The Crisis in Evidence-Based Medicine: Limitations of RCTs and the Rise of Personalized N-of-1 Trials (2026)

Within modern evidence-based medicine, case series and case reports are frequently dismissed as “anecdotal” and therefore regarded as scientifically weak. In oncology in particular, the prevailing paradigm holds that only large randomized controlled trials (RCTs) can establish the validity of a treatment. While such trials are essential for determining the efficacy and safety of a single intervention under controlled conditions, the history of medicine consistently demonstrates that discovery rarely begins with randomization—it begins with observation. Clinicians notice unexpected responses in individual patients; outcomes emerge that defy conventional expectations; and, over time, patterns begin to form.

It is also important to recognize the structural limitations of the current research ecosystem. Large RCTs are often impractical for off-label or repurposed strategies, particularly those involving inexpensive agents, dietary interventions, or lifestyle modifications. These approaches typically lack the commercial incentives—such as patent protection and sufficient return on investment—required to justify the substantial financial commitments of multinational pharmaceutical sponsors. As a result, potentially valuable therapeutic signals may remain under-explored or systematically overlooked.

Evidence-based medicine (EBM) aspires to ground clinical decision-making in rigorous scientific data. However, its implementation has been increasingly challenged by commercial influences, methodological constraints, and an incomplete accommodation of individual variability. Although RCTs remain the established gold standard, they are often costly, time-intensive, and not optimally suited to personalized care. In this context, there is growing interest in complementary approaches, including N-of-1 trials and real-world evidence, which may offer more flexible, patient-centered pathways for evaluating treatment effectiveness in heterogeneous populations.

The Corruption of Evidence-Based MedicineThe core idea of EBM—formally studying treatments to avoid flawed human perception—is sound. It has delivered successes, such as clarifying when angioplasty benefits acute heart attacks but not chronic heart disease (e.g., COURAGE and ORBITA trials). (1)

Yet leaders in the field have grown disillusioned. Richard Horton, editor-in-chief of The Lancet, stated in 2015 that "much of the scientific literature, perhaps half, may simply be untrue." Marcia Angell, former editor-in-chief of The New England Journal of Medicine (NEJM), wrote in 2009: "It is simply no longer possible to believe much of the clinical research that is published." Arnold Relman, another former NEJM editor, described the medical profession as "being bought by the pharmaceutical industry" in teaching, research, and practice.
Commercial interests drive this corruption:

• Industry-funded trials — are 70% more likely to show positive results than government-funded ones.
• Selective publication — suppresses negative findings. For antidepressants, 36 of 37 favorable studies were published, but only 3 of 36 unfavorable ones.
• Rigged outcomes — Pre-2000, unregistered endpoints allowed cherry-picking results (57% positive rate); post-registration, it fell to 8%.
• Advertorials and reprints — Journals earn massive revenue from drug companies buying reprints of favorable studies (e.g., 41% of The Lancet's income).
• Direct bribery — Over 50% of influential journal editors receive industry payments, averaging $27,564 personally plus research funds.
• Publication bias — 28% of started trials go unpublished, with pharma-sponsored ones 5 times more likely to be shelved.

The result: a biased evidence base that exaggerates drug benefits. As former journal editors now admit, EBM has become "lucre-based publishing." Until commercial conflicts are removed—particularly from universities and physicians—EBM remains unreliable.Limitations of Randomized Controlled TrialsRCTs excel at establishing average effects in controlled settings but falter in real-world application, especially for lifestyle, nutrition, or repurposed interventions. (2)

• They are expensive, slow, and logistically complex → often requiring millions of dollars and years to complete.
• They prioritize strict protocols over generalizability → excluding diverse patients and overlooking individual responses.
• In metabolic health — where fewer than 12% of Americans are metabolically healthy — RCTs often fail to capture genetic, lifestyle, or environmental variations.
• Many top drugs benefit fewer than 1 in 4 patients → leaving most without effective solutions.

A 2024 mega meta-analysis of 102 COVID-19 treatments found no significant difference in effect sizes between RCTs and observational studies (RR 1.00 [0.92-1.08]). A prior Cochrane review of 1,583 meta-analyses reached the same conclusion: observational studies match RCTs on average. This challenges the blanket dismissal of non-randomized evidence.Better Alternatives to Traditional RCTsMore pragmatic approaches can generate robust evidence faster and cheaper:

1. Pragmatic Clinical Trials — Real-world settings with broad inclusion for greater applicability.
2. N-of-1 Trials — Randomized periods of treatments in a single patient, ideal for personalization.
3. Synthetic Control Arms & Real-World Evidence (RWE) — Historical data or registries replace placebos, addressing ethical and cost issues.
4. Case Series/Retrospective Reviews — Quick hypothesis generation from existing data.
5. Bayesian Adaptive Trials — Flexible designs that update as data arrives.
6. Digital/Remote Trials — Wearables and apps for scalable monitoring.
7. In Silico Trials — AI simulations to predict outcomes and optimize designs.

A hybrid pathway—case series → RWE cohorts → pragmatic/adaptive trials—builds evidence progressively without demanding massive upfront RCTs.The Rise of N-of-1 Science: A Case StudyHarvard medical student Nick Norwitz (PhD from Oxford) exemplified N-of-1 experimentation by eating 720 eggs (24/day) over 30 days. On a low-carb baseline, his LDL cholesterol dropped slightly; adding 60g carbs daily led to an 18% reduction overall. (3)
The experiment highlighted that dietary cholesterol impacts blood levels minimally for most people—but responses vary (e.g., in "lean-mass hyper-responders"). More broadly, it showcased N-of-1's power: testing interventions on oneself with tracked biomarkers to discover personal truths.
Technology fuels this trend:

• Wearables (Oura, Fitbit)
• Continuous glucose monitors
• Home lab tests (InsideTracker)

These enable "citizen scientists" to establish baselines, test hypotheses, and refine strategies.Stanford's Michael Snyder calls N-of-1 "the future," emphasizing longitudinal personal data for specific recommendations. Aggregating many N-of-1 trials could create vast real-world datasets, accelerating discoveries beyond traditional research.
Challenges remain:

• Potential bias
• Placebo effects
• Misinterpretation on social media

N-of-1 complements, not replaces, population studies. Funding disparities—billions for drugs vs. minimal for lifestyle/nutrition—hinder progress, pushing researchers toward public engagement for support.Toward a Healthier FutureEvidence-based medicine's promise has been undermined by corruption and RCTs' inherent limitations. Cleaning the evidence base requires eliminating financial conflicts, especially in academia.
Meanwhile, embracing pragmatic alternatives—particularly N-of-1 trials empowered by technology—offers a path to truly personalized medicine. Your life is already an N-of-1 experiment. With rigorous tracking and critical thinking, anyone can become a citizen scientist, discovering what optimizes their health in a one-size-fits-none world.

References:

1. The Corruption of Evidence Based Medicine. https://fung.substack.com/p/the-corruption-of-evidence-based
2. Randomised controlled trials (RCTs) are often costly, slow, and logistically challenging
3. N of 1 Trial: Harvard Med Student Eats 720 Eggs in 30 Days, Highlighting a Trend in N=1 Science
4. Randomised controlled trials (RCTs) are often costly, slow, and logistically challenging - ChatGPT