Chapter 1
When Medicine Reverses Course: The Surprising Truth About Medical Treatments
The medical world is no stranger to dramatic reversals. Imagine following your doctor's advice for years, only to discover the treatment was ineffective or even harmful all along. This phenomenon-medical reversal-affects millions of patients annually and costs billions in healthcare dollars. Harvard-trained physicians Vinayak Prasad and Adam Cifu tackle this troubling issue in their eye-opening book, which has become required reading in many medical schools. The work has garnered attention from major medical journals and mainstream publications like The New York Times, with many physicians citing it as transformative to their practice. What makes this book particularly compelling is how it challenges the fundamental assumption that modern medicine consistently follows evidence-based principles. Through numerous examples spanning every medical specialty, the authors demonstrate how frequently doctors continue using practices for decades before discovering they don't work-and offer practical solutions for both patients and the medical system.
Chapter 2
The Anatomy of Medical Reversal: When Yesterday's Cure Becomes Today's Mistake
Medical reversal occurs when a currently accepted therapy is overturned-found to be no better than the therapy it replaced or no better than doing nothing at all. This phenomenon represents not the natural progression of medicine where superior treatments replace effective ones, but rather the sobering discovery that established practices were never truly effective to begin with. The impact of these reversals extends beyond individual treatments to shake the very foundations of medical practice and patient trust.
A compelling example is flecainide, widely prescribed in the early 1990s to prevent sudden cardiac death by suppressing extra heartbeats. The logic seemed ironclad and followed a seemingly rational progression: premature ventricular contractions (PVCs) predict death, and flecainide effectively suppresses PVCs. However, the landmark 1992 CAST trial delivered a shocking revelation: flecainide actually increased mortality rates by 2.5 times despite successfully reducing PVCs. Similarly, atenolol, a beta-blocker enthusiastically prescribed for hypertension for decades, was shown in 2004 to lower blood pressure effectively but failed to achieve its primary goal - extending patients' lives. These cases highlight how intermediate endpoints (like lower blood pressure) don't always translate to meaningful clinical outcomes.
Medical procedures, despite their more invasive nature and higher costs, face equally troubling scrutiny. Coronary stents for stable angina - tiny metal scaffolds inserted to prop open narrowed heart arteries - represent a particularly controversial case. While patients like Anthony Baker often report feeling better after the procedure, comprehensive research shows these individuals are no less likely to have heart attacks, won't live longer, and typically experience a return of chest pain within twelve months. The COURAGE trial, involving 2,287 patients, definitively demonstrated that stents offered no advantage over optimal medical therapy alone. Vertebroplasty for spinal fractures offers another sobering example - initially celebrated as miraculous based on compelling patient testimonials but later questioned when rigorous studies showed it performed no better than sham procedures where patients received only an injection of salt water.
The systematic nature of medical reversal becomes apparent through comprehensive analysis. When the authors examined every article in the New England Journal of Medicine from 2001-2010 testing current standards of care, they uncovered a disturbing pattern: of 363 articles examining established practices, 146 (40%) demonstrated the standard of care was ineffective. This wasn't an isolated finding - parallel research from the British Medical Journal revealed that approximately 50% of medical practices have unknown effectiveness, while a review in Mayo Clinic Proceedings identified 146 medical reversals in recent years.
The implications extend far beyond academic concern. These reversals cost billions in wasted healthcare spending, expose patients to unnecessary risks, and erode public trust in medical institutions. For practicing physicians, they create a challenging paradox: how to balance being current with being cautious, knowing that today's breakthrough might become tomorrow's cautionary tale. This realization should concern not just medical professionals but anyone who interacts with the healthcare system, highlighting the critical importance of evidence-based medicine and continuous scientific scrutiny of accepted practices.
Chapter 3
When Feeling Better Isn't Enough: The Deceptive Power of Subjective Outcomes
Medical reversals involving subjective outcomes like pain relief create the strongest disbelief among patients and doctors. When treatments meant to improve how people feel are shown ineffective, patients often respond with "What do they know? It worked for me," while doctors insist "I know my patients feel better." This skepticism stems from the powerful experience of improvement contradicting scientific findings.
