The Treatment Was Worse Than the Disease

On December 12, 1799, George Washington returned to Mount Vernon after several hours on horseback in snow, hail, and freezing rain. The next day he developed a sore throat and hoarseness, and in the early morning of December 14 he awoke struggling to breathe and swallow. Doctors were called, and what happened next is now one of the most talked-about medical cases in American history.

Washington did not lack medical care. In fact, he received a lot of attention from three skilled physicians, who used the treatments they thought best for someone as sick as he was, such as repeated bloodletting and other methods meant to reduce inflammation and restore balance. Late that night, Washington died.

Even now, more than 200 years later, we are not sure exactly what disease killed Washington. Some experts have suggested acute bacterial epiglottitis, severe pharyngitis, peritonsillar infection, or other causes of upper-airway blockage [1]. What is clear is that Washington lost a large amount of blood, by most estimates about 80 ounces, through repeated bloodletting while he was already very sick. His doctors did not mean to harm him. They thought removing blood would control the inflammation, but it made him worse. Their actions followed a long-standing medical tradition, supported by accepted theories, experience, and the authority of many generations of doctors.

It is easy to look back at this story and feel superior because of our modern knowledge. Today, we understand things like oxygen delivery, blood volume, shock, infections, airway care, IV fluids, antibiotics, and advanced life support. Washington’s doctors did not have this knowledge. But calling them primitive misses the real point. They were trained doctors using the best knowledge they had to try to save their patient. The real question is not why doctors in 1799 believed bloodletting would help. Instead, we should ask which treatments we use today with the same confidence that future doctors might find hard to believe.

Medicine’s Graveyard of Good Ideas

Medicine has always operated between what we know and what we do not. Patients get sick now, but clear scientific answers can take years or even decades to find. Doctors have to watch, guess, treat, and adjust as they go. This process has led to some of our greatest achievements. Anesthesia made surgery possible. Antibiotics turned deadly infections into treatable ones. Insulin changed the outlook for type 1 diabetes. Intensive care, organ transplants, dialysis, cancer treatments, and heart procedures now save people who would not have survived in the past. Admitting mistakes in medicine does not mean ignoring the real progress made.

But history also shows many treatments once thought to be good medicine that are now known to be harmful. Bloodletting is the most famous example because it now seems clearly wrong. For centuries, it fit into the accepted theory that illness came from an imbalance of bodily fluids, so removing blood was supposed to help. Some patients got better after bloodletting simply because many illnesses get better on their own, and these cases made the practice seem effective. When patients died, doctors often blamed the severity of the illness, not the treatment. Without proper studies, doctors’ experiences kept confirming what they already believed.

Mercury provides another sobering example. Mercury compounds were used for centuries to treat syphilis and other diseases despite their substantial toxicity. Records show that mercury was used in many ways, including ointments, pills, injections, and even fumes [2]. Patients often suffered from too much saliva, mouth sores, lost teeth, nerve damage, and other serious problems. Still, some of these side effects were seen as proof that the treatment was working.

The use of mercury lasted so long not because doctors were unintelligent, but because accepted theories, tradition, authority, and stories from practice all supported it. When a practice is new, its advocates must answer the question: why should it be adopted? After widespread adoption, the burden shifts, and skeptics must instead prove why it should be abandoned. Practice becomes habit, habit becomes a standard, and the standard eventually acquires moral and institutional authority.

The physician who follows prevailing practice appears responsible even when evidence is incomplete, while the physician who questions it may appear reckless. That asymmetry helps explain why medical practices can persist even after their scientific foundations begin to weaken.

When More Treatment Meant Better Treatment

Few episodes demonstrate the seductive logic of aggressive treatment better than the history of breast cancer surgery. For much of the 20th century, the Halsted radical mastectomy represented the dominant surgical approach to breast cancer. The procedure removed the breast, pectoral muscles, and extensive regional lymphatic tissue. Within the prevailing model of cancer biology, the reasoning seemed compelling: if cancer spread progressively outward from a primary tumor, then removing more tissue should provide better local control and therefore better survival.

