“Jeopardy!” host Alex Trebek’s announcement that he has stage 4 pancreatic cancer has once again turned the spotlight on this cancer, much as Steve Jobs did in 2004. Make no mistake about it: Pancreatic cancer is a bad disease. More than 40,000 Americans die from it each year — more than from breast, prostate, or ovarian cancer. Only lung and colorectal cancer cause more deaths.

In a perspective article in the New York Times, Dr. Diane M. Simeone, who directs the Pancreatic Cancer Center at New York University Langone Health, suggested that early detection is a solution. But just because a disease is bad doesn’t mean early detection will make things better. In fact, by casting a net that pulls in more people, early detection can make a bad disease worse.

That’s what happened in randomized trials of ovarian cancer screening. (Screening means checking seemingly healthy people for signs of hidden disease.) Screening led more women to have surgery to remove their ovaries, and more women to experience complications from that surgery, but it did not change the risk of dying from ovarian cancer.


Early detection efforts always have the downside of turning more people into patients. Whenever doctors look for early forms of disease, they regularly find more people with it than they would have otherwise. In some, the detected disease is destined to intensify and cause symptoms, or even death. Earlier treatment may help them — or it may not. Others are overdiagnosed, told they have the disease yet it is not destined to ever bother them. Many more have findings that falsely suggest disease and require a battery of tests to attest to their health.

Screening must involve many to potentially help a few.


Anatomy is relevant when talking about treatment for pancreatic cancer. While the ovaries are relatively accessible, the pancreas is not. It sits deep in the abdomen — just getting there is a challenge. Removing a cancerous part of the organ is more challenging still. Surgeons must work around blood vessels that are tightly connected to the pancreas. Injuries to them can be difficult to repair and can lead to catastrophic bleeding.

The pancreas contains ducts that deliver to the intestines digestive fluids from both the liver and the pancreas. So this operation has the additional challenge of preserving and re-plumbing these ducts, which are essential for digestive function. Leaks are common and digestive fluids can wreak havoc on surrounding tissues.

There are few operations more dangerous than pancreatic surgery. Within a month of it, somewhere between 4 percent and 11 percent of patients die (the variability reflects both surgeon and hospital expertise). The risk of dying climbs even higher three months after surgery. Complications are the norm — roughly one-third of people undergoing pancreatic surgery experience a major surgery-related complication.

It is definitely not a good thing to subject more people to pancreatic surgery. But that’s what will happen if screening for this type of cancer takes off.

To be fair, relatively few extra individuals will undergo surgery. But less-consequential harms will accrue to many more, who will be subjected to more scans, more tests, and more biopsies. Some will have complications from the extra investigation. All will be made to worry — as will their families.

People will also be made poorer in the process. A few will be bankrupted by it.

It’s easy to think that early detection is the solution to every disease. But that’s not true. What is true is that early detection has become a great way to drum up more business for medical centers, physicians, test manufacturers, and the pharmaceutical industry.

It’s tempting to want to do something, particularly when a celebrity develops a bad disease. But don’t call for early detection, call for better treatment. And call for more acceptance: recognizing that good people — doing all the right things — can nevertheless develop bad diseases.

H. Gilbert Welch, M.D., is a general internist in Thetford, Vermont, and was a professor of medicine at Dartmouth for 28 years. He is the author of “Less Medicine, More Health: 7 Assumptions That Drive Too Much Medical Care” (Beacon Press, 2015).

  • Screening saves lives, if we look at other cancers it works. Pancreatic Cancer has been the “forgotten” cancer in terms of funding for years and still lacks in terms of dollars invested in all research when compared to many major cancers. What we have been doing for the past 50 or so years has not worked so maybe it is time to look at things differently and take a different approach. I think your comments are one sided, I respect a fair debate on this topic but I think your statements against early detection are incorrect and very misleading. Early detection is a integral part to better treatments if we are going to make progress with this cancer.

  • “don’t call for early detection, call for better treatment” If people are detected before disease is metastasized and before symptomatic, some pancreatic therapies may well be more effective then when at stage 4. There haven’t been the early markers for detecting pancreatic cancer in studies (that I could find) to determine how well current therapies would work for stage 1. Early Biomarkers may well aid in effective therapy that doesn’t require a gastly Whipple procedure. A recent paper by Mike Goggins et al at JHs (Oncology, April 3, 2019) found that genetic testing revealed the finding of germ line mutations was a better indicator (e.g. went on to pancreatic cancer) than just surveillance of people with strong family history and no evidence of mutations because they were NOT genetically tested. Numerous genetic/cellular changes occur between stage I and IV. Let us use the valuable cyst fluid from pancreatic cysts and pancreatic juices to determine early stages and the critical biomarkers that may well then provide availability of screening with scientific sound scientific basis vs “what is in the eye of the beholder with MRI, CT etc.”

  • Welch has a more clear-eyed view of the evidence. Consider Simeone’s claim in the NY Times that “Screening of those at high risk is of demonstrated value.” In the paper she cites, it’s hard to see how screening helped people deemed high risk because of a family history of pancreatic cancer:

    • Screening detected “suspicious” lesions in more than half (112 out of 214 screened because of family history).

    • 13 with suspicious lesions underwent *major* surgery (distal pancreatectomy, total pancreatectomy, or Whipple procedure).

