Prostate Cancer for Marketers

How 23andMe can earn the trust of customers

If you saw an ad on TV from a drug company that said “Ask your doctor about prostate cancer,” would you run to your doctor to get tested? Probably not, because you know what their motives are. This summer, 23andMe is running video ads on YouTube that feature a 50-ish man explaining that his 23andMe data showed that he is genetically at risk for prostate cancer.

What should he do about that?

In this essay, I argue that given what we know about prostate cancer and testing, the right answer is “absolutely nothing,” and the company should retire that ad immediately. I’ll also show that the company is starting to provide actual benefit to patients, if it can build their trust.

Through the DNA Looking Glass

Before we meet him, this man ordered a kit from 23andMe consisting of a vial and a return pouch. He spat into the vial and sent it back to 23andMe for DNA sequencing, which now costs only $99. Weeks later, he logs in and sees some tiny fraction of his genome displayed online with a bunch of interpretive info, mostly along the lines of how he fits into various genetic clusters and what diseases he might be at risk for.

23andMe tells him he is genetically at risk for prostate cancer. That sounds serious. In the United States, prostate cancer is the second largest cause of cancer-related death in men. 80% of men over 80 years old have prostate cancer. No one wants to die of something that could be prevented, so the next logical step is to go see the doctor.

The PSA Test

His urologist prescribes a simple blood test for Prostate-Specific Antigen, which should tell if he has prostate cancer.

That sounds reasonable. Each year in the US, about 240,000 men get diagnosed with prostate cancer as a result of a positive PSA test. Not wanting to have cancer — who does? — it’s time to take the next step and get a biopsy.

Weeks later, the biopsy shows cancer. Oh dear. How can you stop now? Almost no one does. Ninety percent of these patients go on to get treatment.

When The Treatment is Worse than the Disease

Patients trust their doctors to help them get rid of their cancers. In the case of prostate cancer, the urologist presents the patient with a range of options, and together they choose the one that’s most expensive.

Do they work? Unfortunately, treatments extend life in fewer than 3% of cases. Dr Otis Brawley, Chief Medical Officer at the American Cancer Society, in his excellent talk, Failure is the System, says:

One study that suggests [treatment] might save lives said that we had to treat 48 men in order to save one life. Radical prostatectomy in the United States has a one percent risk of death. You can kind of start doing the math. There’s not a lot of people saved.

Not only that, but “saving” people in this context can often mean the difference between dying at age 88 vs 87 (but living with side effects for decades). Prostate cancer grows slowly. Most men die with prostate cancer, not of it. The US Preventive Services Task Force recommends against any PSA screening for any patient:

Prostate cancer is a serious health problem that affects thousands of men and their families. But before getting a PSA test, all men deserve to know what the science tells us about PSA screening: there is a very small potential benefit and significant potential harms.

A very recent study again reinforces the benefits of not getting screened or treated. And according to the Clinical Guidelines Committee of the American College of Physicians:

The benefits of screening with the prostate-specific antigen (PSA) test are outweighed by the harms for most men.

Fear sells, and scaring people into getting PSA tests is the wide end of a $50 billion global prostate-treatment funnel that leaves many patients worse off than before. Those who have side effects often go on to have procedures to correct them — another profit center for hospitals. 23andMe is doing its fair share to shuttle people into the wide end of that funnel, supporting an industry that thrives on doing more harm than good.

Developing a Better Test

Making evidence-based decisions is difficult, because if you’re the one person who benefits, you may not care about others who don’t. And urologists do treat a very few, very specific forms of prostate cancer that are hard to diagnose, fast-growing, and lethal. Those treatments can give people 10-30 years more of life. But at the moment, our tests aren’t very accurate, and that’s where we need to improve.

DNA testing shows us what we inherited from our parents when we were born. We can compare our genes to those of others who have diseases and try to compute our statistical likelihood of getting a particular disease. But — it’s complicated. While genes provide blueprints for building various kinds of proteins, we’re still in the early days of understanding what genes actually do and how they do it. Very few diseases are a direct result of a defect in a specific gene. Genes act together in complicated ways, get turned on and off by their environment, and have stubbornly resisted being responsible for specific diseases.

