth the squeeze. With LDL cholesterol, for example, we know that lowering very high levels improves outcomes. But once someone is already in the normal range, he says, the absolute risk reduction from pushing it even lower is much smaller. “We don’t know what the benefit of being in that ‘optimal’ realm is. And therefore, people aren’t able to make intelligent decisions about what they’re willing to do to get there,” he says.

This points to a fundamental divide on how readily medicine should expand its attention from disease to health itself. “Doctors do a whole lot more good treating sick people than treating healthy people,” argues Cifu. When medicine makes the jump from treating illness to intervening in healthy people, “that’s where the level of evidence has to be higher, because when you start mucking around with healthy people, you’re just so much more likely to do harm.” That potential harm includes not only side effects and expense, but what he calls the “cognitive risk of ‘I’m unwell. I could be doing things better.’”

Iyengar counters that evidence of prevention’s benefits is hard to produce when trials overwhelmingly study people who are already sick and commercial funding favors interventions with a product at the end. “Who is it going to benefit if the [intervention] is just lifestyle factors or changes? That’s not a product,” Iyengar says. “Who is going to fund the study?”

For now, Cifu doesn’t begrudge anyone their non-evidence-based health pursuits, as long as they’re low-risk and out-of-pocket. “There are people for whom this is a hobby, and it makes them happy to do it,” he says. “I pay money to have access to a ceramics studio. If someone wants to pay money to have access to a functional medicine practitioner doing this stuff, that’s fine. Of course, the benefit of my hobby is that my friends get to receive my pottery.”

2. Is medicine for dysfunction or just disease?

Things get thornier when we move from people who feel healthy to those who don’t. What should medicine offer someone suffering from fatigue, headaches, gut problems, or brain fog when the standard workup comes up empty?

For Cifu, the first obligation is a thorough but bounded diagnostic search – ruling out the common things followed by the less-common possibilities. But at some point, the search has to end. “We can’t clear the differential diagnosis for every problem of every person because it’s impossible.” From there, the doctor’s job is to manage symptoms where possible and stay alert to changes. “The care is attention,” he says. “It’s, ‘I’m still here caring for you.’”

For a patient with vague symptoms and normal bloodwork, Iyengar said he might pursue advanced gut or stool testing, but also a deeper investigation of family and psychosocial history, environmental exposures, occupation, daily habits, even something as mundane as frequent contact with cleaning chemicals.

When I asked Cifu whether conventional medicine sometimes stops looking for the causes of symptoms too soon, he was adamant: “I actually think our problem is the opposite. We so frequently go absolutely crazy in our evaluations. Everything from how we’re trained to where our money comes from incentivizes us to do too much.”

This points to a deeper disagreement about medicine’s purpose: Is it just for diagnosing disease? Or does it extend to identifying modifiable contributors to poor function – to helping people feel better?

For conventional medicine, once a reasonable differential has been exhausted, the responsible answer is often simply, “We don’t know.” Which may be epistemically honorable but doesn’t feel particularly helpful. The integrative model casts a wider net and uses a broader toolkit to pursue better function. This approach has obvious appeal if you’re struggling with mysterious symptoms.

Cifu’s objection is that the more determined you are to find an explanation, the greater the risk that eventually you’ll land on one, one that likely has nothing to do with your symptoms.

3. Where does medicine end?

If some patients may benefit from interventions that fall outside the bounds of evidence-based medicine, should those interventions be brought under the physician’s umbrella, or kept clearly separate?

Cifu is open to patients trying things like acupuncture or naturopathic care when conventional medicine has run out of answers, provided it’s unlikely to cause physical harm, financial harm, or delay necessary care. But he wants a bright line between the two domains.

“It’s a matter of figuring out how to [suggest that as an option] without endorsing it,” he says. “If my office were in a strip mall, I’d want you to have to go out to the parking lot and enter a separate door to see them rather than access both of us from the same waiting room. That would actually be kind of important to me.”