    • 5 of the 13 were overtreated, the authors said, because pathology revealed only low-grade abnormalities. And 4 of 13 had postop complications (fistulas, bleeding).

    • Screening detected putative pancreatic cancer in 3 of 214 people, who then underwent surgery, but only 2 of 3 turned out to be cancer.

    The authors concluded that the value of screening in this group “is still not clear.” For a different high-risk group, CDKN2A mutation carriers, they said screening was “relatively successful” (the 5-year survival rate was 24% among those with screening-detected cancers).

  • Peoples of the World:
    Listen and follow Drs. Welch’s and Huang’s advise. They know what they are talking about. I know this for a fact. Life is worth living happily. Dying is an equal opportunity event for all of us and should not be mismanaged for the sake of a few days of unjustified suffering…

  • Three of my first degree maternal relatives died if pancreatic cancer. By the time they were diagnosed, no treatment could save them. They were all dead within 3-4 months of diagnosis.
    Most of what is stated is true but no where near 100%.
    Thats the reason screening Chest CTs are recommended for smokers or those with significant smoking history. Because catching a cancer early, saves lives.
    Same with colonoscopies. Removing polyps before they are cancerous.

    MRI, CT ,ERCP etc are used before pancreatic surgery so its not as if surgery is the initial approach.

    Some cancers are very aggressive and grow/spread rapidly, some do not, such as prostate cancer (although the debate is still a tive on screening). Ask the women who’s lives were saved by mammogram screening. Should they have died because of the many false scares? Not when it’s you!

    We need much more research on pancreatic cancer and perhaps as more celebrities succumb to this cancer, it will finally happen and we will be able to catch this monster early.

  • This is quite true. Also, being told that one has some disease condition can also activate nocebo effects (via negative expectations). For example, the following recent randomized study examined how the knowledge of a genetic risk can be physiologically more damaging than having that risk itself:

    Turnwald, B. P., Goyer, J. P., Boles, D. Z., et al. (2019). Learning one’s genetic risk changes physiology independent of actual genetic risk. Nature Human Behaviour, 3(1), 48.

  • Thank you for your comments Sui Huang. I think susceptibility is more likely than randomness. Immune system supression of varying degrees is likely contributive. Just like Epstein-Barr’s causative role in Burkitts lymphoma is enhanced by the victims fight against chronic malaria; leaving the immune system exhausted.
    The long term use of the HPV vaccine will be enlightening going forward. One could chart the rise of cancers with the 60’s free-love and the concomitant spread of sexually transmitted viruses over that time period. For example, breast cancer is likely caused by a virus, yet nobody is actively looking. What one eats and drinks has little to do with getting cancer. The idea is to keep healthy, choosing good quality food, keeping stresses in check as much as possible to prepare for the fight ahead.

  • Thank you Prof. Welch for your relentless warning against the perils of screening and early detection. Already in medical school we are conditioned on the reflex of “early detection is the solution to every disease” indeed, leaving no room for reflection.

    But Dr. Welch: the downside of early detection and treatment is worse than just the complications of the treatment, or merely the associated psychological burden that you so eloquently articulate in your books.

    For, the very treatment itself, be it chemo, radiation or -as we and others now show, surgery, impart so much tissue stress and triggers so much uncontrolled regenerative activities that they actually promote progression – and cancer recurrence. Treatment is a double-edged sword.

    Since we treat every cancer, and sadly a majority of invasive cancer recur, we do not have control experiments in humans to study this phenomenon. But animal studies in all kinds of forms for all kinds of tumor now are accumulating showing treatment-induced generation of cancer stem cells and treatment induced recurrence – confirming old suspicions based on epidemiological data.

    STATnews has reported on this:
    and – even for surgery, here:

    That therapy is not only potentially harmful because of collateral damage, but also can backfire (producing more cancers, mets) much as bombing countries harboring terrorists can produce more terrorists is explained here for a wider audience:
    or in more detail, here:

  • Bravo to Gil Welch, who continues to be a leader in the field and one of wisest commentators on screening. Anyone who is considering geting screened should heed his warning. Pancreatic cancer is rare, despite the seemingly high death rate. Until we have better treatments, we hsve to accept that some people will be hit with a devastating cancer, and screening is not the answer.

  • Yes Pancreatic cancer is a terrible disease BUT it must have a CAUSE so what is it? And yes we can make more drugs to TREAT but not to cure. So it seems prudent to investigate the role of some oncovirus as causative. It’s where medicine needs to go but Big Pharma is in charge and treatments are a multi-billion dollar industry.

    • Dear Carol: The horrible thing about cancer is that many may not have a cause, or one beyond our control. Even oncoviruses are not the actual controllable cause: they may be every where, but only strike randomly. Cells play lottery zillions of time, and some particular “wrong” number they draw will, through a chain reaction, in the most unfortunate case, lead to cancer.
      Like freak accidents, of course, there is little control and very limited ways we can prevent cancer. Not flying will guarantee that you don’t die in a plan crash. Some prevention are more meaningful than others – in cancer the only one with substantial effect I would say are: avoid smoking, obesity and perhaps red meat.

      Very little research money goes into prevention. If eating two tomotoes a day would significantly reduce cancer we would never know (just one fictive example). For, no one makes billions of $ from selling tomatoes (as opposed to drugs).

Comments are closed.

A roundup of STAT’s top stories of the day in science and medicine

Privacy Policy