Over the past 70 years, dozens of widely practiced medical treatments have been reversed by evidence — often after harming millions of patients. This is not a story of incompetent doctors or corrupt science, but of how institutional inertia, financial incentives, biological complexity, and the honest difficulty of generating rigorous evidence conspire to entrench practices long past their expiration date. Understanding these patterns is not merely historical — it is essential to evaluating current medical practice, where mounting evidence suggests several widely defended treatments may join this list within a generation.

The phenomenon has a name: medical reversal, defined as the abandonment of a clinical practice after higher-quality evidence reveals it to be no better than a prior or lesser standard of care. A landmark 2013 analysis in the Mayo Clinic Proceedings found that of 363 medical practices tested in randomized trials published in a single major journal over a decade, 146 (40%) were reversals — established practices shown to be ineffective or harmful. Medicine corrects itself, but slowly, painfully, and often only after enormous damage.

---

1. Doctors once told you smoking was safe

From the 1930s through the mid-1950s, tobacco companies systematically recruited physicians to sell cigarettes. Camel's 1946 campaign declared "More doctors smoke Camels than any other cigarette," while Lucky Strike claimed 20,679 physicians endorsed their product. Philip Morris ran campaigns citing physician "research" showing their cigarettes cleared throat irritation. These advertisements appeared in JAMA, The New England Journal of Medicine, and The Lancet itself — JAMA accepted tobacco ads until 1953.

The practice seemed defensible because modern epidemiology barely existed. Eighty percent of middle-aged British men smoked; physicians were smokers themselves. Without prospective cohort studies, the link between cigarette use and disease could only be assessed anecdotally. The tobacco industry exploited this gap masterfully, funding contradictory research and manufacturing doubt for decades.

The evidence turned decisively with Richard Doll and A. Bradford Hill's 1950 case-control study in the BMJ, which found lung cancer patients were 14 times more likely to be heavy smokers. Their British Doctors' Study, launched in 1951 and followed for 50 years, demonstrated that smoking cuts lifespan by a decade and kills more than half of persistent smokers. The 1964 U.S. Surgeon General's Report, reviewing over 7,000 articles, declared smoking causally linked to lung cancer. Yet the tobacco industry fought for decades longer — the AMA accepted an $18 million deal to avoid stating cigarettes cause cancer, and internal industry documents revealed executives knew the evidence was "overwhelming" as early as 1962.

Tobacco has killed more than 20 million Americans prematurely since 1964, per the 2014 Surgeon General's report. Eight million deaths were prevented by control measures, underscoring the staggering toll of decades of delay. Adult smoking rates fell from 42.4% in 1965 to under 12% by 2021, but smoking remains the number one preventable cause of death in the United States, still killing 480,000 Americans annually.

2. The lobotomy won a Nobel Prize

Frontal lobotomy — severing connections in the brain's prefrontal cortex — was performed on approximately 40,000 Americans and over 100,000 people worldwide between the late 1930s and mid-1950s. Walter Freeman personally performed 3,000–4,000 procedures, sometimes 25 in a single day, using an ice-pick-like instrument hammered through the eye socket. Nearly 60% of American lobotomy patients were women. The procedure peaked at 5,000 per year in the U.S. by 1949, the same year Egas Moniz received the Nobel Prize for discovering its "therapeutic value."

In fairness, psychiatric institutions in the 1930s and 1940s housed over 450,000 patients with no effective pharmacological treatments. Asylum conditions were desperate. Moniz's early reports claimed improvement in 7 of 20 patients, and the procedure was performed at Johns Hopkins, the Mayo Clinic, and Massachusetts General. The Nobel Prize provided the ultimate seal of legitimacy.

The turning point was pharmacological, not ethical. When chlorpromazine (Thorazine) was introduced in 1954, psychiatric practice transformed almost overnight — by 1964, 50 million people worldwide had been prescribed the drug. Follow-up studies by Maurice Partridge and others documented catastrophic outcomes: permanent emotional blunting, personality destruction, incontinence, and vegetative states. Freeman was banned from surgery in 1967 after a patient died when he paused for a photograph mid-procedure. The lobotomy stands as medicine's most cited cautionary tale about the gap between therapeutic enthusiasm and rigorous evidence.

3. Hormone therapy was prescribed to prevent what it actually caused

By the 1990s, Premarin was the number-one prescribed drug in the United States, with 46 million prescriptions filled in 2000. An estimated 42% of postmenopausal women aged 50–74 used hormone replacement therapy, not just for hot flashes but for disease prevention. The American Heart Association, American College of Physicians, and ACOG all endorsed estrogen for preventing heart disease and osteoporosis.

The recommendation was grounded in compelling observational data. The Nurses' Health Study (1991) showed 30–50% reductions in coronary heart disease among estrogen users, and estrogen demonstrably improved lipid profiles. What the observational studies couldn't reveal was healthy-user bias: women who chose HRT tended to be wealthier, thinner, and more health-conscious.