Iyengar’s approach is the opposite. If his patients are going to seek out wellness services anyway, he’d rather bring some of them inside his practice, where they can be offered more safely. His clinic offers massage, IVs, and Botox alongside primary and urgent care; he describes his goal as an all-in-one health ecosystem that blends “prevention, wellness, and health in one spot.” (He draws the line at peptides, which he considers too poorly studied and risky to offer responsibly.)

In my searches for functional-medicine sources, I’ve found an increasing number of them offering “longevity” and aesthetic services. Even functional clinics that don’t offer Botox still often promote services like ozone therapy and nutritional IVs as legitimate health interventions, despite thin evidence.

This is a stubborn pickle. An unbending commitment to the evidence-based toolkit can leave suffering patients at sea with nothing left to try. But if the line between declining something (peptides), offering it (Botox), and promoting it as treatment (nutritional IVs) is up to each individual doctor or clinic’s discretion, this puts a lot of onus on the lay person to select someone whose instincts she can trust, a task most of us don’t have the training or skills to do.

4. Which incentives distort care the most?

Both models contain perverse financial incentives, both obvious and covert. Cash-pay functional medicine can reward practitioners for finding ever more things to test, treat, supplement, and optimize. At the industry’s upper reaches, some of its most prominent figures profit directly from products and services they recommend to their enormous audiences — hardly the personalized, precision medicine they claim to champion.

On the flip side, Iyengar says, “Sick care is truly the biggest industry in America. It incentivizes ongoing treatment, not cures or prevention. And currently, no one’s incentivized to change it.” It’s not that individual doctors want their patients to stay sick, but that the system pays much more readily for treating illness than for the time-intensive work he considers necessary to prevent it (as well as research supporting that work).

All this adds up to a dizzying landscape for the individual patient to navigate, one pulsating with questions of epistemology, money, incentives, blind spots, discernment, and the subtle dimensions of healing that don’t lend themselves to easy quantification.

What’s a health writer to do? What can she confidently assert and advise?

I don’t want to believe or spread a story that’s not true, but I also don’t want to make it my mission to aggressively debunk every act of magic that’s led to real relief from suffering. Like the bowl of disappearing water, healing can be real even when we don’t understand the precise mechanism at play. Maybe the answer begins with being less certain. We can take the healing seriously, even as we hold the explanation for it lightly.

My choice is this: In my life and my writing, I’ll seek out doctors who use evidence as a guardrail while acknowledging its limits; who are open and attuned to the powers of expectation, relationship, and meaning, without confusing them for proof; and who keep a clear line between beauty treatments and medicine. What turns me off, in any direction, is unearned certainty.

Others will make different choices, based on their own calculus. Maybe the wisest thing medicine can do is help us make those choices with our eyes open.

Share

Mo Perry is a national award-winning freelance health journalist and professional stage actor based in the Twin Cities, Minnesota.

Photo by Little Annabell on Unsplash

My career has been marked by medications that have altered the course of medicine. The GLP-1s have not been as revolutionary as statins, protease inhibitors, CFTR correctors and potentiators, checkpoint inhibitors, or the like. But, as obesity is a few thousand times more common than cystic fibrosis, the impact is, at least, comparable.

The history of the GLP-1s is interesting. The drugs are somewhat famously inspired by a peptide found in the saliva and venom of the Gila monster. For a lizard that rarely eats, a peptide that inhibits GI tract activity (decreases gastric emptying) and suppresses appetite is pretty useful. The first GLP-1, exenatide (Byetta), was approved in 2005. It was a crappy drug that I almost never used. Exenatide needs to be injected twice a day and causes so much nausea that the associated weight loss seems beside the point.

Semaglutide (a long-acting GLP-1 receptor agonist) was approved at the end of 2017. Tirzepatide (a combination GLP-1 receptor agonist and GIP – a glucose-dependent insulinotropic polypeptide) came out in 2022. I’m pretty sure that the people who appreciate the effectiveness of these drugs most are people who have struggled with obesity for decades – either personally or as a primary care doctor.

Up until 2017, people with decades of obesity and their doctors had little hope. Sure, there were drugs, but these led to only modest weight loss and failed to improve the outcomes that weight loss should be associated with. Bariatric surgery was (is) effective, but it is, well, surgery. And there is diet and exercise.