Consider arthroscopic knee surgery for degenerative meniscal tears. Despite 700,000 procedures performed annually in America at a cost of $4 billion, 2013 studies showed the surgery was no more effective than physical therapy alone or even sham surgery where doctors merely pretended to operate. This widely trusted procedure demonstrates the remarkable power of the placebo effect-improvements in symptoms without physiological intervention.
The placebo effect has been scientifically validated and even shown to involve real physiological changes, including the brain's release of endorphins (natural morphine-like substances). However, placebo effects vary widely between individuals, are most common with pain treatments, and tend to be relatively short-lived.
This phenomenon extends beyond pain to other subjective symptoms. In a New England Journal of Medicine study on asthma treatments, researchers compared albuterol inhalers against placebo inhalers, sham acupuncture, and no intervention. While only albuterol objectively improved lung function, patients reported equal subjective improvement from all three active interventions compared to doing nothing.
This raises uncomfortable questions about placebos in modern medicine. The authors argue that patients seek proven therapies, not treatments that "might work," and that placebo interventions often carry unacceptable costs, risks of harm, or delay effective treatments. Harvard professor Ted Kaptchuk's research challenges the assumption that deception is necessary for placebo effects. In his study on irritable bowel syndrome, patients knowingly given sugar pills labeled as placebos with "self-healing properties" showed significant symptom improvement compared to those receiving no treatment.
The lesson is clear: in each reversal involving subjective outcomes, patients genuinely felt better-but from the idea of treatment rather than the treatment itself. This becomes problematic when placebo therapies are expensive, carry risks, involve deception, or delay more beneficial treatments.
Chapter 4
The Numbers Game: Why Surrogate Endpoints Often Mislead
Surrogate endpoints are measurable markers used as substitutes for clinical endpoints that matter to patients. While objective, these markers are invisible to patients themselves-like HbA1c levels for diabetes, blood pressure readings for hypertension, or cholesterol numbers for heart disease. They're popular in research because they're easier, faster, and less expensive to study than clinical outcomes like heart attacks, strokes, or death. This appeal has led to their widespread adoption in clinical trials, particularly in areas where measuring final outcomes could take years or decades.
The fundamental problem is that improving a surrogate marker doesn't necessarily translate to improved health outcomes. Take cholesterol management-while statins successfully lower cholesterol and demonstrably save lives, other cholesterol-improving drugs have failed spectacularly. Extended-release niacin and fenofibrate, for instance, both improved lipid profiles impressively in laboratory tests but failed to reduce actual cardiovascular events in large clinical trials. The ACCORD and AIM-HIGH trials showed these drugs changed the numbers but didn't reduce heart attacks, strokes, or deaths, despite billions spent on prescriptions.
Bevacizumab (Avastin) for metastatic breast cancer provides a particularly compelling illustration of this problem. It received accelerated FDA approval in 2008 based on impressive progression-free survival improvements-from 5.9 to 11.8 months. The scans looked better, tumors shrank, and blood markers improved. However, actual survival barely changed (26.7 vs. 25.2 months), and patients experienced significant side effects including hemorrhage and heart failure. Three years later, the FDA made the rare decision to reverse its approval after additional trials confirmed that while bevacizumab improved CT scan measurements, patients weren't living longer or better lives.
Even seemingly logical surrogates like blood flow measurements can be misleading. For patients suffering cardiogenic shock after heart attacks, intra-aortic balloon pumps became standard treatment based on their ability to improve cardiac output and hemodynamic measurements. The theory was elegant - better blood flow should mean better survival. Despite inclusion in international guidelines for decades, the 2012 IABP SHOCK II trial of 600 patients showed they made absolutely no difference in mortality or any other meaningful outcome, leading to a dramatic change in practice recommendations.
Hospitalization represents another deceptively complex endpoint-both undesirable in itself and frequently used as a surrogate for mortality. Lars Hemkens and colleagues analyzed hundreds of trials and discovered that in one-third of studies, treatments that decreased hospitalization actually increased mortality or vice versa. For example, certain heart failure medications reduced hospital admissions while showing concerning trends in mortality. This surprising finding challenges the common assumption that reducing hospitalizations necessarily indicates better patient outcomes.
