Bench to bedside

Why So Many Promising Discoveries Never Reach Patients

Most promising biomedical discoveries never reach a patient, and the reason is rarely that the science was wrong. A finding can be real, elegant, and reproducible and still stall in the long gap between a result in the lab and a decision in the clinic.

Most promising biomedical discoveries never reach a patient, and the reason is rarely that the science was wrong. A finding can be real, elegant, and reproducible and still stall in the long gap between a result in the lab and a decision in the clinic. That gap is a series of obstacles, each with its own logic. Who pays for the next step. What counts as proof. Who is rewarded for finishing. What regulators must require to keep people safe. Those four pressures explain the attrition better than any story about a single failed experiment.

I have watched this from several vantage points. My early publications were about mechanism, the kind of biology that looks like a lever someone could one day pull. Later I worked in drug development, then on getting evidence into the room where clinicians decide. The spaces between those roles are where good ideas tend to fall.

The discovery is the cheap part

A laboratory result is the opening move toward an answer. Showing that a receptor or a gene behaves a certain way in cells, as I did in published work on alpha2A-adrenergic receptors and on a calcium-channel gene in type 2 diabetes, tells you where a lever might be. It says nothing about whether moving that lever in a living person is safe, useful, or better than what already exists.

Each step after the discovery costs far more than the one before. Cell work is cheap. Animal work is dearer. Human testing is dramatically more expensive again. The money and time required climb steeply at the moment the chance of failure is still high, a hard combination for anyone deciding what to fund.

That mismatch has a name in translational circles: the valley of death. It is the stretch between a published finding and a candidate ready for human testing, where grants for basic science have run out and commercial investment has not arrived. Many sound ideas expire there. No funding stream was built to carry them across.

Evidence has to clear a high bar, and it should

The standard for helping patients is never "this is plausible." It is "this works, it is safe enough to justify its risks, and it beats the current option in a fair comparison." That bar sits high on purpose. Medicine's history is full of treatments that seemed obvious and proved useless or harmful.

Biological messiness makes the bar harder to clear. Many common conditions are clusters of related problems with different causes rather than single diseases. A meta-analysis I co-authored on ethnic differences in insulin sensitivity and insulin response speaks to this. A population is rarely uniform, and an average can hide the truth inside it.

Heterogeneity is the quiet killer of clean trials. If a treatment helps one subgroup and does nothing for another, a study that pools everyone can produce a flat, unconvincing average. A real effect disappears into the noise. The discovery was not false. The trial design simply could not see it.

Incentives reward the start of the journey, not the finish

The people who make a discovery and the people who carry it the last distance are usually rewarded for different things. Academic careers advance on novel findings and publications. Implementation, the slow work of fitting a proven idea into routine care, earns little of the recognition that funds a lab or a promotion.

So the most celebrated work clusters at the beginning of the pipeline, where novelty lives, and thins out toward the end, where the patient is. This reflects no one's bad faith. It is what happens when the scoreboard measures discovery and stays silent about delivery.

Drug development sits inside the same arithmetic. Bringing a therapy through years of testing and review demands sustained investment with no guarantee of return, which is why so few candidates that enter human testing ever reach approval. Having worked in global development at Novo Nordisk on GLP-1, insulin, and combination therapies, I can say the constraint is seldom a shortage of clever ideas. It is the long labor of proving one to a standard that holds.

Regulation is a safety system, not a wall

Regulators are often cast as the reason good things move slowly. The requirements exist because the cost of being wrong falls on patients. An agency asking for more evidence is usually asking the same question a careful physician would: how do we know this helps more than it hurts? My training in clinical investigation and in medical-device regulation left me with respect for that question rather than impatience with it.

The genuine friction is that the rules were largely built for pills and devices, and newer tools do not always fit the mold. Software that learns and updates raises honest questions about how you keep proving it works after approval. That is a design challenge for the evidence framework. Oversight itself is sound.

The last mile is research too

The final gap is the one people forget. A treatment that exists is not the same as a treatment a patient receives. A guideline a clinician cannot apply during a short, crowded appointment changes little.

This is the problem I turned toward in digital health: putting validated evidence in front of a clinician at the moment of decision, then testing whether that changes care. With EASY Diabetes, a clinical decision-support approach I co-developed, we evaluated it the way you would test a therapy, through a registered randomized controlled trial (NCT03258268). Delivery earns the same rigor as discovery.

What actually helps

Some of the gap is fixable. Funding designed to span the valley of death, instead of stopping where basic science ends, lets sound candidates reach the testing they need. Trials built to detect effects in subgroups, instead of drowning them in an average, recover signals that pooled designs miss.

Rewarding implementation as real scholarship would pull talent toward the underserved end of the pipeline. Evidence frameworks updated for software and adaptive tools would let regulators stay protective without forcing new technologies through rules written for older ones. None of this is exotic. It means attending to the spaces between the steps.

This article is general education, and it is not medical advice. For questions about your own care, please talk with a qualified clinician who knows your history.

References and sources

  1. Translational valley of death (Clin Transl Med)
  2. Clinical trial success rates, Phase I to approval (Biostatistics)
  3. EASY Diabetes decision-support RCT (ClinicalTrials.gov NCT03258268)
  4. Research-to-practice gap and implementation science (Worldviews Evid Based Nurs)

How this was researched. This explainer is built from the primary sources listed above and reflects Dr. Tojjar's own critical appraisal of that evidence. It explains and evaluates research and does not provide medical care.

This article is for general education and is not medical or professional advice. For guidance about your own health, talk with a qualified clinician.

Cite this article

Tojjar, D. (2024). Why So Many Promising Discoveries Never Reach Patients. Dr. Damon Tojjar. https://readingtheevidence.org/articles/why-discoveries-do-not-reach-patients/

Back to all insights