Don’t even start with, “Well, you and your patients could have focused on diet and exercise.” Behavioral change is hard, and weight loss in your 40s and beyond, when you have been overweight since your teens, seems next to impossible. I can counsel diet and exercise with the best of them, but I have always estimated that the number to counsel to achieve weight loss is around 2000. I have, only half-jokingly, suggested that given these numbers, it is not something we should even do. I wouldn’t prescribe a drug that is this ineffective. This snarky comment of mine is always met by stories of the one person who comes back 6 months later, 25 pounds lighter, with an A1C 1 point lower. Not that there is any proof of causation in that story – and you never hear the follow-up when that patient presents one month later with painless jaundice. (Sorry, that was kind of dark).

A few clinical observations.

Maintenance of GLP-1 weight loss

The GLP-1 skeptics point out that these drugs are a treatment for obesity, not a cure. This is absolutely true. (Though, of course, the same can be said for diet and exercise.) Data suggests that most people who come off their GLP-1 regain the weight lost.

My experience is that the two types of people maintain their weight loss.

First are the record-keeping obsessives. These are the people who can track everything – food diaries, calorie counts, calories burned in exercise, fiber intake… – and recognize that they can’t let any of these things change after drug discontinuation. Most of these people tell me some version of, “I know why I have lost the weight, I know it will be harder to keep doing it off the drug, but I am committed to doing that.” It takes quite a lot to succeed, but some people, in the right situation, can manage.

Second are those whose use of the drugs provides a total reset. A case study may be best to explain this.

R and S are a married couple in their 60s. When their kids were born, they were in their early 30s, and they both had a BMI in the high 20s. Both slowly gained weight in their 30s, with the pace picking up in their late 40s and 50s. The explanation was obvious: falling basal metabolic rate, less exercise – mostly related to work and childcare responsibilities -- and more regular alcohol use. They describe adopting a COVID cocktail that they did not de-adopt post-COVID.

They both began tirzepatide with BMIs of 37-40. 18 months later, they discontinued the medication with BMIs of ~ 26. By the time they came off the drug, they had completely changed their lifestyle: different weekly meal plans, daily exercise together, cocktails limited to a Sunday evening Negroni. It’s been 18 months of weight stability.

Openness to discussing obesity in the office

Prior to the GLP-1s, I seldom discussed obesity in the healthy overweight – those people with BMIs between 30 and 40 who were biochemically normal – normal lipids, A1C, LFTs, blood pressure. I absolutely think these weights are associated with worse outcomes, but with little to offer, and knowing that most people already beat themselves up about their weight, I was happy to leave it unacknowledged.

Now with therapy available, I have made it my practice to bring treatment up with anybody with a BMI > 35. (Again, I am talking about people without other health issues that would warrant more aggressive weight management). I find the diversity of reactions interesting.

I’m glad you brought it up; I’ve been thinking about it.

No, I have no interest in looking like some gaunt influencer.

Huh. I haven’t considered it. Would it make sense for me?

Fewer of these people end up taking the medications than I would think.

GLP-1-related health concerns

I am not talking about medullary thyroid cancer, hair loss, “Ozempic face,” or muscle wasting. The effect on cancer seems to be a wash, and the other side effects seem to me mostly the expected effect of living in a calorie deficit for a time. (I am no Pollyanna. I am not sure we won’t eventually find some important side effect, though I expect the benefits of these drugs, broadly defined, will outweigh the risk.)

What I am talking about are the findings that weight loss reveals that send people rushing in for appointments. Visible xiphoid processes, palpable umbilical hernias, and partner-detected lipomas have reached epidemic proportions.

A new appreciation for the first law of thermodynamics

The first law of thermodynamics states that energy cannot be created or destroyed; it can only be transformed or transferred from one form to another. For years, my counseling around diet and exercise would be dismissed with, “I don’t overeat.” When I pointed out that losing weight is all about calorie balance, people pushed back.

I get it; we do all have different basal metabolic rates. One can eat a small amount and still gain weight. We cannot all be Stanley Tucci, making a living eating in Italy while remaining trim at 65.

