The Strange Economy of Medicine: Why We Know Less Where We Spend More

What if the problem is not too little medicine, but too little knowledge?

Most people assume healthcare fails because society does not spend enough on it, or because the wrong people cannot afford access. But there is a more unsettling possibility: we are often paying for medicine faster than we are learning what medicine actually does. That means the real crisis is not only cost or access. It is epistemology, the economics of knowing.

This is a difficult idea to absorb because medicine feels like the most evidence-driven domain in modern life. We imagine a system in which treatments are tested, validated, and then deployed at scale. Yet many of the things we routinely trust, prescribe, or recommend are far less settled than the public thinks. At the same time, some of the most promising interventions are weakly studied precisely because they are cheap, ordinary, or unpatentable.

That is the core paradox: the healthcare system is highly optimized to sell certainty and poorly optimized to produce it.

---

The illusion of obvious medical value

When people hear that medical spending often shows only a tiny average effect on health outcomes, the intuitive reaction is disbelief. Surely there must be a mistake. Surely the miracle of modern medicine cannot be so small. But the deeper shock is not that medicine sometimes fails. It is that even the things we think we know well are often harder to pin down than our confidence suggests.

Consider antibiotics. They are a classic symbol of medical triumph, the kind of thing people point to as proof that medicine obviously saves lives. And yet when you zoom out from dramatic anecdotes to population-level effects, the picture becomes less clean. Some interventions do have clear benefits in specific contexts, but the aggregate story is often far less spectacular than the myth.

This does not mean antibiotics are useless. It means something subtler and more important: medical value is highly conditional. A treatment can be transformative in one setting, marginal in another, and harmful in a third. The public story tends to flatten this variability into a simple narrative of progress. Real medicine is messier. It works less like a magic wand and more like a finely tuned instrument that only matters when used at the right time, on the right patient, for the right problem.

That creates a dangerous cognitive habit. If the visible examples of medicine are the dramatic success stories, we may infer that medicine is broadly powerful everywhere. But if the actual distribution of benefit is uneven, then much of what healthcare does may be low-yield, redundant, or even unnecessary. The symbol of medicine is not the same as the average effect of medicine.

The public sees medicine through its peaks. Policy must judge medicine by its average.

---

The hidden problem: we do not learn enough from what cannot be patented

If the first paradox is that medicine may be overestimated in general, the second is that the most economically useful knowledge is often the least likely to be rigorously developed. Why? Because the modern research system is deeply shaped by incentives, and patents are one of its strongest engines. If a drug can be patented, there is a business case for funding trials, marketing results, and building an evidence base.

But what about things that cannot be patented? Vitamin D. Acupuncture. Diet patterns. Exercise protocols. Cheap generic compounds. Old drugs used in new ways. These are exactly the kinds of interventions that could matter enormously because they are scalable, accessible, and low-cost. Yet they are often stuck in a valley of neglect. Nobody has a strong private incentive to spend tens of millions proving whether they work, because the reward cannot be monopolized.

This creates a bizarre asymmetry. A marginally improved patented drug may receive lavish testing because a company can capture the upside, while a cheap intervention with broad public health potential may remain poorly studied because the benefits would be shared by everyone. In other words, the market rewards exclusive knowledge, not necessarily useful knowledge.

Think of it like transportation research. If every breakthrough in train scheduling could be patented and sold privately, we would likely know a lot more about train scheduling. But if the benefit of a better schedule simply spread across the whole city, private firms would underinvest. Medicine has the same problem, except the stakes are bodies, suffering, and death.

The result is not merely an uneven research pipeline. It is a distorted map of reality. We tend to have the most elaborate evidence for what is easiest to monetize, not necessarily for what matters most in public health. That means some of the biggest opportunities may be hiding in plain sight, not because they are scientifically uninteresting, but because they are commercially inconvenient.

---

A market for answers is not the same as a market for truth

These two paradoxes are connected by a deeper issue: healthcare is not just a delivery system for treatments, it is a production system for beliefs. Every prescription, trial, guideline, and reimbursement rule signals what the system thinks it knows. But if the incentives reward treatment volume and patentable novelty more than knowledge generation, then the system can become fluent in action while remaining weak in understanding.

This is where the tension gets serious. In many domains, we imagine that more spending buys more quality. In medicine, much spending buys more activity. But activity is not the same as insight. A hospital can generate a vast amount of interventions without learning whether those interventions improve outcomes enough to justify their use.

This helps explain a frustrating feature of healthcare debates: both sides often talk past each other. One side sees medicine as underutilized life-saving science, the other as bloated and ineffective bureaucracy. The truth may be that both are partially right because they are observing different layers of the same system. There are undeniably situations where medicine is indispensable. But there is also enormous routine care whose value is uncertain, marginal, or poorly measured. The problem is not simply that medicine exists. It is that medicine is applied in a world where our confidence often outruns our evidence.