The surgery was impressive anatomically, but it was very hard on patients. Women often ended up with severe disfigurement, loss of function, swelling, pain, and emotional trauma. The real question was not whether the radical mastectomy removed more tissue (it clearly did) but whether taking out more tissue actually led to better results for patients.

Randomized trials eventually challenged the assumption. The National Surgical Adjuvant Breast and Bowel Project B-04 trial compared radical mastectomy with less extensive approaches, and at 25 years of follow-up investigators found no significant survival advantage from the radical procedure in the populations studied [3]. The NSABP B-06 trial subsequently demonstrated that breast-conserving surgery followed by irradiation could provide long-term survival comparable to mastectomy for appropriately selected women with invasive breast cancer [4]. Long-term findings from the Milan trial similarly demonstrated that breast-conserving surgery could achieve survival comparable to radical mastectomy in women with small breast cancers [5].

This lesson applies to much more than breast cancer. Medicine has often mixed up doing more with doing better. We still fall into this trap. More tests can lead to false alarms and unnecessary findings. More drugs can cause side effects and bad interactions. Pushing for stricter targets can create new problems. More invasive procedures can harm patients without helping them live longer or better. Medicine has limits, and sometimes the smartest move is to recognize that more treatment is not always better.

When the Monitor Improved and the Patient Died

Perhaps the clearest modern example of an intervention successfully correcting a physiological abnormality while worsening the outcome comes from cardiology. Physicians recognized that premature ventricular depolarizations after myocardial infarction were associated with an increased risk of sudden cardiac death. The reasoning that followed appeared almost irresistible. If ventricular ectopy identified patients at greater risk, and antiarrhythmic drugs could suppress ventricular ectopy, then suppressing those arrhythmias should reduce mortality.

The drugs performed their intended physiological function extremely well. Encainide and flecainide suppressed ventricular ectopy, and the electrocardiographic abnormality improved. Unfortunately, the patients did not. The Cardiac Arrhythmia Suppression Trial (CAST) demonstrated that patients receiving encainide or flecainide experienced substantially more arrhythmic deaths and nonfatal cardiac arrests than patients receiving placebo, despite successful suppression of the arrhythmias physicians were trying to eliminate [6]. Subsequent reporting confirmed the excess mortality associated with these drugs in this population [7].

CAST should remain required reading for physicians because it exposes one of medicine’s most persistent intellectual traps: confusing a surrogate endpoint with a patient-centered outcome. The association between ventricular ectopy and mortality was real, and the ability of the drugs to suppress that ectopy was also real. The problem was the assumption connecting those two observations. Suppressing the marker did not eliminate the underlying risk; instead, it introduced a new lethal risk. The monitor looked better while the patient became more likely to die.

Modern medicine produces huge amounts of data. We track blood pressure, oxygen, blood sugar, kidney function, cholesterol, inflammation, heart function, tumor size, lung capacity, bone density, and many other numbers. These measurements have changed medicine, but they can also make us focus on numbers instead of people. A test result can show risk without causing it. An abnormal value might be linked to bad outcomes, but fixing the number does not always help the patient. In the end, the patient—not the lab result or scan—matters most.

A Lesson from the Intensive Care Unit

Critical care provides another striking example. Hyperglycemia is common during severe illness and has repeatedly been associated with adverse outcomes. An influential randomized trial published in 2001 reported reduced morbidity and mortality among surgical intensive care patients treated with intensive insulin therapy [8]. The biological reasoning was attractive: hyperglycemia was associated with poor outcomes, insulin lowered glucose, and maintaining near-normal glucose seemed more physiological than allowing substantial hyperglycemia.

The findings had an enormous influence. Intensive care units created new protocols, nurses adjusted insulin doses, and blood sugar was checked often. Tight blood sugar control became a key part of ICU care in many hospitals. This pattern was familiar: an observation led to a theory, a study supported it, and soon the idea became standard practice. Then, in 2009, the NICE-SUGAR trial, which randomized more than 6,000 critically ill adults, found that intensive glucose control increased 90-day mortality and substantially increased severe hypoglycemia [9]. Once again, medicine confronted the difference between correcting an abnormal physiological value and improving the patient.