Not surprisingly, I’ve had dozens of patients express doubt that GLP-1s would even work because, “I don’t overeat.”

I never hear this after the person has gone on the drug and lost weight.

Share

Do you think a healthy pregnant woman in 2026 should get a COVID-19 shot?

Because our goal here is to “showcase a range of ideas and opinions about all things bio-medicine”, we thought it would be informative to post Dr. Rind’s answer and give Dr. Prasad the chance to expand his answer. This piece ended up being even more of a back-and-forth than we expected.

If you finish this and think, “I don’t get to see this sort of agreeable disagreement anywhere these days,” subscribe!

Adam Cifu

Subscribe now

David Rind

I agree with many of Dr. Prasad’s answers in his article, but there was one that I disagreed enough to respond to: his conclusion that a healthy pregnant woman in 2026 should “absolutely not” get a Covid-19 vaccination. This is a fraught topic, and it is one my organization, The Institute for Clinical and Economic Review (ICER), tackled in its July 2026 report on Covid vaccination. The evidentiary support for any unreferenced sentences in this piece can be found in that Report.

I want to start by asserting two truths. First, the mRNA vaccines rolled out in the US as a result of Operation Warp Speed were lifesaving in 2021. If you do not agree with this, it is pointless to continue reading what I have to say. The second is that, in 2026, any benefits of COVID vaccination are much, much smaller. This is because the population is no longer immunologically naïve. Even if you believe universal vaccination should be required, you should agree that the absolute benefits today are much smaller than they were five years ago.

For most people, getting a COVID vaccine reduces an already very small risk of serious disease while carrying a relatively high risk of side effects severe enough to interfere with work or activities for a couple of days. This is a tradeoff, and reasonable people can choose to get vaccinated or not. The evidence that vaccination reduces transmission at the population level is relatively unconvincing. Vaccinated individuals can transmit COVID even if they are asymptomatic, and being vaccinated may increase the likelihood of COVID infection being asymptomatic. Better evidence is needed before vaccination can be claimed to protect others in society against COVID in 2026.

In the 2024-2025 season, younger adults in the US were hospitalized with Covid-19 at a rate of about 20-40 hospitalizations per 100,000 people. ICER estimated that vaccination might reduce this by about 6 hospitalizations per 100,000.

Some groups, though, are at much higher risk. The rate of hospitalization for those 85 and older is 1200 per 100,000 people. Vaccination in those over age 65 is likely less effective than in younger people on a relative basis, but the best recent evidence that ICER found strongly suggests that vaccination would reduce hospitalizations by 75 per 100,000 and deaths by 22 per 100,000 older adults. Older adults often have fewer side effects with COVID vaccination and might well choose to be vaccinated for these risk reductions.

Pregnant women have consistently been found to have higher risks of severe Covid-19 than non-pregnant women. However, the risk of severe Covid-19 has been decreasing in all groups year over year, so the absolute benefit of vaccination is uncertain and quite possibly small. The American College of Obstetricians and Gynecologists (ACOG) continues to strongly recommend COVID vaccination for all pregnant individuals during any trimester at the earliest opportunity.

ICER’s conclusions about the evidence lead me to question both the ACOG recommendation and Dr. Prasad. Infants less than 6 months of age are at higher risk of hospitalization for COVID than any other age group until people reach the age of 75. Do we know if maternal vaccination is beneficial?

On this question, ICER concluded we needed to look to biologic plausibility – typically a form of evidence low on the evidentiary hierarchy. Early third-trimester vaccination for COVID has been demonstrated to optimize cord blood antibodies. Higher levels of antibodies are associated with greater protection in infants; this is a type of “dose response”, which increases the likelihood that this association is causal. Additionally, observational data show that influenza and pertussis vaccines in pregnant women can protect young infants against those diseases.

What about harms? There have been many studies of COVID vaccination in pregnant women that have found it to be safe. However, fever during embryogenesis has been shown to be teratogenic in animals; the evidence in humans has been somewhat inconsistent. Given this concern, the frequency of fevers after COVID vaccination, and the desire to optimize placental antibody transfer, vaccination early in the first trimester seems riskier than other options and may offer fewer benefits. In contrast, early third-trimester vaccination is unlikely to be harmful even if it results in fever, and it optimizes cord blood antibody levels.