No perfectly reliable surrogate endpoint yet exists, and recent history provides ample reason for skepticism. The path to preventing medical reversal requires rigorous evidence that practices improve hard clinical endpoints that matter to patients, not just surrogate markers that are easier to measure. This means longer, larger, and more expensive trials - but also more reliable results that truly guide better patient care.
Chapter 5
Screening Dilemmas: Finding Disease That Might Never Matter
Christopher D'Amico's experience with PSA testing left him conflicted. Initially diagnosed with prostate cancer through a PSA test that led to robotic surgery, he later learned that expert recommendations had shifted dramatically-from routine screening to having risk-benefit conversations, and finally to avoiding PSA testing altogether. This left him wondering whether the test had saved his life or unnecessarily diminished its quality through the side effects of incontinence and impotence.
Unlike medical therapies, screening tests affect millions of healthy people, potentially turning them into patients unnecessarily. For a screening test to be truly effective, it must: 1) find cancers early, 2) reduce deaths from the specific cancer, and 3) improve overall survival. While most screening tests accomplish the first goal, evidence for the second and third is surprisingly weak.
PSA testing may reduce prostate cancer deaths but hasn't shown overall survival benefits. Mammography shows variable benefits for breast cancer mortality but no proven improvement in overall survival. Colonoscopy remains untested in randomized trials, though related screening methods show reduced cancer-specific mortality. Only CT screening for lung cancer in heavy smokers has demonstrated improved overall survival, though with high false-positive rates.
The key problem is overdiagnosis-when screening detects cancers that would never have caused symptoms or death. Currently, we cannot distinguish between deadly cancers and harmless ones. For prostate cancer, we treat approximately 40 cancers to save one life. For breast cancer, only about 13% of cancers found by mammography would have been life-threatening.
Population data reveal critical insights about screening effectiveness. When screening works properly, we should see increased early cancer detection followed by decreased advanced cancer rates. However, real-world evidence shows concerning patterns: after mammography and PSA testing became widespread, early cancer diagnoses rose dramatically and stayed elevated, while advanced cancer rates declined only minimally. For breast cancer, early diagnoses increased by 132 cases per 100,000 women annually, while advanced cancers decreased by only 8 cases-suggesting significant overdiagnosis.
These patterns indicate we're finding many cancers that would never have caused harm, subjecting patients to unnecessary treatments and anxiety without proportional benefits.
Chapter 6
Systems Failure: When Healthcare Protocols Don't Deliver
Systems interventions represent a relatively new focus in medicine, recognizing that good care requires more than just a skilled doctor but an entire healthcare ecosystem. These interventions include protocols like contact precautions, checklists, procedural rules, quality measures, and economic incentives. Unlike businesses that constantly evaluate and abandon ineffective systems changes, healthcare often adopts these interventions without adequate study plans, making them difficult to modify even when evidence shows they don't work.
Carl Murphy, a police veteran battling leukemia, found himself subjected to contact precautions after testing positive for vancomycin-resistant enterococcus (VRE) colonization. The protocol required healthcare workers to don gowns and gloves before entering his room-a process that visibly frustrated staff, was inconsistently followed, and seemed to reduce how often doctors visited him. Though Carl tolerated this intervention believing it protected other patients, studies published after his death showed these precautions don't actually decrease VRE or MRSA transmission in hospitals.
Many systems interventions exist in a gray zone-adopted based on weak evidence, persisting for years while being increasingly questioned, but never conclusively rejected. The "four-hour rule" for antibiotic administration in pneumonia patients became a national quality measure without rigorous testing, leading to unintended consequences like treating patients who didn't have pneumonia. Similarly, the door-to-balloon time initiative for heart attack patients showed dramatic improvements in timing metrics but no change in mortality rates, despite massive implementation costs.