Here is a useful mental model: imagine healthcare as a library where the most popular books are not necessarily the most reliable, and the rarest books are not necessarily the least valuable. The reading room is crowded. The catalog is incomplete. And the cataloging budget is tied to whether a publisher can profit from the book. Under those conditions, volume tells you very little about truth.

In healthcare, the loudest evidence is not always the best evidence. Sometimes it is just the best funded.

---

The real frontier is not more treatment, but better epistemic infrastructure

If the diagnosis is right, the solution is not merely to spend more on healthcare or to slash spending indiscriminately. The real challenge is to build epistemic infrastructure, systems that learn faster, cheaper, and more honestly than the current arrangement permits.

What does that mean in practice? It means designing institutions that answer three questions better:

1. Does this intervention work?
2. For whom does it work?
3. At what cost relative to simpler alternatives?

Those questions sound obvious, but healthcare often answers them indirectly, slowly, or not at all. A treatment may be adopted because it is mechanistically plausible, culturally prestigious, or commercially profitable, not because we have a robust sense of its real-world effect size. That is especially dangerous for interventions that are cheap and diffuse, because nobody has much incentive to run the boring studies that would settle the matter.

The solution is not only randomized trials, though they matter. It is a broader shift toward learning-oriented health systems. That includes pragmatic trials embedded in real-world care, better comparative effectiveness research, public funding for non-patentable interventions, and reimbursement rules that reward evidence creation, not just intervention volume.

There is also a cultural shift required. Patients and clinicians need to become more comfortable with the phrase “we do not know yet”. In most industries, uncertainty is a temporary inconvenience. In medicine, it is often the starting point. A mature healthcare culture does not pretend certainty where there is none. It builds better ways to reduce uncertainty.

This matters because the biggest gains may not come from dramatic new treatments. They may come from discovering that some widely used practices are unnecessary, while some neglected low-cost practices are surprisingly powerful. In that sense, the future of medicine may be less glamorous than we imagine and far more valuable.

---

The actionable lesson: follow the incentives behind the evidence

The most useful question a patient, policymaker, or intelligent layperson can ask is not simply, “Does this treatment exist?” It is, “Who had the incentive to prove this works?” That question cuts through a lot of noise.

If a treatment is heavily studied, ask whether it was studied because it is important or because it is profitable. If an intervention seems suspiciously under-researched despite widespread use, ask whether that is because it is ineffective or because no one stood to gain from proving it effective. This does not mean every unpatentable therapy is secretly valuable. It means the absence of evidence often reflects the structure of incentives, not the actual structure of nature.

This lens also helps explain why some medical debates become so polarized. People often assume that disagreement reflects differing interpretations of the same evidence. But sometimes the evidence base itself is structurally biased. One side may be overconfident in well-funded but narrow studies. The other may overreact by embracing weak claims simply because the mainstream seems dismissive. A better response is disciplined skepticism in both directions.

The right posture is not cynicism. It is institutional humility. We should be highly confident that medicine can save lives in the right circumstances, and highly cautious about assuming that the current research ecosystem knows which circumstances those are as well as it should.

Key Takeaways

• Do not confuse activity with value. A treatment being common, expensive, or technologically advanced does not mean it has a large average effect.
• Ask who benefits from proving something works. Patentable interventions attract research money; nonpatentable ones often do not, even when they may matter greatly.
• Treat medical certainty as conditional. Many interventions are effective only for specific patients, contexts, and timing.
• Support learning systems, not just delivery systems. The healthcare system should be rewarded for generating knowledge, not merely for delivering more care.
• Use skepticism evenly. Be wary both of hype around expensive interventions and of dismissing low-cost ones that have been neglected by the research market.

---

The deeper reframe: healthcare is a knowledge economy disguised as a service economy

The most important shift is conceptual. We usually talk about healthcare as if it were primarily a service industry, something like plumbing with better branding and higher stakes. But that framing misses the central feature of the field. Healthcare is actually a knowledge economy: a system for discovering, validating, transmitting, and applying truths about biological reality.

When that knowledge economy is distorted, the consequences are profound. We may overvalue interventions because they are visible, subsidized, or profitable. We may undervalue interventions because they are cheap, generic, or hard to own. And we may keep mistaking institutional motion for scientific progress.

That is why the question is not simply how much we spend on healthcare. It is what kind of knowledge our spending is able to produce. A system that cannot efficiently learn about cheap, common, nonpatentable, or context-dependent interventions is a system that will forever be partially blind. It will still produce miracles. It will still save lives. But it will also waste enormous resources chasing certainty where the incentives are strongest and ignoring it where the incentives are weakest.

The real scandal is not that medicine is imperfect. Every human system is. The scandal is that we have built a system that often learns least where learning would help most. Once you see that, the debate changes. It is no longer about whether medicine is good or bad in some abstract sense. It is about whether our institutions are designed to discover truth, or merely to monetize the appearance of it.

And that, ultimately, is the question that should haunt every discussion of modern healthcare: how many of our medical beliefs are actually discoveries, and how many are just the byproducts of a badly structured market for answers?