Since COVID itself results in fevers, women contemplating pregnancy may want to try to have had the most up-to-date vaccine prior to pregnancy and then get revaccinated early in the third trimester. We currently have limited evidence and biologic plausibility to support this approach, but it is what I tell my patients, my friends, and my family members. ICER’s report explicitly called out the need for randomized trials of third-trimester vaccination for COVID to definitively answer questions about safety and efficacy.

I think ACOG should have made more targeted recommendations, but while we wait for better evidence, I think Dr. Prasad is wrong to say that in 2026 a healthy pregnant woman should “absolutely not” get a COVID vaccine.

Vinay Prasad

I appreciate Dr. Rind, and ICER in general, but on this issue, I maintain that a pregnant woman in 2026 should “absolutely not” get a COVID-19 shot.

What is our core disagreement? People who choose to get COVID shots differ from those who don’t in ways that go beyond the vaccine. If you believe this – which I do – you cannot use observational data to make your argument. Not for efficacy, nor safety. Proof they differ comes from studies that compare the two and find ludicrous differences: different rates of car accidents or all-cause mortality. If healthier, richer, better educated people get (more) shots, even safety risks can be hidden.

Dr. Rind cites a rate of 20/100k hospitalizations among young people in 2024-25. The CDC tracker he links to tracks “COVID-associated hospitalization.” This is technically not the same as tracking kids who are hospitalized because they can’t breathe because of COVID in the lungs. It includes kids who are hospitalized for all sorts of reasons (broken bone) who incidentally have COVID (when swabbed on intake). Second, his numbers are out of date. 2025-26 is down to 8 per 100,000 among kids 0-4 (presumably the target of passive maternal antibodies) at its peak.

These data don’t separate healthy kids from kids with comorbidities. A child with in utero birth defects has a much higher chance of hospitalization (with or from) COVID than a perfectly healthy, morphologically normal baby on a 20-week ultrasound.

Dr. Rind says that ICER estimates COVID boosters reduce the rate of hospitalization in young adults from 20 to 14 per 100,000 (6 per 100,000). The 6 per 100k, or 30% RRR, is an imaginary number. Dr. Rind has no randomized evidence to support this number. He is citing a target trial from the VA. Target trials are unreliable; they showed Paxlovid saved lives, glp1s reduced Alzheimer’s, and other claims that were debunked with randomized studies. In this case, the fundamental challenge is that people who receive both vaccines are still different than those who get flu only. The authors are familiar characters to those of us in COVID debates, and to be blunt, I don’t trust them. Their falsification tests vary from publication to publication and often exclude the greatest falsification test. Ask them to publish a Kaplan-Meier curve of all-cause mortality for the matched cohorts, which is the ultimate falsification test to exclude healthy user bias. Extrapolating this relative risk reduction to maternal antibody transfer to babies is dubious. Finally, even assuming his RRR, Maternal antibodies last at most 3-9 months, conferring even smaller benefits.

Dr. Rind says many studies show vaccinating pregnant women is safe – again citing observational data. In my opinion, we have no idea what the true safety profile is in pregnant women. There was a randomized controlled trial that was going to answer this question, but it was aborted (conveniently for the manufacturer). Those truncated RCT data are on the package leaflet, and they allow no strong​ conclusion to be drawn. We don’t know if the product increases birth defects; we don’t know if the product increases maternal hypertension. Pregnant women deserve better than this limited table.

Every year the covid-19 vaccine is different. It is possible that a future antigen will stimulate the immune system in a more problematic way. With pregnancy, the priority of safety and no one knows the safety of the next vaccine.

Dr. Rind admits that his level of evidence is biologic plausibility. If you can take healthy people and administer medical interventions based on biologic plausibility, why should that not extend to other situations in medicine? Should I be able to give a fat 8-year-old a glp1 agonist because of biological plausibility? How about a PCSK9? You can construct similar chains of logic here.