The intensive insulin therapy in ICUs case exemplifies systems intervention reversal. In 2001, a single-center unblinded trial showed normalizing blood sugar improved survival in surgical ICU patients. Professional societies quickly embraced strict blood-sugar targets, and hospitals implemented quality interventions to enforce these standards. When rigorously tested in 2009 through a 42-hospital randomized trial with over 6,000 patients, the strict blood-sugar control actually increased deaths by 2.6 percentage points at 90 days-meaning one death for every 40 patients treated.
Despite catchy phrases about "the science of systems," we've failed to learn that systems interventions, like medical treatments, require rigorous evidence before adoption.
Chapter 7
Self-Prescribed Reversals: When Our Own Choices Lead Us Astray
While previous sections focused on doctor-recommended treatments that don't work, we also readily find ineffective remedies ourselves. Americans increasingly use self-prescribed treatments-vitamins, minerals, dietary supplements, and alternative procedures like acupuncture. Over half of all Americans use dietary supplements, spending more than $30 billion annually on these products, making these potential reversals particularly impactful since they affect anyone who chooses them, not just those with specific medical conditions.
Many popular complementary therapies have been reversed through rigorous testing. Glucosamine and chondroitin, despite $700 million in sales by 2004, showed no improvement in joint pain compared to placebo in a 2006 trial of 1,500 people and in a 2010 analysis of ten trials. The supplements failed to outperform sugar pills even among patients with severe arthritis who were considered most likely to benefit. Echinacea, used by nearly 20% of Americans for immune support, failed to reduce cold duration in a 2005 study where volunteers were exposed to cold virus. Even at doses far exceeding recommended amounts, the herb showed no meaningful effect on infection rates or symptom severity.
Acupuncture, despite its 3,000-year history and widespread acceptance in Eastern medicine, showed "little truly convincing evidence" of pain reduction in comprehensive reviews. While some studies suggest modest benefits for specific conditions like lower back pain and migraines, these effects are often indistinguishable from placebo responses and careful attention from practitioners. Multivitamins, taken by over 40% of postmenopausal women in one study, showed no benefit for cancer, heart disease, stroke, or mortality in both observational and randomized trials spanning decades and involving hundreds of thousands of participants.
People adopt complementary therapies through a process similar to traditional medical treatments, driven by compelling marketing and word-of-mouth testimonials. These remedies typically begin with an engaging story, satisfy needs unmet by traditional medicine, and have influential supporters and profit-motivated promoters. Unlike highly regulated medical treatments, complementary therapies in the U.S. face no requirement to prove effectiveness before reaching the market, thanks to the 1994 Dietary Supplement Health and Education Act.
Dietary recommendations are particularly confusing because they're nearly impossible to study properly due to the complexity of human nutrition and difficulty maintaining controlled conditions over long periods. Even the PREDIMED trial, which showed modest benefits of a Mediterranean diet in preventing strokes, required thousands of high-risk participants followed for years to demonstrate a small effect. The study found only three fewer cardiovascular events per thousand people per year. Michael Pollan's simple advice-"Eat food, mostly plants, and not too much"-offers reasonable guidance in this uncertain landscape, emphasizing whole foods over processed products and moderation over strict rules.
The supplement industry's rapid growth, combined with minimal regulation and oversight, creates perfect conditions for ineffective treatments to flourish. Many products gain popularity through celebrity endorsements and clever marketing rather than scientific evidence. This phenomenon demonstrates how readily we embrace unproven remedies when they're packaged with compelling narratives and promises of natural healing.
Chapter 8
Protecting Yourself: A Patient's Guide to Avoiding Medical Reversal
The medical system's structure and culture makes it likely you'll be offered unproven therapies that may eventually be reversed. "White-coat silence" occurs when normally informed, confident people fail to ask important questions in doctors' offices, often due to anxiety, time pressure, or perceived power dynamics. This phenomenon affects patients across all education and socioeconomic levels, making self-advocacy particularly important.
To protect yourself, follow these detailed steps when offered medical treatments:
1) Identify which endpoints matter to you-usually morbidity (will this relieve symptoms or prevent disability?) and mortality (will this help me live longer?). Don't be misled by surrogate endpoints like blood sugar numbers or tumor shrinkage. For example, while a diabetes medication might lower blood sugar effectively, what matters is whether it prevents heart attacks or extends life. Similarly, some cancer treatments might shrink tumors without improving survival or quality of life.
