Doctors know how to die. The system won’t let the rest of us.

I stumbled on this piece from downbad_ on HN about how doctors die. It’s one of those articles that resurfaces every few years because it hits a nerve: physicians, with their insider knowledge, often refuse the aggressive treatments they give their patients. They choose DNRs. They skip the ICU. They die at home, not in a hospital bed surrounded by beeping machines.

The essay is beautifully written and emotionally devastating. It also avoids a far messier problem.

Doctors can pull this off because they know the system. They know which treatments are statistical long shots. They have colleagues who will level with them and families who trust their judgement. The rest of us? We’re at the mercy of a healthcare system engineered to keep the machine running, not to translate our final wishes into reality.

PainSignal data makes this painfully concrete. We track real-world problems reported by healthcare workers, and right now there’s a persistent, high-severity issue: patient code status is inconsistently documented across systems. In plain English, what your advance directive says in one hospital may not show up in the EHR at another. When you show up in the ED unresponsive, the default is to code you. That’s the “futile care” the article talks about—not because anyone is malicious, but because the digital plumbing doesn’t connect.

This isn’t a niche edge case. The problem has a severity score of 5 out of 5. That’s the highest PainSignal rating, meaning it’s actively harming patients and driving clinicians up the wall. It also has an opportunity score of 62/100, which is decent—meaning there’s a clear business case for solving it. In fact, we track an app idea called CodeSync that aims to centralize real-time code status documentation, and it’s one of 396 healthcare app ideas currently mapped on the platform. The demand is there.

But the chaos goes deeper. If you talk to nurses—the people actually running the codes—they’ll tell you the system is breaking them. PainSignal shows problems like unsafe patient-to-nurse ratios, medication errors from lack of real-time verification, and charting overload that steals time from actual patient care. These aren’t abstract complaints; they’re specific, recurring, and getting worse. The severity trends on nurse staffing issues are rising. That means more burned-out nurses, more mistakes, and more patients getting aggressive treatment simply because no one has the bandwidth to pause and confirm what the patient actually wanted.

This is where the builder angle gets interesting. The HN article frames the problem as an ethical or cultural one: doctors know too much, families are in denial, we need better conversations about death. All true. But there’s also a massive operational layer that’s completely unaddressed. If you’re a vibe coder or an indie hacker looking for a problem to solve, you could build something that directly reduces futile care. Think about it—CodeSync is essentially a CRUD app with HIPAA compliance wrapped around it. It’s not rocket science. It’s just a deeply underserved gap because the incumbents (Epic, Cerner) move like molasses.

I’m not saying an app will fix America’s death denialism. But consider the scale: PainSignal tracks 624 problems in healthcare alone. Many of them are simple data synchronization failures that lead to catastrophic outcomes. When a patient’s code status is buried in a PDF scanned into the chart, you might as well not have one. That’s a software problem, not a philosophical one.

And there’s a subtler point the article misses: doctors often die well because they opt out of the hospital entirely. They go home, on hospice, with a morphine drip and family around. That’s a luxury that depends on having a home, a support system, and a healthcare proxy who can advocate for you. For huge swaths of the population—the poor, the isolated, the ones without a PCP who knows their name—that path doesn’t exist. They will die in the ICU because that’s where the system puts them by default. Building tools that surface a patient’s actual wishes, even when they can’t speak, is a massive accessibility unlock.

One more thing: the article implies that doctors’ training hammers “save lives” so hard that they have to unlearn it personally. Our data suggests the opposite for nurses. They’re often the ones pushing for less aggressive care because they see what it looks like up close—the broken ribs from CPR, the family drama, the moral distress. PainSignal captures problems like “Nurses face moral distress from providing futile care” with high severity. If you’re building for clinical decision support, the nurses are your power users. They want the tools to document and elevate patient preferences, but they’re buried in charting. Give them a one-click solution to flag a code status discrepancy during shift change, and you’ll save lives—or rather, deaths.

So here’s the takeaway: the HN article is right that doctors die differently, but wrong to treat it as a quirk of medical culture. It’s an indictment of a system that hasn’t built the infrastructure to honor anyone’s wishes. The good news? That infrastructure is hackable. The problems are well-defined, the severity is off the charts, and the incumbents are asleep. If you’ve been looking for a healthcare problem worth solving, this is it. Don’t build another wellness app. Build the thing that makes sure my DNR actually shows up when I’m dying.