The Women's Health Initiative, the largest randomized trial of its kind — 16,608 women, initiated in 1991, costing $625 million — shattered the consensus on July 17, 2002. The estrogen-plus-progestin arm was stopped early after revealing a 29% increase in heart disease, 26% increase in breast cancer, 41% increase in stroke, and a doubling of pulmonary embolism risk. HRT prescriptions collapsed 32% within nine months. The USPSTF downgraded its recommendation from "B" (recommended) to "D" (recommended against) for chronic disease prevention.

The aftermath was its own tragedy. The pendulum overcorrected: millions of symptomatic younger women were denied effective treatment. A Yale study estimated that between 2002 and 2012, as many as 91,000 postmenopausal women may have died prematurely from avoidable conditions after appropriate HRT was withheld. Current guidelines now support HRT for symptomatic women under 60 or within 10 years of menopause — a nuanced position that took nearly two decades to reach.

4. Opioids for chronic pain (Vitamin D can help)

The opioid prescribing epidemic began with a 101-word letter. In 1980, Porter and Jick published a five-sentence letter in the NEJM observing that among hospitalized patients given narcotics, addiction was rare. This letter — cited 608 times by 2017, with 72% of citations using it to claim addiction was rare and 81% failing to note its inpatient-only context — became the intellectual scaffolding for an entire industry of aggressive prescribing.

Purdue Pharma launched OxyContin in 1996, training sales representatives to tell physicians the addiction risk was "less than one percent" and funding over 20,000 pain-related educational programs. The "pain as the 5th vital sign" movement, adopted by the VA in 1999 and mandated by the Joint Commission in 2001, created systemic pressure to prescribe. OxyContin prescriptions grew from 670,000 in 1997 to 6.2 million by 2002. By 2010, the U.S. was consuming over 98% of the world's hydrocodone supply.

The reckoning was slow and devastating. Purdue pleaded guilty to federal criminal charges in 2007 ($635 million in fines) and again in 2020 ($8.3 billion settlement). The 2016 CDC Guideline for Prescribing Opioids formally recommended nonopioid therapy as preferred for chronic pain. But the damage was done: from 1999 to 2023, approximately 806,000 Americans died from opioid overdoses. The annual economic burden exceeds $1 trillion. The crisis has since shifted from prescription opioids to illicit fentanyl, but the origins are unmistakably iatrogenic — a pharmaceutical catastrophe rooted in a misrepresented letter, aggressive marketing, and institutional complicity.

5. Vioxx killed tens of thousands before Merck pulled it

Vioxx (rofecoxib), a COX-2 selective NSAID approved in 1999, was prescribed to over 20 million Americans and generated $2.5 billion in annual sales by 2003. The drug addressed a real problem: traditional NSAIDs caused an estimated 16,500 deaths per year from gastrointestinal bleeding. COX-2 selectivity promised pain relief without stomach ulcers.

The VIGOR trial (2000) confirmed fewer GI events but revealed a fivefold increase in heart attacks compared to naproxen. Merck proposed the "naproxen hypothesis" — that naproxen was cardioprotective, not that Vioxx was dangerous — and continued marketing aggressively. Internal emails later revealed Merck scientists had been "seriously discussing a potential MI risk" since 1996–1997. The NEJM issued an extraordinary "expression of concern" in 2006, stating the VIGOR authors had withheld critical cardiovascular data.

The APPROVe trial, a polyp-prevention study, was halted in September 2004 after showing doubled cardiovascular risk. Merck withdrew Vioxx on September 30, 2004. FDA scientist David Graham estimated 88,000 to 140,000 excess heart attacks during Vioxx's market life, with roughly 38,000–60,000 deaths — a toll he compared to "two to four jumbo jetliners crashing every week for five years." Merck faced 60,000 lawsuits and paid $4.85 billion in settlements. Subsequent investigation revealed Merck had ghostwritten 16 of 20 published clinical trial papers about Vioxx (Krumholz et al., BMJ, 2007).

6. Routine episiotomy increased the injuries it was supposed to prevent

At its peak in 1979, episiotomy — a surgical cut to the perineum during childbirth — was performed in approximately 61% of all vaginal deliveries in the United States, amounting to over 2 million procedures annually. The practice was championed after Dr. Joseph DeLee's 1920 advocacy for routine cutting in all first-time mothers, based on unproven claims that it prevented severe tearing, shortened labor, and protected the pelvic floor.

These claims were taught in obstetrics textbooks for decades without rigorous evidence. The 1984 West Berkshire Perineal Management Trial in the BMJ was the first major RCT to show that restricting episiotomy resulted in less perineal trauma. The Argentine Episiotomy Trial (1993, The Lancet) — a multicenter study of 2,606 women — found severe perineal trauma was actually more common in the routine episiotomy group. The definitive systematic review by Hartmann et al. in JAMA (2005) concluded: "Evidence does not support maternal benefits traditionally ascribed to routine episiotomy. In fact, outcomes with episiotomy can be considered worse."