While your genome can tell that you might get cancer some day, you’re probably more interested in the cancer you have or don’t have right now. Today, we diagnose cancer using four main methods:

Feeling lumps, using fingers, as we do with breasts and prostates.

Imaging shows tumors but not the beginnings of tumors or what makes them or where they will spread next.

Antibodies show the body’s immune response to cancer. Works better for some cancers than others and provides limited, indirect information.

Biopsy lets us look at potential cancer cells and test them using a variety of visual and biological tests. It’s the best tool we have, but it’s less accurate than we would like.

Most cancers have many different sub-types and evolve along with our ability to treat them. Because our tools are so indirect, it’s rare that a single test or score tells you much of anything.

Proteomics is different. A proteomic screen will detect cancerous cells individually and help us develop the “big molecule” drugs to selectively kill them. Getting this to work in practice will take much of the next twenty years. If this interests you, please watch Danny Hillis talking about the Proteome as a diagnostic tool. What they are doing at Applied Proteomics is far more important to understanding our health than what 23andMe is doing.

The Three “Benefits” of 23andMe

If 23andMe isn’t going to be very helpful in diagnosing cancer, could it be helpful in making any health decisions? According to their web site, 23and me has three health uses. Let’s look at each of these claims.

It can be used to screen for potential birth defects. The way we currently do this is to look at family histories. That’s helpful, but it doesn’t spot carriers — people who may be carrying a genetic defect but not expressing it. In this case, a genetic test does give you more information. However, genes can only give you probabilities; they don’t actually tell you anything about the baby. What tells you about the baby is amniocentesis — doing a genetic screen on the baby’s amniotic fluid. This is far more important than testing the parents. And it’s done routinely already, looking for Down Syndrome and other defects. We don’t need to spit in a vial and wait six weeks to learn that we should test the fetus’s amniotic fluid. We should just do it.

It can be used to know the risk of various diseases you might get. For example, it can tell you whether you might possibly get Parkinson’s, age-related macular degeneration, or Type 2 diabetes. The problem is, anyone might get these diseases, the same as many men might get prostate cancer. You can try to avoid them, but you’re not going to start treating any of these diseases before you get them. It’s not clear that knowing your genetic factors will lead you to do something different that may extend your life. Genetic data pales in comparison to sequencing your proteome and your microbiome, which will tell us what’s happening now.

It can be used to help determine drug response and dosing. As I wrote about in my last book, Pull, there are genes that tell us which groups of people respond to certain classes of drugs. These are large important classes, and knowing the Cytochrome P450 (CYP) genetic signals tell us much about drug metabolism in patients. At least 57 genes contribute to the P450 complex. Today, 23andMe can tell how your body will respond to Warfarin (a blood thinner known as Coumadin), and that’s huge progress. Drug companies and doctors want to adjust your dose depending on your CYP profile. 23andMe is clearly heading in the right direction here.


The race is on to find better ways to detect true life-threatening cancers and do something about them, but for many cancers, we’re not there yet. No evidence-based doctor would recommend getting your genes sequenced so you can see if you’re likely to get prostate cancer, much less getting a PSA test. As the evidence comes in, we must force ourselves to be rigorous and not fall prey to economic incentives to test and treat. As Otis Brawley says, the system isn’t failing — “Failure is the system.”

Even in their advertising, 23andMe should make evidence-based decisions and try to help customers. Someday, 23andMe and its competitors will be part of the solution for health care — I hope they will do a better job of earning the trust of consumers than the drug companies have.

Come learn more about evidence-based decisionmaking at

David Siegel



Provocateur, professional heretic, slayer of myths, speaker of truthiness to powerfulness, and defender of the Oxford comma.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
David Siegel

Provocateur, professional heretic, slayer of myths, speaker of truthiness to powerfulness, and defender of the Oxford comma.