I believe in a simple rule. Do not inject pregnant women with anything unless you have robust randomized evidence that there is a net clinical benefit to mother and child. That standard is not met here. Not even close. In the meantime, not only do I think pregnant women absolutely should not get a COVID shot – I think it is negligent to give it.

Final point: if you powered an RCT of pregnant women getting a COVID booster starting in 2027 (the first year we could run it), following neonatal outcomes, the sample size would be so astronomical that this might be one of those rare situations where I would argue that randomization is unethical due to futility. Maybe it is time to forget about COVID.

David Rind

Pregnant women need to make a choice in the face of inadequate data. I think the evidence points in a direction opposite that of Dr. Prasad’s FAQ answer but, to be clear, I am less certain in my conclusions than he is in his.

Dr. Prasad is clearly not a fan of the VA study or its authors. When it was published, I was startled that the NEJM would publish an observational study and skeptical that it could improve our understanding of COVID vaccine efficacy; then I read the article. I would encourage everyone interested in this question to do the same.

Share

David Rind is an academic primary care physician at Beth Israel Deaconess Medical Center and Chief Medical Officer of the Institute for Clinical and Economic Review. He was previously Vice President, Editorial and Evidence-Based Medicine at UpToDate.

Vinay Prasad is, never mind, you know him.

• The Hubris of Screening for Disease — Again

  ---

Prognostic Value of Blood-Based P-Tau217 Levels for Progression to Cognitive Impairment

• Why Blood Tests Are a Poor Screening Tool for Alzheimer’s

Critical appraisal teaching is the reason to study the POET II trial.

It is the biggest head-shaker of the year. My issue is the author’s choice of primary and secondary endpoints. A rule of evidence is that you choose as the primary endpoint the most important outcome. Secondary outcomes are chosen to bolster or support the primary.

POET II authors got this backward; but readers of the trial can still make conclusions from the evidence—it’s just different from what the authors conclude.

Background: The medical issue and question is the treatment of bacterial endocarditis, which is infection of the heart valves. It’s about as bad as an infection gets. You can easily die form endocarditis because the infection can destroy the valve, or, the nidus for infection in the heart can keep seeding the blood and body with bacteria, and then sepsis can cause death. The other problem with endocarditis is that it can be tough to eradicate because there is often a full-on mass (or vegetation) on a heart valve. Sometimes a surgeon has to cut the valve out and replace it.

The specific question asked in the Danish-led POET II trial was whether a shorter course of antibiotics could be used.

In POET I, the Danish team had shown that switching from IV to oral antibiotics could be done at a point when certain criteria had been met. These criteria were then called the POET criteria, and, basically, were signs that the infection was going in the right direction. Things like falling white count and inflammation markers as well as a lack of fever.

Subscribe now

The POET II trial enrolled only patients with endocarditis who had a) received the standard 2 to 4 weeks of abx, depending on pathogen and complexity of infection and b) met POET criteria and were considered to be stabilized.

Patients were then randomized to either response-tailored therapy (stopping antibiotics) or continuing standard therapy duration. There were about 500 patients; half had the antibiotics stopped and half continued.

The primary endpoint was days alive without antibiotics within 6 months. This was tested for superiority. The primary safety endpoint was death, unplanned heart surgery or embolic events within 6 months, tested for non-inferiority. Relapse of bacteremia or endocarditis was a key secondary endpoint.

Pause there and think for a moment about the primary endpoint. If you are like me, you might wonder why days alive have to be modified by without antibiotics. More on that in a moment.

Results of POET II

• The main primary endpoint of days alive without abx was 183 days in the tailored therapy arm vs 169 days in the standard duration arm. This 13 days difference was statistically significant.

• A primary safety endpoint event occurred in 8.2% with tailored therapy vs 10.7% with standard therapy. ARD: −2.4 percentage points; 95% CI, −7.7 to 2.7; P<0.001 for noninferiority.

• For the key secondary endpoint of relapse of infection: 5.1% tailored vs 1.6% standard therapy. HR 3.3 (1.07-10.1) and p =0.04.