Tens of millions of women underwent unnecessary episiotomies, experiencing increased severe lacerations, postpartum pain, sexual dysfunction, and anal sphincter injury. U.S. rates have fallen to approximately 12%, and the WHO, ACOG, and FIGO now recommend against routine use. Yet rates remain above 70% in Romania, and high across much of Latin America and Asia — a stark reminder of how slowly evidence penetrates practice globally.

7. The low-fat era may have fueled the obesity epidemic (Vitamin D can help)

In 1977, the McGovern Committee's Dietary Goals for the United States told Americans to increase carbohydrate intake to 55–60% of calories and decrease fat to below 30%. The 1980 Dietary Guidelines and the 1992 Food Pyramid institutionalized this message. The food industry responded by flooding the market with "low-fat" products that replaced fat with sugar and refined carbohydrates. Americans reduced their fat percentage but increased total calories by approximately 300 per day, driven almost entirely by added carbohydrates.

The scientific basis was Ancel Keys' Seven Countries Study (1958–1984), which showed correlations between saturated fat and heart disease, and the Framingham Heart Study's identification of cholesterol as a cardiovascular risk factor. Keys appeared on the cover of Time in 1961. A 2016 JAMA Internal Medicine exposé revealed that Harvard nutritionist Mark Hegsted — a key adviser to the McGovern Committee — had been paid by the sugar industry in the 1960s to write papers downplaying sugar's role and blaming fat (Kearns, Schmidt & Glantz).

The evidence unraveled across multiple fronts. The WHI Dietary Modification Trial (2006, JAMA) — 49,000 women over 8 years — found a low-fat diet did not significantly reduce heart disease, stroke, or cardiovascular disease. The Siri-Tarino meta-analysis (2010, American Journal of Clinical Nutrition) of 21 studies and 350,000 subjects found "no significant evidence" linking saturated fat to cardiovascular disease. The PREDIMED trial (2013, NEJM) showed a high-fat Mediterranean diet reduced cardiovascular events by 30% compared to a low-fat control. Meanwhile, U.S. obesity rates more than doubled from 13.4% in 1960 to 30.9% in 2000, and type 2 diabetes prevalence quadrupled. The 2015 Dietary Guidelines finally dropped the cholesterol cap, and the strong scientific consensus now favors replacing refined carbohydrates — not dietary fat — as the primary nutritional target.

8. Millions of children lost their tonsils for nothing (Vitamin D can help)

In 1959, the United States reached a peak of approximately 1.4 million tonsillectomies per year — the most commonly performed surgery in the country, overwhelmingly on children. By the 1960s, 61% of New York City schoolchildren had already had the procedure. The rationale was the "focal theory of infection": tonsils were believed to be portals through which infections could spread to cause rheumatic fever, endocarditis, and nephritis.

In the pre-antibiotic era, this reasoning had some basis — strep throat genuinely could lead to rheumatic fever and heart valve damage. But the indications expanded far beyond evidence. Jack Wennberg's small-area variation studies (1973) found neighboring Vermont towns where 60% versus 20% of children had been tonsillectomized, with no difference in health outcomes. The NIH Consensus Panel (1978) concluded there was insufficient evidence for routine tonsillectomy. The landmark Paradise et al. RCT (NEJM, 1984) established that only children with the most severe, recurrent infections (7+ episodes per year) showed meaningful benefit — and even then, many controls improved spontaneously.

A 2018 JAMA Otolaryngology study of 1.2 million Danish children found tonsillectomy was associated with increased long-term risks of respiratory, infectious, and allergic diseases. A UK study the same year concluded that nearly 90% of the 37,000 annual pediatric tonsillectomies were "unlikely to benefit" the patient. Today, the primary indication has shifted to obstructive sleep apnea (about 80% of pediatric tonsillectomies), where evidence is stronger. The procedure's history illustrates how a treatment can persist for decades on tradition and financial incentive long after its original rationale has evaporated.

9. PSA screening found cancers that never needed finding

The PSA blood test, identified by Richard Ablin in 1970 and approved by the FDA for screening in 1994, became nearly universal in U.S. men over 50 by the late 1990s, with over 30 million tests performed annually. Every year, more than 1 million American men underwent painful needle biopsies and upward of 100,000 had radical prostatectomies.

The logic was the same that drove mammography and other cancer screening: early detection saves lives. Prostate cancer was the second most common cancer in men, killing 30,000 Americans annually. But a critical fact was overlooked: autopsy studies consistently show prostate cancer in 18–39% of men who die of entirely unrelated causes. PSA testing was detecting a vast reservoir of indolent disease that would never cause symptoms.