For someone like me, who has cared for critically ill patients for decades, this lesson is especially important. The body’s responses during shock, infection, trauma, breathing failure, and organ problems are extremely complex. Some changes are harmful, some may help, and some are good in one situation but bad in another. In the ICU, it is tempting to try to fix every abnormal number because so much can be measured and changed. But people are not machines, and not every value needs to be pushed back to normal.

This does not mean that understanding the body is unimportant or that studies answer every question. It means that theory, experience, and research all need to challenge each other. Focusing only on how the body works can mislead us, and using study results without thinking about real patients can also go wrong. Good medicine happens in the space between these approaches.

When Common Sense Was Wrong

Some practices persist not because of sophisticated biological theories but because they simply seem like common sense. Prolonged bed rest after myocardial infarction was one such practice. If the heart had been injured, minimizing physical exertion seemed prudent. Patients recovering from acute coronary events were therefore historically confined to bed for extended periods, sometimes for weeks.

The approach had intuitive appeal but carried high physiological costs. Immobilization promotes deconditioning, venous stasis, loss of muscle mass, orthostatic intolerance, pulmonary complications, and psychological dependence. By the middle of the 20th century, physicians including Samuel Levine and Bernard Lown challenged prolonged immobilization and advocated earlier chair rest and mobilization after coronary thrombosis [10]. Modern cardiac care eventually moved dramatically away from the assumption that prolonged inactivity was inherently protective.

This lesson is not always obvious. Harmful medicine is not always about doing too much. Sometimes, the mistake is holding back too much instead of giving an unnecessary drug or surgery. Medicine can overdo things with treatments, but also by keeping patients in bed, monitoring too much, making too many rules, running too many tests, or treating normal behavior as illness. The common thread is not aggression but certainty without sufficient proof.

When Established Medicine Reverses Itself

Medicine’s ability to reverse course is one of its greatest strengths, but medical reversal also reveals an uncomfortable reality. Practices can become widely established before definitive evidence demonstrates that they are ineffective or harmful. Prasad and colleagues examined a decade of publications in a major medical journal and identified 146 examples in which new evidence contradicted established medical practices [11]. Such studies cannot determine the exact frequency of reversal across all of medicine because the denominator depends heavily on how practices are selected and defined, but the underlying phenomenon is indisputable.

Medical practices are often reversed for several reasons. Early studies may be small, observational, or limited to specific patient groups. Sometimes, doctors focus on test results instead of real outcomes. Treatments can catch on because they seem logical, and excitement can spread faster than solid evidence. Once a practice is common, jobs, hospital rules, insurance payments, performance measures, and patient expectations can all depend on it. At that point, stopping the practice is often harder than starting it was.

A careful review of studies on established medical practices shows that reversals happen in modern medicine too, not just with old or rare treatments [12]. The key point is not that doctors should doubt everything. Total distrust is just as unreasonable as blind faith. The real lesson is that calling something the “standard of care” does not mean it cannot be wrong.

A Brief Lesson from Covid-19

Covid-19 showed how medicine and public health must sometimes act with great uncertainty. A new virus was spreading quickly, hospitals were overwhelmed, and choices had to be made before solid evidence was available. It made sense to be cautious. But it is worth looking at how quickly some temporary recommendations started to seem more certain than the evidence really allowed.

Physical distancing illustrates the problem. A major 2020 systematic review examining SARS-CoV-2, SARS-CoV, and MERS-CoV supported physical distancing and suggested that greater distance could provide greater protection, but the evidence was entirely observational, with no randomized trials, and the authors acknowledged important limitations [13]. The familiar six-foot rule nevertheless became a remarkably precise rule in public life. Later evidence emphasizing aerosol and airborne transmission indicated that risk could not be reduced to a fixed radius around each person; particularly indoors, ventilation, exposure duration, crowding, and vocalization were also critically important [14].