2) Ask what type of studies show the treatment improves these endpoints. Randomized trials across multiple hospitals provide the strongest evidence, especially those with long-term follow-up. Consider whether the trial participants were similar to you in age, gender, and health conditions. Pay attention to who funded the research - industry-sponsored studies tend to show more favorable results than independent research. Request specific study references and, if possible, review them yourself or with another healthcare provider.
3) Ask about the "number needed to treat" (NNT)-how many people must receive this treatment for one person to benefit. Even our best therapies often require treating 20-50 patients to help one. For instance, statins for primary prevention might have an NNT of 50, meaning 49 people take the medication without benefit. Understanding this helps set realistic expectations and makes informed decisions easier.
4) When making your decision, carefully weigh the likelihood of benefit against potential side effects, costs, and lifestyle impact. Consider both short-term and long-term implications. For preventive treatments in healthy people, be especially skeptical of unproven interventions. Remember that aggressive screening and prevention can lead to overdiagnosis and unnecessary treatments.
5) Find a doctor whose philosophy matches yours and who's willing to answer questions. Use their experience ("How long have you been using this treatment?" "What results have you personally observed?") and expertise (ask to see relevant studies). Look for physicians who practice evidence-based medicine and are comfortable discussing uncertainties. Don't hesitate to seek second opinions for major decisions.
The authors emphasize the vital importance of participating in clinical trials conducted by impartial sponsors like the National Institutes of Health or Veterans Administration. These organizations conduct trials solely to answer important clinical questions, without profit motives. Your participation helps generate reliable evidence and benefits future patients facing similar decisions. They particularly recommend trials comparing existing treatments rather than just testing new ones, as these often provide the most practical insights for patient care.
Remember that medical knowledge evolves constantly, and staying informed about current evidence helps you make better healthcare decisions. Consider joining patient advocacy groups and following reputable medical news sources to stay updated on developments in your areas of concern.
Chapter 9
Toward a Better Medical Future: Solutions for Systemic Change
Having established that medical reversal is common, harmful, and not inevitable, the authors propose several solutions. Since reversal stems from adopting therapies without sufficient evidence, improving evidence quality is the key intervention point. This requires randomized trials showing improvements in meaningful outcomes before new practices gain approval, plus systematic testing of existing unproven therapies.
Medical education needs fundamental reform. The current structure, built on early 20th century foundations, creates inefficiency and indoctrines students with reductionist thinking over empiricism. Students learn to believe therapies work because of their mechanisms rather than evidence from trials. A reformed medical education would start before medical school, with students arriving already knowing biochemistry, physiology, and anatomy. Medical school would adopt an "encounter-based model" building from patient interactions, focusing on clinical reasoning, literature searching, and critical appraisal from day one.
Academic medical centers need restructuring to reduce conflicts of interest. Thought leaders who guide discussion in their fields often charge significant fees to advise pharmaceutical companies and device manufacturers. Universities become dependent on trial income, making it difficult to question trial designs from companies their financial health relies upon. The ultimate solution would be removing industry from trial design altogether, creating an independent body to prioritize trials based on clinical importance rather than profit potential.
The authors propose two ambitious reforms: establishing a new ethical standard for drug and device developers based on the burden of proof concept, and encouraging patients and doctors to commit to clinical trials through the "nudge principle." This approach makes trial participation the default option while allowing people to opt out, potentially answering countless medical questions quickly and cost-effectively.
In a world with higher evidentiary standards and streamlined trial recruitment, patient care would look different but better. There might be fewer treatment options, but only because ineffective or harmful ones would be eliminated. Patients who exhaust proven remedies wouldn't receive unproven guesses but would instead have access to clinical trials-the safest way to receive new treatments, with built-in safeguards.
While the authors advocate for rigorous evidence, they recognize that some medical practices must exist without randomized trial support, particularly in rare conditions, unique cases, and diagnostic evaluations. Their vision combines burden-of-proof philosophy with nudge principles to create a system where every patient receives proven treatments and untested practices undergo proper trials.
The time to begin this transformation is now.