The PLCO trial (NEJM, 2009) — 76,693 men followed for 13 years — found no mortality benefit from screening. The European ERSPC trial (NEJM, 2009) found a 21% mortality reduction, but to prevent one death, 1,055 men needed to be screened and 37 cancers needed to be treated. Ablin himself published a New York Times op-ed in 2010 titled "The Great Prostate Mistake," declaring his discovery had become "a profit-driven public health disaster." The USPSTF issued a "D" rating against routine screening in 2012, later softened to a "C" (shared decision-making) for men 55–69 in 2018. The estimated overdiagnosis rate is 20–50% of all screen-detected cancers. One in five men who undergo radical prostatectomy develops long-term urinary incontinence; two in three experience long-term erectile dysfunction — devastating consequences for cancers that in many cases posed no threat to life. Active surveillance has since become standard for low-grade tumors, but the prostate cancer treatment industry still generates an estimated $13 billion annually.

10. Bed rest was the worst thing you could prescribe for back pain (Vitamin D can help)

For most of the 20th century, physicians routinely prescribed one to three weeks of strict bed rest for acute low back pain — the fifth most frequent reason patients visited a doctor. The same logic applied to heart attack patients, who were confined to bed for six weeks or more through the 1960s, and to pregnant women with complications, roughly 20% of whom had restricted activity prescribed as recently as the 2010s.

The intuition was powerful: pain signals damage; rest prevents further damage. But Deyo, Diehl, and Rosenthal's 1986 RCT in the NEJM — randomizing 203 patients to 2 versus 7 days of bed rest — found no difference in outcomes, but patients prescribed 7 days missed 45% more work. Malmivaara et al. (NEJM, 1995) showed ordinary activity produced faster recovery than either bed rest or exercises. The definitive review came from Allen, Glasziou, and Del Mar (The Lancet, 1999), who analyzed 39 RCTs of bed rest for 15 different conditions involving 5,777 patients and found bed rest improved no outcomes in any condition while significantly worsening outcomes in nine.

Prolonged bed rest causes rapid muscle atrophy, bone loss (10% in the calcaneus after 17 weeks), cardiovascular deconditioning, venous thromboembolism, pressure ulcers, and depression. All major clinical practice guidelines across more than 13 countries now discourage bed rest for back pain and recommend staying active. Enhanced Recovery After Surgery (ERAS) protocols mandate early mobilization for virtually all surgical procedures. Yet the broader lesson is sobering: a treatment recommended universally for centuries, based on nothing more than intuition, was never tested in a randomized trial until the 1980s.

Five more that nearly made the list

Arthroscopic knee surgery for osteoarthritis was performed on over 650,000 patients per year at a cost of $3 billion annually until the Moseley sham-surgery trial (NEJM, 2002) showed outcomes identical to placebo. Medicare stopped covering it in 2004, but variants persist.

Antibiotic overprescribing for ear infections, sinusitis, and colds — accounting for 47 million unnecessary prescriptions annually — remains the primary driver of antimicrobial resistance, which the WHO identifies as a top-10 global health threat.

Fen-phen (fenfluramine/phentermine), prescribed to nearly 6 million Americans for weight loss by 1996, caused valvular heart disease in approximately 30% of users and was withdrawn in 1997 after a $7.65 billion settlement.

Antidepressants for mild depression were massively expanded on the back of the "chemical imbalance" narrative, which the Moncrieff et al. 2022 umbrella review in Molecular Psychiatry found unsupported by any consistent evidence; the Kirsch 2008 meta-analysis showed the drug-placebo difference was below clinical significance thresholds for all but the most severely depressed patients. Statins for primary prevention in low-risk patients remain contentious: the NNT to prevent one death over five years in low-risk populations ranges from 138 to 400 depending on the analysis, while side effects including myalgia and new-onset diabetes are nontrivial.

---

Part two: fourteen practices that may be judged next

The following treatments and recommendations are still widely defended by mainstream medicine. Each has mounting contradictory evidence, identifiable institutional resistance to change, and a plausible path to reversal. History suggests we should take these signals seriously.

1. Knee meniscus surgery remains a $4 billion (RCT found it does not help)

Approximately 700,000 arthroscopic partial meniscectomies are performed annually in the United States — the most common orthopedic surgery — at a direct cost of roughly $4 billion per year. The FIDELITY trial (Sihvonen et al., NEJM, 2013) randomized 146 patients to real versus sham surgery and found no difference at one, two, or five years. The five-year follow-up (BJSM, 2020) actually showed APM may accelerate osteoarthritis development. Seven meta-analyses confirm these findings. The BMJ issued a strong recommendation against the procedure in 2017. Yet utilization declined only 55% in the Medicare population between 2013 and 2021, and hundreds of thousands of unnecessary procedures continue — driven by hospital reimbursement that is 365% higher than surgeon fees, the "see it, fix it" mentality, and patient expectations when MRI reveals a tear that is usually an incidental, age-related finding.

