[Lancet Classic] The 85% Leak: How Avoidable Waste Cripples Biomedical Progress
Avoidable Waste in the Production and Reporting of Research Evidence
This seminal Viewpoint identifies systemic "avoidable waste" in biomedical research, estimating that over 85% of investment is lost through four cumulative stages: poor question selection, inadequate study design, non-publication, and unusable reporting. It proposes a framework of institutional reforms to ensure research translates into actual clinical benefit.
TL;DR
Biomedical research is a massive global investment, yet this paper reveals a staggering reality: over 85% of research funding is wasted. This isn't due to "failed experiments" (which are a natural part of science) but due to systemic failures in how we choose questions, design studies, and report results. The authors argue for a fundamental shift toward transparency, systematic evidence synthesis, and patient-centered prioritization.
Background: The "Butterfly" Problem
Medical research is often critiqued for its "butterfly behavior"—researchers flitting from one exciting new drug or hypothesis to the next without ever fully exploiting or even publishing the results of their previous work. This leaves clinicians and patients in a state of forced ignorance. As one patient with myeloma asked: "Why was I forced to make my decision knowing that information was somewhere but not available?"
The Four Stages of Cumulative Waste
The core contribution of Chalmers and Glasziou is the identification of a "leaky pipeline" where waste accumulates at four distinct stages.
1. Choosing the Wrong Questions
Research agendas are often disconnected from clinical reality. In osteoarthritis, for example, 80% of trials focus on drugs, while patients overwhelmingly prefer research on surgery, physiotherapy, and coping strategies.
- The Insight: Research is currently driven by "what is fundable" or "what is interesting to researchers," rather than "what reduces the disease burden."
2. Unnecessary or Poorly Designed Studies
Many researchers launch new trials without checking if the question has already been answered. A shocking study found that only 46% of trial authors were even aware of relevant systematic reviews when designing their new studies.
- The Flaw: Designing studies in a vacuum leads to redundancy and methodological errors (like inadequate allocation concealment).
3. Failure to Publish (The File Drawer Problem)
Positive results are published; "disappointing" results are buried. Only about half of the research presented at conferences ever makes it to a full, peer-reviewed paper.
- The Consequence: This creates a biased evidence base that overestimates treatment benefits and hides harms.
4. Biased or Unusable Reports
Even when published, reports often lack the "recipe." If a trial shows a "low-fat diet" or "stroke unit" works but doesn't describe the intervention in detail, clinicians cannot replicate it.
Figure 1: The cumulative leak in research evidence production.
Methodology: Quantifying the Loss
The authors utilized data from various systematic reviews of the "epidemiology of research." By compounding the estimated 50% loss at each stage (design, publication, reporting), they arrive at the mathematical conclusion that the final usable output is less than 15% of the initial investment.
| Stage | Estimated Efficiency | Cumulative Yield |
|---|---|---|
| Initial Investment | 100% | 100% |
| Poor Design/Method | ~50% Loss | 50% |
| Non-Publication | ~50% Loss | 25% |
| Unusable Reporting | ~50% Loss | 12.5% |
Critical Analysis: Why This Matters Today
This paper acted as a catalyst for the REWARD (REduce Waste And Reward Diligence) movement. Its brilliance lies in shifting the "blame" from individual "bad scientists" to a systemic incentive structure that rewards novel primary research over high-quality synthesis and reporting.
Key Recommendations for the Industry:
- Funders: Should require systematic reviews of existing evidence before approving grants for new trials.
- Regulators: Must mandate the registration of all protocols at inception to prevent "outcome switching."
- Journals: Should follow The Lancet's lead in requiring every new study to be "set in context" (i.e., how does this trial change the existing meta-analysis?).
Conclusion: A Call for Evidence-Based Research
The authors conclude that while some waste is inevitable in the "search for truth," the current levels are a correctable scandal. By treating the production of research as a rigorous process—subject to the same evidence-based standards as the medical treatments themselves—we can double or triple the value of health research without spending an extra dollar.
Future Outlook
Since this paper, we have seen the rise of Open Science, ClinicalTrials.gov, and the CONSORT reporting guidelines. However, the "butterfly behavior" persists in the age of AI and high-speed publishing. The challenge remains: how do we incentivize researchers to value completion and clarity as much as discovery?
Takeaway: High-quality research is not just about 'significant' p-values; it’s about accessible, usable, and contextualized evidence.