The authors concluded:

Among patients with infective endocarditis on the left side of the heart, the use of a response-tailored antibiotic strategy resulted in a longer time alive without antibiotic therapy than standard-duration therapy and met the criterion for noninferiority with respect to safety but was associated with a higher incidence of relapse of bacteremia or infective endocarditis.

Media and Press Conference at the European Society of Cardiology Congress

News articles had headlines such as this: POET II Supports Shorter Antibiotic Course in Infective Endocarditis.

In the news conference at ESC, POET II primary author, Henning Bundgard said: “It seems that from POET I and POET II we have challenged dogmas… showing that less intensive antibiotic therapy after stabilization is superior and safe,” And we think that defining the stabilization point is a kind of personalizing endocarditis treatment.”

Following the presentation, the trial discussant said “despite some limitations, including the need for more information on the strict clinical stabilization criteria used in the study, the balance favors the shortened antibiotic course.”

He called it a positive study, and antibiotic therapy for endocarditis “should move from the fixed-duration therapy to [a] personalized, tailored treatment decision, but in selected patients and in selected centers with high specialization in the field of endocarditis.”

My Comments

The first thing to say is my surprise that a Danish team could be so wrong. Normally, I see Danes as highly intelligent right-minded people. But this interpretation of the data is about as wrong as it gets.

I covered this on the This Week in Cardiology podcast last Friday, and we covered it here on the Sensible Medicine podcast but I wanted to make sure readers were exposed to the mistakes made in trial interpretation.

When we treat infections, especially serious ones in the heart and bloodstream, there are a couple of important endpoints—and one is not days without antibiotics, but recurrence of infection or death.

And in this trial, the shorter course of abx led to a 3x (up to 10x in the upper bound of the confidence interval) more recurring infection or bacteremia.

Most of us have been patients and been on antibiotics, I surely have, and whether I take 1 week or 2 weeks of tablets is far less important than if I have recurring bacterial infection.

Plus, we are not dealing here with the sniffles. Recurrent bacterial infections is really dangerous. Infection of the heart, too, is not like a urinary tract infection or minor cellulitis.

The bottom line is that you don’t need to be an academic or infectious disease specialist to take from this trial data that the shorter course is clearly the worse strategy.

Why the authors weigh being off antibiotics more highly than not having a recurrent bacterial infection is a matter for speculation. I hope that it is not antibiotic stewardship run amok.

No matter, our job as users of the evidence is to find the best treatment for our patients. In this case it is the standard not shorter course of antibiotics.

Subscribe now

What does make me optimistic are bright medical students actually trying to collect data so we can know the truth about the effects of these changes in healthcare.

Adam Cifu

“Primary care today just isn’t the same as it used to be,” is something I have heard from plenty of primary care physicians. These physicians began their practice in an era that, from where I sit as a second-year medical student, looks almost unrecognizable. Most physicians are employees now, not owners: a recent report found that, as of the start of 2026, roughly 82 percent of practicing physicians were employed by a hospital or other corporate entity, up from around 25 percent in 2012.

Amid the drastic wave of upscaling, consolidation, and corporatization that has swept across US healthcare in recent decades, private equity (PE) has become the most recent villain. And indeed, the existing evidence on PE in healthcare is not encouraging. A widely cited study in JAMA found that hospitals acquired by PE firms saw a 25.4 percent increase in hospital-acquired adverse events relative to matched controls, driven largely by falls and central line infections. Research on procedural specialties has documented meaningful price increases once PE enters a market.

PE can threaten hospitals, physicians, and patients. In a highly publicized case in Massachusetts, for example, the PE-backed Steward Health Care sold hospital land and entered into a sale-leaseback agreement with a real estate investment trust. The capital raised from the sale likely returned to the PE investors rather than being invested in the individual Steward hospitals. Then, facing high rent on the land they sat on, the hospitals fell behind on invoices to medical supply vendors, eventually leading to bankruptcy. A recent paper in BMJ found a pattern between real-estate investment trusts, including PE, acquisition of hospitals, and resulting bankruptcy. There is also evidence that physician turnover rises after PE acquisition. Taken together, this body of literature suggests that PE investments in healthcare may do more harm than good.