2. Stenting for stable heart disease does not prevent heart attacks (Vitamin D can help)

The COURAGE trial (NEJM, 2007) randomized 2,287 patients with stable coronary disease and found stenting plus optimal medical therapy produced no reduction in death, heart attack, or hospitalization versus medication alone — a finding confirmed at 15-year follow-up. The ORBITA trial (The Lancet, 2018), the first sham-controlled PCI study, showed no improvement in angina or exercise capacity versus a placebo procedure. The ISCHEMIA trial (NEJM, 2020) — 5,179 patients across 37 countries — confirmed no benefit even in patients with severe ischemia. Despite this evidence, the Lown Institute documented 229,000 unnecessary stents placed on Medicare patients between 2019 and 2021, costing $2.44 billion. Only 3% of recorded physician-patient conversations included all elements of informed decision-making. The "plumbing model" of heart disease — the intuitive but incorrect analogy of unclogging a pipe — persists because it feels logical, because interventional cardiologists train to perform procedures, and because PCI is one of the highest-revenue procedures in cardiology.

3. Routine mammography's benefits may not justify its harms (Vitamin D can help)

The Nordic Cochrane Review (Gøtzsche & Jørgensen, 2013) analyzed eight trials involving 600,000 women and found that the three trials with adequate randomization showed no significant reduction in breast cancer mortality. The Canadian National Breast Screening Study (Miller et al., BMJ, 2014) — 25-year follow-up of 89,835 women — found a hazard ratio for breast cancer death of 0.99: functionally identical between screened and unscreened groups, with 22% of screen-detected cancers representing overdiagnosis. The Cochrane review estimated that for every 2,000 women screened over 10 years, 1 avoids dying of breast cancer, 10 healthy women are treated unnecessarily, and over 200 experience significant psychological distress from false positives. The Swiss Medical Board recommended in 2014 against new mammography programs (Biller-Andorno & Jüni, NEJM, 2014), provoking an uproar. Resistance stems from massive institutional investment in screening infrastructure, the political impossibility of appearing to reduce cancer screening, and the deeply embedded cultural narrative that equates detection with prevention. The path forward likely involves risk-stratified screening rather than age-based universal programs.

4. SSRIs for mild depression rest on a collapsing theoretical foundation (Vitamin D can help)

The Kirsch et al. meta-analysis (PLoS Medicine, 2008), using all FDA submission data, found the drug-placebo difference for SSRIs was just 1.8 points on the 53-point Hamilton scale — below the 3-point threshold for clinical significance set by NICE. Clinical significance was reached only for the most severely depressed patients. Turner et al. (NEJM, 2008) revealed that of 74 FDA-registered antidepressant studies, 94% of published trials appeared positive, versus only 51% of all trials — selective publication inflated apparent efficacy by 32%. In 2022, Moncrieff et al. published a systematic umbrella review in Molecular Psychiatry finding "no consistent evidence" supporting the serotonin hypothesis of depression, leading the Royal College of Psychiatrists to remove all references to "chemical imbalances" from their website. Yet one in six Americans takes antidepressants, with prescribing rates still climbing. NICE guidelines now recommend against SSRIs as first-line for mild depression, but access barriers to therapy, the pharmaceutical industry's multi-billion-dollar revenue, and deeply embedded cultural expectations ensure that most prescribing continues in primary care, where mild-to-moderate cases predominate. Withdrawal effects — occurring in roughly 56% of patients, with nearly half rating them as severe (Davies & Read, 2019) — create a self-perpetuating cycle that makes deprescribing difficult.

5. Long-term proton pump inhibitors were never meant to be taken forever (Vitamin D can help)

PPIs were originally approved for 4–8 weeks of use for conditions like GERD and peptic ulcers. Today, approximately 113 million PPI prescriptions are filled annually in the U.S., and studies consistently find that 40–65% of hospitalized patients and 40–55% of primary care patients taking PPIs lack documented indications. In one VA study, 35% of patients exceeded the approved duration by a median of 346 days.

Observational data links long-term use to chronic kidney disease, C. difficile infection, bone fractures, magnesium depletion, and vitamin B12 deficiency. The ARIC study (Neurology, 2023) found cumulative PPI use exceeding 4.4 years associated with higher dementia incidence. A Danish nationwide cohort of nearly 2 million individuals (Alzheimer's & Dementia, 2024) found increased dementia rates with PPI use, though the prospective ASPREE trial found no such association — a contradiction that highlights the limitations of observational evidence. The American Geriatrics Society Beers Criteria advise against PPI use beyond 8 weeks except for specific conditions. The critical barrier to change is acid rebound: discontinuation triggers severe rebound symptoms that patients interpret as proof they need the medication, creating physiological dependence that perpetuates indefinite use. Structured tapering protocols exist but are rarely implemented.

