More Than We Measured

Inside the system, and connecting the dots

By Rujjwal, physician and full-time wonderer. September 2026, about 9 minutes.

When I was growing up in Nepal, a country still under developed by modern standards, books and the internet were my window into a much bigger world. My math, physics, and biology books told me about galaxies and black holes, quantum physics, gene editing, modern cancer treatments and biologics. But none of it felt tangible in the society around me. I saw it on TV and read about it in magazines and books from the West, and I came to a quiet conclusion: humans had basically figured it all out. If we didn't have something in Nepal, I assumed it was a matter of geography and lack of resources, not of knowledge.

Then I grew up, and the deeper I learned, the more questions appeared.

I moved to the West and saw what the pinnacle of human development looks like up close. It's remarkable, but it isn't the flawless place I had pictured from afar. There are inefficiencies in health systems and governments,abuse and corruption is not absent, and the exploitation of nature, people, and environment, and we are still divided by the color of our skin and by our religious/political beliefs. None of this is unique to the West. Wherever we live, we humans paint our world with both our goodness and our flaws. What surprised me was simply that the perfection I had imagined from far away didn't exist up close. Even the people I look up to are humans like me, with their own flaws, emotions and insecurities. And beyond society, the biggest questions about the nature of reality and consciousness remain wide open. We've turned physics into cell phones, computers, and AI, yet we still don't fully understand much of the mechanics underneath.

So no, we haven't figured it out. Recorded human civilization is only about 5,000 years old, and most of our understanding has exploded in just the last hundred. But every answer seems to lead to a deeper, harder question.

Nowhere do I feel this more than in medicine, where I work every day.

How we design a drug

Most modern drugs begin in a lab. Scientists identify a target, usually a receptor or an enzyme tied to a disease, design a molecule to act on it, and then study the downstream effects in the area they care about. It's a powerful approach, and it works.

Take NSAIDs like ibuprofen, which almost all of us have used. We know they block cyclooxygenase (COX) enzymes, reducing the production of prostaglandins, the chemical messengers that drive inflammation and make nerves more sensitive to pain. That's exactly what we aimed for, and it delivers. Side effects like nausea and stomach pain were studied and reported. But over time, we learned NSAIDs can also cause stomach bleeding, harm the kidneys, and raise the risk of heart attacks and strokes. One of them, Vioxx, was pulled from the market in 2004, after millions of people had already taken it. Even aspirin was used for about seventy years before anyone could explain how it worked.

Or take proton pump inhibitors (PPIs), among the most commonly used medications in the world. They shut down the proton pump in the stomach's acid-producing cells and reduce acid secretion, which is what we wanted. But proton pumps aren't only in the stomach. Bone-remodeling cells called osteoclasts rely on them too, and researchers now suspect PPIs may interfere there, one proposed reason long-term users show higher fracture risk. PPIs came to market in the late 1980s. The first cases of dangerously low magnesium linked to them weren't described until 2006.

So here's my question: are we sure a drug does only what we designed it to do? Drugs are screened against known targets before approval, but the body runs on countless enzymes, receptors, and signaling systems at the same time. No panel can cover them all, and trials are built to measure what we're looking for, over months rather than decades. We can't find what we never think to measure. Some effects only surface years later, after millions of people have taken the drug, and some may not look related at all.

The drug that did more than it was aimed at

The clearest recent example is a drug millions of people are taking right now. GLP-1 drugs, the medicines behind names like Ozempic, Wegovy, Mounjaro, and Zepbound, were built around a gut hormone that helps control blood sugar, and they began as diabetes treatments. Then they proved powerful for weight loss, and their use exploded. By one survey, about 1 in 8 US adults have taken one at some point, and more than 15 million currently have a prescription.

Then came the surprises. Patients began describing a quieting of "food noise," the constant background chatter about eating. Some mentioned they'd lost interest in drinking. Researchers took notice. A study of more than 600,000 US veterans found that people taking these drugs were less likely to develop substance use disorders involving alcohol, nicotine, cannabis, cocaine, opioids, and other drugs, and those already struggling with addiction had fewer overdoses, hospitalizations, and drug-related deaths. In absolute terms, that's about 7 fewer people per 1,000 developing a substance use disorder over three years.