The same need for nuance applied to masks. Randomized evidence was more complicated than either side of the subsequent political debate often acknowledged. The DANMASK-19 trial did not demonstrate a statistically significant reduction in infection among people advised to wear surgical masks in the studied setting, although its confidence intervals did not exclude benefit and the trial was not designed primarily to evaluate source control [15].

A cluster-randomized trial in Bangladesh found that an intervention promoting community masking produced a modest reduction in symptomatic seroprevalence, with the clearest benefit associated with surgical masks [16]. A later Cochrane review found substantial uncertainty and methodological limitations across randomized studies of physical interventions for respiratory viruses [17].

The point is not that distancing or masks never help reduce spread. The point is that emergency measures should clearly show the uncertainty behind them. A precaution is not a proven law of biology. It should have been normal to ask whether six feet was better than three, whether outdoor masking made sense, whether cloth and surgical masks were the same, or whether mandates were worth their costs. Science gets stronger, not weaker, when new evidence questions temporary recommendations.

When Guidelines Become Commandments

Clinical guidelines are among modern medicine’s great achievements. No physician can independently evaluate every paper relevant to every disease, and well-constructed guidelines synthesize enormous bodies of knowledge into practical recommendations. Used appropriately, they improve care and reduce unnecessary variation. The problem begins when a guideline stops guiding and starts commanding.

This change usually happens slowly. First, someone notices something and forms a theory. Studies support the theory, then professional groups make recommendations. Hospitals turn these into protocols, electronic records turn protocols into checklists and alerts, insurers tie payments to following the rules, and administrators use them to measure performance.

In the end, a conclusion based on limited evidence can start to feel like a hospital law. Clinicians begin asking, “What does the protocol require?” rather than “What does this patient need?” Those questions frequently produce the same answer, but not always. Guidelines describe populations, whereas physicians treat individuals. Age, comorbidities, competing risks, unusual physiology, prior treatment, patient preference, and goals of care can make a generally appropriate recommendation inappropriate for a particular person.

The original formulation of evidence-based medicine explicitly integrated the best available external evidence with individual clinical expertise [18]. Evidence-based medicine was never intended to eliminate judgment. Yet bureaucratic healthcare systems can turn evidence-based medicine into compliance-based medicine, where adherence to a metric matters more than understanding why the metric exists.

After many years caring for very sick patients, I have found that the hardest cases are often those that do not fit the textbook. The patient with sepsis does not always follow the algorithm. An older patient with several chronic illnesses cannot be treated as if each disease is separate. Real medicine starts when the guidelines are not enough.

The Modern Bloodletting Problem

Doctors in the past used tools like lancets, leeches, mercury, and major surgeries. Today, we have computerized checklists, advanced scans, lab markers, genetic testing, artificial intelligence, complex models, and endless clinical guidelines. Our tools are much more powerful, but better technology does not change human nature. In fact, technology can sometimes spread mistakes faster than ever before.

If one doctor uses a questionable treatment, only that doctor’s patients are affected. But if questionable treatment is built into a national care pathway, it can affect millions. If a recommendation is added to electronic records, it can show up in hundreds of hospitals at once. Standardization is very helpful when the standard is right. But if the standard is wrong, mistakes spread quickly.

This is not a criticism of technology, guidelines, public health, research, or evidence-based medicine. It is a warning against abandoning critical thinking. Skepticism is not bad for science; careful questioning is one of its most important tools. Every recommendation should be open to change when new evidence appears, and every doctor should keep enough independence to ask if what they are doing is truly helping the patient.

The Cost of Doing Something

Medicine rewards action. Patients understandably seek physicians because they want help, and physicians are trained to identify problems and intervene. Hospitals are organized around interventions, and healthcare reimbursement has historically rewarded procedures, tests, visits, and treatments more readily than thoughtful restraint. There is also enormous emotional pressure to act when a patient is deteriorating.

Critical care physicians like me understand this impulse particularly well. We intubate, ventilate, dialyze, cardiovert, resuscitate, infuse, drain, monitor, and support failing organs. These interventions save lives every day. During my career, I have watched patients survive diseases and degrees of organ failure that would have been uniformly fatal only decades earlier. Anyone who uses the history of medical error to argue that modern medicine does not work has learned precisely the wrong lesson from that history.