6. Acetaminophen barely outperforms placebo for osteoarthritis (Vitamin D can help)

The PACE trial (Williams et al., Lancet, 2014) — 1,652 patients across 235 primary care centers — found neither regular nor as-needed paracetamol improved recovery compared with placebo. Machado et al. (BMJ, 2015) conducted a systematic review finding acetaminophen's effect on osteoarthritis pain was just 3.7 points on a 100-point scale — far below the minimum clinically important difference of approximately 9 points. Da Costa et al. (Lancet, 2017) concluded bluntly: "We see no role for single-agent paracetamol for the treatment of patients with osteoarthritis irrespective of dose." Meanwhile, the Machado review found recommended doses were associated with a 3.8-fold increased risk of liver function abnormalities. The 2019 OARSI guidelines now conditionally recommend against acetaminophen for OA, and the ACR followed suit. Yet it remains in many national guidelines as a first-line option, largely because the alternatives — NSAIDs with cardiovascular and renal risks, opioids with addiction risks — are worse. This is a case where a treatment persists not because it works, but because the therapeutic vacuum left by its absence feels intolerable.

7. Dietary guidelines on sodium and saturated fat face serious challenges

The PURE study (Dehghan et al., Lancet, 2017) — 135,335 individuals from 18 countries — found that high carbohydrate intake (above 60% of calories) was associated with 28% increased mortality, while total fat intake was associated with 23% lower mortality. Saturated fat was associated with 14% lower mortality and 21% lower stroke risk. On sodium, O'Donnell et al. (NEJM, 2014) found a J-shaped curve: intakes between 3–6 g/day showed the lowest risk, with both higher and lower intakes associated with increased cardiovascular events — directly contradicting guidelines recommending below 2.3 g/day.

These findings are contested. Harvard's Nutrition Source has criticized PURE's methodology, noting that high-carbohydrate diets in the study often reflected poverty (white rice in Bangladesh, not whole grains). The TOHP trial 20-year follow-up using gold-standard 24-hour urine collections found a linear, not J-shaped, sodium-mortality relationship. The truth likely lies in the middle: current guidelines probably overstate the harm of moderate fat and sodium intake, but the alternative is not unlimited consumption. The deeper problem is that dietary guidelines have been shaped by industry lobbying — the sugar industry funded studies blaming fat, and food manufacturers profit from both low-fat and low-sodium product lines — making evidence-based revision extraordinarily politically difficult.

8. Vitamin D recommendations may be off by an order of magnitude

The IOM's 2011 report set the vitamin D RDA at 600 IU/day (800 IU for those over 71), targeting a serum level of 20 ng/mL. In 2014, Veugelers and Ekwaru published a statistical reanalysis in Nutrients identifying what they called a fundamental error: the IOM had confused the lower 95% prediction limit for study averages with the prediction for individual values. Using the IOM's own data correctly, they calculated that 8,895 IU/day would be needed for 97.5% of individuals to reach the target — roughly 15 times the published RDA. Heaney et al. (Nutrients, 2015) independently confirmed that approximately 7,000 IU/day from all sources was needed, still well below the IOM's own tolerable upper intake of 4,000 IU/day and the Endocrine Society's safe range.

An estimated 1 billion people worldwide are vitamin D deficient. Multiple observational studies link adequate vitamin D to reduced risks of cancer, autoimmune disease, respiratory infections, and all-cause mortality, though most remain correlational. The IOM dismissed extraskeletal evidence as "inconsistent and inconclusive," but critics argue the bar was set impossibly high. The Endocrine Society already recommends 1,500–2,000 IU/day — double to triple the IOM target. The fundamental barrier to change is institutional: acknowledging a statistical error of this magnitude would undermine confidence in the IOM process itself. Health Canada has reportedly agreed to undertake an independent reanalysis, but results remain pending.

9. Rising C-section rates have exceeded any medical justification (Vitamin D can help)

Global C-section rates have climbed from approximately 7% in 1990 to 21.1% by 2018 (Betrán et al., BMJ Global Health, 2021), projected to reach 28–33% by 2030. The WHO recommends 10–15%, and a 2015 JAMA analysis of 194 countries found maternal and neonatal benefits plateau at roughly 19%. Brazil's private hospitals perform C-sections at rates exceeding 80%. The U.S. rate has stabilized around 32%.

Keag et al. (PLoS Medicine, 2018) documented increased long-term risks from C-sections for both mother and baby, including placenta previa and accreta in future pregnancies, childhood obesity, and asthma. Blustein and Liu (BMJ, 2015) linked C-sections to altered microbiome development. The drivers are financial (hospitals and physicians earn more from C-sections), medicolegal (fear of malpractice for not performing one), and logistical (schedulable during business hours). Countries like Sweden and Finland maintain rates of 15–17% with excellent outcomes, demonstrating that lower rates are achievable without compromising safety. The path to reversal requires equalizing reimbursement between vaginal delivery and C-section, strengthening midwifery infrastructure, and implementing the WHO's 2018 Robson Classification for monitoring rates by clinical indication.