However, there’s a twist. PE investors see primary care practices as entirely different businesses than inpatient or specialty care.

My colleagues and mentors at Brown have been investigating whether the overwhelmingly negative findings of PE’s impact on healthcare extend to primary care. In Health Affairs, we found that PE acquisition of primary care practice increased the number of billed services and the number of patients each physician saw. A meaningful share of that increase came from preventive care, particularly the Medicare Annual Wellness Visit. Patients in acquired practices were more likely to see a greater number of distinct physicians and APPs, with the latter growing at a faster rate than in matched controls. In JAMA Health Forum, we looked specifically at acute care outcomes such as all-cause hospitalizations, potentially avoidable hospitalizations, and ED visits in the traditional Medicare population.

We didn’t find any compelling short-term improvement in health outcomes, with the only significant finding being a small decrease in all-cause ED visits. Even this finding did not persist across sensitivity tests nor was it clinically meaningful. And in MCRR, we found no meaningful effect on the delivery of low-value care after PE or hospital acquisition. Our findings on increased utilization and cost of care align with the broader literature that shows PE-driven consolidation is associated with billing for higher-intensity visits, increased patient volume, and higher prices due to a practice’s greater market power.

In summary, PE’s effect on primary care seems to be more billing and more preventive care documentation, without clear harm to acute outcomes or an effect on low-value care.

A few questions that are worth sitting with: Does more billing mean more services were delivered, or does this represent more thorough documentation for the same services that were already being provided? And if it does mean more services, is this clinically beneficial? It’s also probably too early to see PE’s long-run effects on primary care because most of the deals are recent (after 2019). Since most of our evidence comes from traditional Medicare (a fee-for-service system), it tells us less about what happens under Medicare Advantage, in the employer-sponsored insurance space,1 or other value-based payment designs, where the incentives to over-bill or under-treat run in different directions.

Zooming out, capital investment in health care isn’t going anywhere, and I’d argue it’s hard to judge if PE is worse than the hospital-based consolidation that preceded it. PE does not have to be the villain. Capital is essential, and when deployed well, has real power to do good. However, when investors’ focus is on short-term return, typically five to seven years, the incentives don’t always align with what patients – or physicians – need over the long run.

The deeper issue is market power – on both the insurer and provider sides – and our patients ultimately absorb the cost of that power, whether through bills or premiums. The average annual premium for family coverage through an employer hit $26,993 in 2025, up 6 percent from the year before, outpacing both wage growth and inflation. Regardless of who owns the practice down the street, it is ordinary households footing the bill for the system’s growth. The right response is not a blanket ban on PE ownership, but more transparency into all ownership structures, more measured interpretation of the evidence as it accumulates, and less appetite for one-size-fits-all policy before we understand what is happening.

What does this mean for future physicians? Perhaps we won’t know anything different since most of us will likely never experience being a physician/small-business owner, but will I increasingly feel like a cog in a corporate machine? What does this mean for physician retention, job satisfaction, and the quality of a patient-physician relationship? I worry that in a setting where workforce shortages are already acute, consolidation from PE or other large capital investment may further worsen retention in primary care.

I think about my father’s educators, and why primary care physicians of that era are remembered with such nostalgia. It was never really about the tools they had. It was about autonomy and about a physician-patient relationship that wasn’t mediated by anyone’s ROI calculation. Whether the essence of that model of care can survive the current wave of consolidation, PE or otherwise, remains to be seen.

Share

Meehir Dixit is a second-year medical student at the Pritzker School of Medicine, University of Chicago. He obtained a B.A. in Public Health from Brown University, where he spent time during his undergraduate studies and gap year leading and collaborating on various projects within health services research, ranging from private equity in primary care to contract terminations in the Medicare Advantage program.

He is deeply grateful to Dr. David Deshpande, Professor Yashaswini Singh, Dr. Ira Wilson, Margherita Rampichini, and his parents, Dr. Ninad Dixit and Dr. Trupti Dixit, for their crucial support in shaping this essay and his journey thus far.