Still, experience shows that every treatment has a cost. Breathing machines can save lives but also damage lungs. Central lines allow lifesaving treatments but can cause infections, clots, and other problems. Blood thinners prevent clots but increase bleeding risk. Antibiotics cure infections but can cause side effects, promote resistance, and disrupt the microbiome. Sedatives help patients tolerate breathing tubes but can lead to confusion and weakness. Scans can find treatable problems but also lead to more tests, procedures, worry, and expense.

So, the real question in medicine is not just whether we can do something, but whether doing it will actually help the patient. This seems obvious in writing, but history shows it is much harder to remember in real-life care.

What Will Physicians Say about Us?

Imagine a physician in 2076 studying medicine as practiced in 2026. Which of our treatments will seem primitive? Which operations will appear unnecessarily aggressive? Which drugs will seem astonishingly overprescribed? Which laboratory targets will appear arbitrary? Which screening programs will be judged to have generated more procedures than meaningful benefit? Which diagnostic categories will have disappeared because we understand disease biology completely differently?

We do not know the answers, and that is exactly why asking the question is important. Admitting uncertainty is not a weakness. It shows discipline and thoughtfulness. History teaches us to be wary when someone claims that all the big questions are settled, because progress often comes from reopening those questions.

This does not mean that all knowledge is temporary or that every medical recommendation should be doubted equally. Modern medicine has many treatments with strong proof behind them. Antibiotics cure bacterial infections. Insulin saves lives in type 1 diabetes. Restoring blood flow saves heart muscle. Vaccines have prevented huge numbers of infections. Trauma surgery saves people who are bleeding, and cancer treatments now cure some cancers that used to be fatal. The lesson is not to lose confidence in medicine, but to stay humble about what is truly proven and what is just likely, traditional, popular, or convenient.

Humility As a Clinical Skill

People often call humility a virtue or a personality trait. In medicine, it should be seen as a skill. Humility shapes how doctors read evidence, listen to patients, react to surprises, rethink diagnoses, and change treatments when they realize they were wrong. A doctor who cannot admit the chance of being wrong can become dangerous.

The best doctors I have known combined decisiveness with humility. They could act quickly when a patient was in danger, but also change their minds just as quickly when new facts came in. They knew that changing your mind because of new evidence is not a weakness. This is what science is about. On the other hand, when certainty becomes part of a doctor’s identity, any challenge to a treatment can feel like a personal attack.

Medical training should teach this lesson clearly. Students spend years learning huge amounts of information, often as if every answer is final because tests demand it. But real patients do not always fit the textbook. They live in the gray areas between chapters, studies, guidelines, and their own biology. Learning to handle uncertainty without freezing up may be one of the most important skills a doctor can learn.

Back to Mount Vernon

George Washington died late on December 14, 1799, with doctors around him who tried hard to save his life. They used the treatments they had and followed what they believed about disease at the time. Their intentions were good; they were confident, and their methods were accepted by their peers. Looking back, we can see how their treatments may have made his condition even worse.

It is tempting to think that modern medicine could no longer make mistakes of this kind. We know much more about disease, our science is stronger, and our treatments would seem miraculous to Washington’s doctors. But technology has advanced faster than human nature. We are still influenced by authority, trends, clever theories, stories, test results, money, workplace pressure, and the comfort of agreeing with others. They believed they were practicing good medicine, just as we do.

The lesson is not to look down on Washington’s physicians. Instead, we should recognize ourselves in them and preserve the capacity to question even our most comfortable assumptions. We should embrace treatments supported by strong evidence, use guidelines intelligently, employ technology enthusiastically when it improves outcomes, and act decisively when a patient’s life depends upon intervention. At the same time, we must remain willing to observe, reconsider, revise, and occasionally admit that what everyone knows to be true may not be true after all.

Right now, there is almost certainly a treatment being used just because everyone assumes it works, a lab result being fixed because everyone thinks abnormal means bad, a drug being continued because stopping it seems risky, or a protocol being followed because it feels uncomfortable to question it. We do not know which of these future doctors will find surprising. That is exactly why humility matters.