10. Aggressive chemotherapy at end of life shortens and worsens the lives(Vitamin D can help)

Approximately 30% of patients with advanced cancer continue receiving chemotherapy toward the end of life, including 2–5% who receive their final dose within 14 days of death. Rates have been increasing over time. A pivotal study by Weeks et al. (NEJM, 2012) found that 69% of patients with metastatic lung cancer and 81% with metastatic colorectal cancer did not understand that their chemotherapy was not curative.

The landmark Temel et al. trial (NEJM, 2010) randomized 151 patients with metastatic non-small-cell lung cancer to early palliative care versus standard oncology care. The palliative care group had better quality of life, less depression, less aggressive end-of-life care, and — remarkably — lived 2.7 months longer (11.6 versus 8.9 months), a benefit comparable to first-line chemotherapy itself. Prigerson et al. demonstrated that chemotherapy in patients with treatment-refractory cancer was associated with worse quality of death for those with good performance status. Multiple SEER-Medicare analyses confirm no survival benefit from chemotherapy administered within two weeks of death. The structural barriers are powerful: fee-for-service payment rewards drug administration, oncologists' prognostic estimates are systematically overoptimistic, and the culture of oncology equates stopping treatment with "giving up." ASCO has recommended early palliative care integration since 2016, but implementation remains grossly inadequate.

11. Colonoscopy's status as "gold standard" rests on surprisingly thin evidence (Vitamin D can help)

The United States is a global outlier in using colonoscopy as the primary colorectal cancer screening modality, performing approximately 15 million colonoscopies annually at $1,000–$3,000 each. Most European countries use stool-based testing (FIT) as primary screening, reserving colonoscopy for positive results.

The NordICC trial (Bretthauer et al., NEJM, 2022) — the first-ever randomized trial of screening colonoscopy, involving 84,585 participants — found an 18% reduction in CRC incidence but no statistically significant reduction in CRC mortality in the intention-to-screen analysis. Only 42% of those invited actually underwent colonoscopy, reflecting real-world uptake challenges. The COLONPREV trial (The Lancet, 2025) directly compared FIT versus colonoscopy and found FIT was non-inferior for CRC mortality at 10 years, with nearly double the participation rate (39% versus 20%). Cost-effectiveness modeling consistently shows annual FIT saves more lives and costs less than colonoscopy every 10 years at the population level. The critical insight is that the best screening test is the one that actually gets done — a less sensitive test with higher uptake outperforms a more sensitive test that half the population declines. Professional resistance stems from the enormous revenue colonoscopy generates for gastroenterologists and hospitals, and from the procedural identity of the specialty itself.

12. Annual physicals do not reduce mortality but do increase diagnoses

The Cochrane Systematic Review (Krogsbøll et al., updated 2019) analyzed 17 randomized trials involving 251,891 adults and found that general health checks did not reduce total mortality, cardiovascular mortality, or cancer mortality over a median 9-year follow-up. What they did increase was the number of new diagnoses — one trial found a 20% increase in total new diagnoses per participant over six years — and the use of medications, without corresponding health benefits.

The consequences of overdiagnosis were demonstrated dramatically in South Korea, where adding thyroid ultrasonography to a national screening program caused thyroid cancer diagnoses to increase 15-fold between 1993 and 2011 while mortality remained completely flat (Ahn et al., NEJM, 2014). Nearly all new diagnoses were small papillary cancers that would never have caused harm, yet 11% of thyroidectomy patients developed hypoparathyroidism and 2% had vocal cord paralysis. The Choosing Wisely campaign has identified hundreds of overused tests, and the European Federation of Internal Medicine has called routine health checks "non-evidence-based yet firmly entrenched." The shift needed is from age-based routine testing to evidence-based selective preventive services — but revenue generation, patient expectations, and defensive medicine create formidable resistance.

13. Water fluoridation faces new evidence on neurodevelopment (Vitamin D can help)

Approximately 200 million Americans (64% of the population) receive fluoridated water — more than the rest of the world combined. The CDC has called water fluoridation one of the ten great public health achievements of the 20th century, and evidence supporting its role in reducing dental caries is substantial.

However, the August 2024 NTP Monograph systematically reviewed 74 epidemiological studies and found that 18 of 19 high-quality studies showed an inverse association between fluoride exposure and children's IQ, concluding with "moderate confidence" that higher fluoride exposures (above 1.5 mg/L) are consistently associated with lower cognitive scores. Bashash et al. (Environmental Health Perspectives, 2017), in a prospective Mexican cohort, found that each 0.5 mg/L increase in maternal urinary fluoride was associated with a 2.5-point IQ decrease in children. Green et al. (JAMA Pediatrics, 2019) found similar associations in a Canadian cohort at concentrations matching the U.S. recommended level of 0.7 mg/L. A federal judge ruled in 2024 that fluoride at current U.S. levels poses an "unreasonable risk" to children's neurodevelopment.

