QUESTION: “What Kills Cancer Patients: The Disease or the Treatment?”

Why ask this now? Cancer is the second leading cause of death in the United States. That fact should force a second question: when people with cancer die, how much is the disease—and how much is the treatment or the damage the treatment caused? The question matters even more because a widely cited analysis argues medical error may rank as the third leading cause of death in America—a signal that our systems under-detect and under-report care-caused harms.

From prevention to treatment-first: two worlds that don’t speak the same language

After spending 35 years in the world of occupational safety and health—where every injury is recorded, logged, and investigated—I entered the world of oncology as a patient. What I found was staggering. In OSHA, prevention is everything. Injuries are classified, tracked, and turned into actionable safety improvements. In cancer care, treatment comes first—and harm, if it’s recorded at all, is often buried in trials, misattributed, or ignored altogether.

I speak from personal experience. I’m a two-time head and neck cancer survivor. I received cisplatin and radiation, and I now live with long-term cognitive impairment that was brushed aside as “chemo fog.” More than five years post-treatment, I learned—on my own—that I’m at heightened risk for stroke due to carotid artery narrowing caused by radiation. That risk was never once explained to me. This kind of omission isn’t rare—it’s systemic.

So here’s the hard question: Do people die from the cancer… or from the treatment and its unspoken downstream harms? The honest answer is both—and the tragedy is, we don’t measure the difference well enough to inform patients properly.

What We Know (and What We Choose Not to Track)

Even during treatment, the risks are sobering. In modern clinical trials, chemoradiation for head and neck cancers carries a 1–3% treatment-related death rate. In real-world settings—outside of trials, where patients are often older or have more complex health conditions—the rate can be even higher. That’s not an outlier if you’re the one affected.

For those receiving palliative chemotherapy, early mortality within 30 days of treatment is tracked as a safety signal. Because sometimes, it’s not the cancer that kills first—it’s the treatment: infections, organ failure, or immune complications triggered by aggressive drugs.

But it’s the years after treatment that tell the more silent story. Radiation to the neck significantly increases the long-term risk of carotid artery stenosis and stroke. Several peer-reviewed studies now recommend ultrasound surveillance starting 2–5 years post-treatment and continuing regularly. Yet many clinics never mention this. There’s no standard survivorship protocol, no routine screenings, and often no acknowledgement of this late—but potentially lethal—risk.

Even in clinical trials, causes of death are often misclassified. A death caused by treatment toxicity or a delayed injury may still be recorded as a “cancer death.” That blurs attribution, misleading both doctors and patients.

The Invisible Harm: Errors, Misdiagnosis, and Missing Data

Diagnostic error is another hidden danger. A 2024 study estimated that nearly 800,000 Americans each year are seriously harmed or die from major diagnostic mistakes. These aren’t rare glitches—they are systemic failures. Yet, these deaths don’t appear in national mortality data because our death certificates aren’t built to capture “medical error” as a cause.

In 2016, a controversial but influential BMJ study argued that medical error would rank as the third leading cause of death in the United States if recorded accurately. Even its critics agreed on the core issue: our reporting system conceals more than it reveals.

What OSHA Would Do (But Oncology Doesn’t)

In workplace safety, an employer must log every serious injury—who it happened to, what caused it, how severe it was, and what was done about it. These logs feed into national databases that drive prevention.

Oncology, by contrast, has no such system.

There is no national “harm log” to capture toxicities, complications, or long-term effects of treatment outside clinical trials. The FDA’s adverse-event reporting database is voluntary and plagued by underreporting. Cancer registries like SEER are excellent at tracking diagnosis and survival—but are not designed to capture treatment harms in any structured or useful way.

Even death reporting is skewed. There is no way to code “treatment-related death” on a standard death certificate. So treatment harm remains invisible, unmeasured, and unaddressed.

Case in Point: Head & Neck Radiation and Stroke Risk

As a survivor, I only learned long after my treatment that head and neck radiation substantially increases stroke risk by damaging carotid arteries. This isn’t theoretical—it’s documented across studies and institutional protocols.

Some centers now recommend routine carotid artery screening beginning a few years after radiation. But the vast majority of survivorship care plans omit it entirely. In the OSHA world, failing to disclose and prevent a foreseeable harm like this would be a serious violation. In cancer care, it’s just standard practice.

How We Can—and Must—Do Better

We need a safety overhaul in cancer care. And we can borrow the blueprint from OSHA.

1. Create a national “Treatment Harm Log.” Just like OSHA forms, this would document each patient’s therapy, complications, and outcomes using standardized fields. Not just in trials—for everyone. The log should track acute toxicities, late effects, diagnostic errors, and outcomes—not just survival.

2. Add a survivorship “Critical Risks” checklist. This should include disease- and treatment-specific risks (like stroke after neck radiation or heart failure after anthracyclines) and require review and signature at key points: diagnosis, end of treatment, year 1, and year 5.

3. Fix death reporting. Death certificates should include a structured field to indicate if treatment contributed to the death—whether through toxicity, procedural harm, or diagnostic error.

4. Mandate adverse-event reporting. Serious oncology complications should be logged and disclosed. Non-reporting should carry consequences, and results should be published in a public dashboard to drive transparency and accountability.

5. Tie institutional and practitioner accountability to harm outcomes. In construction, energy, and manufacturing, companies maintain OSHA 300 logs—detailed injury records that feed into an Experience Modification Rating (EMR). This score affects insurance premiums and determines eligibility for high-risk contracts. Harm has consequences—financial, reputational, operational.

Cancer centers are no different. They are service providers. Yet, unlike these other high-risk industries, they face no equivalent system tied to treatment injuries. That needs to change.

We should implement a Patient Harm Index (PHI)—a counterpart to EMR—that reflects rates and severities of treatment-related harm. Treatment centers should be rated, and individual practitioners too. These ratings should be accessible to patients, insurers, and regulators alike.

6. Revoke blanket legal immunity for healthcare-caused harm. A major barrier to accountability is the legal protection many healthcare institutions enjoy. In many states, patients harmed by treatment—unless it qualifies as narrow-scope “malpractice”—have little recourse. That must change. Real reform requires legal and financial consequences for preventable harm.

The Bottom Line

This isn’t about assigning blame. It’s about acknowledging reality: treatment, while often life-saving, can also cause serious, lasting harm—and sometimes, death. We can’t protect patients from every risk. But we can start by tracking those risks honestly and informing patients fully.

Cancer didn’t kill me. The treatment didn’t either. But parts of me were lost along the way—pieces no one warned me I’d be risking. We owe future patients a better standard: one that tracks harm with the same rigor we track hope