That study was observational, which means it shows an association, not proof. But randomized trials are now following. In May 2026, The Lancet published a trial comparing weekly semaglutide with placebo in people with alcohol use disorder and obesity, all of whom were also offered standard therapy, and the drug-plus-therapy approach reduced heavy drinking.

Why would a blood sugar drug touch addiction? It turns out GLP-1 receptors aren't only in the gut and pancreas. They're also in the brain regions tied to reward, the same circuits that alcohol and drugs act on. A hormone we studied for blood sugar had been speaking to the brain's reward system all along. We just weren't listening for it.

A hormone we studied for blood sugar had been speaking to the brain's reward system all along.

The other edge

The same systems logic cuts both ways. In 2024, researchers reported a possible link between semaglutide and NAION, a sudden loss of blood flow to the optic nerve that can cause permanent vision loss in one eye. The evidence has grown more nuanced since. European regulators concluded in 2025 that it should be listed as a very rare side effect, affecting up to about 1 in 10,000 patients. A 2026 study of nearly 588,000 US veterans with diabetes found about 39 cases per 10,000 GLP-1 users over three years, compared with 29 per 10,000 among people taking a different class of diabetes drug.

There's a twist that shows how tangled the body is. The risk factors for NAION include diabetes, high blood pressure, high cholesterol, and sleep apnea, which are some of the very reasons people start these drugs in the first place. Separating what the drug does from what the patient already carried is genuinely hard.

To be clear, the absolute risk is small, and for most people the benefits are large. Nobody should stop a medication because of a blog post, including this one. The point isn't that GLP-1 drugs are dangerous. It's that a drug aimed at one pathway reached the brain's reward system and possibly the blood vessels of the eye, and we learned about both mostly after millions of people were already taking it.

Why we find these things late

This isn't because anyone is careless. It's built into how we test medicines. A clinical trial is designed around one main question, such as whether a drug lowers blood sugar, and it's sized to answer that question well. It runs for months or a few years, in carefully selected patients. Rare effects, slow effects, and effects in organs nobody was watching may simply not show up in a few thousand people over a short window.

They show up later, when millions of real people with messy, overlapping conditions take the drug for years. That's when observational studies and safety reporting systems start catching patterns. But observational data has its own trap: it can show that two things travel together without proving that one causes the other. That's why the addiction findings, exciting as they are, still need randomized trials, and why the eye findings are still debated.

The honest summary is that we approve drugs based on what we measured, and we learn the rest over time.

Connecting the dots

One drug, four ripples:

A lesson in humility

We aimed at one lock and opened several doors, some we wanted and some we didn't. Maybe that's the real lesson: we keep trying to optimize systems we don't fully understand.

We've built prototype quantum computers that need extreme cold and enormous engineering just to run. Meanwhile, an ordinary leaf, using nothing but sunlight, water, and carbon dioxide, captures light and passes its energy along with almost no loss, and some scientists believe quantum effects may help it do so. We still don't understand how the universe works, or the brain, or consciousness. And yet, with fragments of knowledge, we've done remarkable things. That's what makes us human. We're curious and courageous, and sometimes overconfident, even foolish.

So I want to keep reminding myself of our limits. The unknown is far larger than the known. In nature, disorder is the easy path, and order takes constant energy and effort to hold together. And yet here we are, the product of an extraordinary orchestration of the cosmos, able to witness it and wonder about it.

My younger self, who understood the world through the books I was reading, wasn't wrong about how much we had achieved in a remarkably short time. But I was oblivious to the grandeur of the darkness: the sheer scale of what we don't know, and the things we may never be able to know. That's a deeper topic, and one I'll explore another time.

What else in medicine might be doing more than we measured?

These thoughts are my own and are not medical advice for you. If you take any of the medicines mentioned here, talk with your own clinician before changing anything. If you notice sudden vision changes while on a GLP-1 medicine, seek care promptly.

More about who's writing this is on the About page. If this got you thinking, or you'd push back on any of it, I'd genuinely love to hear it: sharma.rujjwal [at] gmail [dot] com.