The critical nuance: fluoride's mechanism is primarily topical, not systemic — it strengthens enamel on contact, not through ingestion. Europe, where 98% of the population drinks non-fluoridated water, has experienced caries declines comparable to the U.S., relying instead on fluoride toothpaste, varnishes, and targeted treatments. The NTP report acknowledges insufficient data to draw firm conclusions specifically at 0.7 mg/L, and the ADA has actively contested the report's methodology. But the direction of evidence — combined with the availability of topical alternatives — suggests the risk-benefit calculus of systemic water fluoridation may shift meaningfully in coming decades.

14. Formula marketing still undermines breastfeeding worldwide

The WHO recommends exclusive breastfeeding for six months, yet the global exclusive breastfeeding rate is only 48%, and in the United States, just 25.8% of infants are exclusively breastfed at six months. The global infant formula market is worth an estimated $55 billion annually. The 1981 WHO International Code of Marketing of Breast-milk Substitutes was adopted 118–1, with the United States casting the sole negative vote. A 2022 UNICEF/WHO report found formula marketing "remains unacceptably pervasive, misleading and aggressive," with over 50% of surveyed parents having received direct marketing from formula companies.

The evidence favoring breastfeeding is extensive: WHO estimates improved breastfeeding could prevent over 820,000 child deaths annually through reduced respiratory infections, diarrheal disease, and other causes. Maternal benefits include reduced breast and ovarian cancer risk. The Nestlé controversy — beginning with the 1974 exposé "The Baby Killer" and continuing through documented Code violations in 46 countries — demonstrated how free formula samples in maternity wards create dependency by suppressing lactation, and how marketing in low-resource settings without clean water leads to infant illness and death.

This is less a case of "a treatment that doesn't work" than of a massive market distortion: formula is necessary and lifesaving when breastfeeding is medically impossible, but aggressive marketing systematically undermines breastfeeding even where it is feasible and preferable. The structural barriers — $55 billion in industry revenue, inadequate paid maternity leave, hospital partnerships with formula companies, and the deliberate framing of criticism as attacking maternal autonomy — make this among the most politically entrenched items on this list.

---

The patterns that connect past and future ($$, etc.)

Several structural forces appear consistently across both lists, and recognizing them is more valuable than memorizing any individual case.

Financial incentives are the most reliable predictor of resistance to evidence. Every reversed practice on List 1 and every challenged practice on List 2 involves someone making money from the status quo — whether Purdue Pharma selling OxyContin, orthopedic surgeons performing meniscectomies, or formula companies marketing to new mothers. This is not conspiracy; it is the predictable behavior of rational economic actors within poorly designed incentive structures. Fee-for-service payment, in particular, rewards doing more rather than doing right.

Relative risk reduction obscures absolute futility. Statins "reduce heart attacks by 30%" sounds transformative; "treating 250 people for five years to prevent one death" does not. Mammography "reduces breast cancer mortality by 20%" sounds essential; "screening 2,000 women for 10 years to save one life while overtreating 10" invites a different conversation. Until absolute risk communication becomes standard in medical practice, patients will systematically overestimate the benefits and underestimate the harms of interventions.

The placebo effect of procedures is enormous and systematically underestimated. Sham surgery trials — for knee arthroscopy, meniscectomy, vertebroplasty, and cardiac stenting — consistently show that the ritual of surgery itself produces substantial and lasting improvement. This means any uncontrolled surgical case series will show "benefit," making it extraordinarily difficult to convince surgeons that their procedures don't work.

Publication bias corrupts the evidence base. Turner's finding that selective publication inflated antidepressant efficacy by 32%, Merck's ghostwriting of 16 of 20 Vioxx papers, and the CTT Collaboration's refusal to share individual statin data with independent researchers all illustrate the same problem: when the entities funding research also control its publication, systematic distortion is inevitable, not exceptional.

Medical reversal takes 15–30 years from first contradictory evidence to widespread practice change. The Argentine Episiotomy Trial was published in 1993; ACOG didn't revise guidelines until 2006. The COURAGE trial was published in 2007; unnecessary stenting continues in 2026. This lag — driven by training, identity, institutional investment, and patient expectations — means that millions of patients receive outdated care long after evidence has shifted. The most important question is not "what will be reversed?" but "what already has enough evidence for reversal but hasn't changed yet?" Many items on List 2 may already belong on List 1.

The answer is not to distrust medicine but to demand that it live up to its own stated values: evidence over authority, outcomes over intuition, and the courage to stop doing things that don't work — even when they're profitable, familiar, and emotionally comforting.

---

Related in VitaminDWiki

• Medical errors are 3rd most likely cause of death
• Vitamin D fighting COVID-19 meets all Bradford Hill Criteria
  100X more cost effective than vaccination