Washington’s doctors truly wanted to save him. They were trained, experienced, and sure that doing something was better than doing nothing. Now, more than 200 years later, we know that good intentions, common practice, and logical theories are not enough. In the end, medicine must always ask the hardest question: are we really helping the patient? Sometimes, the answer changes over time. When it does, doctors must be ready to change too, because history shows that the line between helping and harming can be much thinner than we think.

References

1. Mackowiak PA. George Washington’s recurrent health problems and fatal infection re-examined. Clin Infect Dis. 2021;72(10):1850-1853.
2. Tilles G, Wallach D. Le traitement de la syphilis par le mercure: une histoire thérapeutique exemplaire [in French]. Hist Sci Med. 1996;30(4):501-510.
3. Fisher B, Jeong JH, Anderson S, Bryant J, Fisher ER, Wolmark N. Twenty-five-year follow-up of a randomized trial comparing radical mastectomy, total mastectomy, and total mastectomy followed by irradiation. N Engl J Med. 2002;347(8):567-575.
4. Fisher B, Anderson S, Bryant J, Margolese RG, Deutsch M, Fisher ER, et al. Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for invasive breast cancer. N Engl J Med. 2002;347(16):1233-1241.
5. Veronesi U, Cascinelli N, Mariani L, Greco M, Saccozzi R, Luini A, et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. N Engl J Med. 2002;347(16):1227-1232.
6. Cardiac Arrhythmia Suppression Trial (CAST) Investigators. Preliminary report: effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Engl J Med. 1989;321(6):406-412.
7. Echt DS, Liebson PR, Mitchell LB, Peters RW, Obias-Manno D, Barker AH, et al. Mortality and morbidity in patients receiving encainide, flecainide, or placebo: the Cardiac Arrhythmia Suppression Trial. N Engl J Med. 1991;324(12):781-788.
8. Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, et al. Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345(19):1359-1367.
9. NICE-SUGAR Study Investigators. Intensive versus conventional glucose control in critically ill patients. N Engl J Med. 2009;360(13):1283-1297.
10. Levine SA, Lown B. “Armchair” treatment of acute coronary thrombosis. JAMA. 1952;148(16):1365-1369.
11. Prasad V, Vandross A, Toomey C, Cheung M, Rho J, Quinn S, et al. A decade of reversal: an analysis of 146 contradicted medical practices. Mayo Clin Proc. 2013;88(8):790-798.
12. Herrera-Perez D, Haslam A, Crain T, Gill J, Livingston C, Kaestner V, et al. A comprehensive review of randomized clinical trials in three medical journals reveals 396 medical reversals. eLife. 2019;8:e45183.
13. Chu DK, Akl EA, Duda S, Solo K, Yaacoub S, Schünemann HJ, et al. Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta-analysis. Lancet. 2020;395(10242):1973-1987.
14. Duval D, Palmer JC, Tudge I, Pearce-Smith N, O’Connell E, Bennett A, et al. Long distance airborne transmission of SARS-CoV-2: rapid systematic review. BMJ. 2022;377:e068743.
15. Bundgaard H, Bundgaard JS, Raaschou-Pedersen DET, von Buchwald C, Todsen T, Norsk JB, et al. Effectiveness of adding a mask recommendation to other public health measures to prevent SARS-CoV-2 infection in Danish mask wearers: a randomized controlled trial. Ann Intern Med. 2021;174(3):335-343.
16. Abaluck J, Kwong LH, Styczynski A, Haque A, Kabir MA, Bates-Jefferys E, et al. Impact of community masking on COVID-19: a cluster-randomized trial in Bangladesh. Science. 2022;375(6577):eabi9069.
17. Jefferson T, Dooley L, Ferroni E, Al-Ansary LA, van Driel ML, Bawazeer GA, et al. Physical interventions to interrupt or reduce the spread of respiratory viruses. Cochrane Database Syst Rev. 2023;1(1):CD006207.
18. Sackett DL, Rosenberg WMC, Gray JAM, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn’t. BMJ. 1996;312(7023):71-72.