You want to do what to me?

When I go to the doctor for my yearly physical (OK, not quite yearly, but…) he puts me through the ringer.  He asks me dozens of questions, follows up with more questions, and does a (thorough!) physical exam.  Then he takes some blood, and sometimes runs some other tests.  The big question is “why”?  What are he and I trying to accomplish?

There isn’t a great set of data out there about “yearly physicals” as such.  But broken down into its components, we can learn something important.  The point of a yearly physical is prevention of disease and monitoring of chronic diseases.  For these, we have data.

Monitoring of chronic diseases includes counseling diabetics on diet and medication, and evaluating them for complications of diabetes.  Regular f0ot exams, urine protein tests, blood pressure measurements, HbA1c levels, and other parameters prevent the complications of diabetes.  Maximizing proper treatment of diseases such as diabetes requires regular tweaking of medications, diet, etc., and regular appointments provide a good setting for this.

The other goal of a doctor visit is prevention of disease.  Prevention of disease is not as simple as it may sound.  It involves identifying target diseases, possible strategies for primary and secondary (and “tertiary”) prevention, and for screening (early detection), developing metrics for success, and testing these.

To make preventing a disease worthwhile, it should probably be relatively common and/or have a significant health cost associated with it.   It should also be a disease for which early detection or prevention provides some sort of benefit.   An uncommon disease worth screening for might be phenylketonuria (PKU).  This is a rare disease, but one with devastating consequences, including severe mental impairment.  If detected early (shortly after birth) a child can grow up to be completely normal with the use dietary changes and sometimes medication.

But the bulk of disease prevention focuses on common diseases, such as diabetes, heart disease, and cancer.  Primary prevention refers to preventing a disease from occurring, for example, treating hypertension to prevent a heart attack.  Secondary prevention is aimed at preventing recurrent disease, for example, preventing a second heart attack in someone with established heart disease.  Screening is the early detection of a disease, with the aim of treating it at an early, more treatable stage.

And this is where things get murky.  I can screen for hypertension by checking a blood pressure.  It’s cheap and easy to do, and treating high blood pressure is inexpensive and has a huge payoff (prevention of strokes, heart attacks, kidney failure, etc.).  But not all screening is so straightforward.

At my yearly-ish physical, my doctor screens me for prostate cancer.  He does this by measuring a blood test called PSA, and by digital rectal exam (“digital” in the sense of “finger”, not bytes).  Cancer is a leading cause of death in men, and prostate cancer is the second leading cause of cancer death in men.  I don’t want prostate cancer, and if I get it, I want to catch it nice and early, right?


Several different organizations put out screening recommendations for prostate cancer.  These recommendations all agree on one thing: it’s a lot more complicated then just running a test.  For example, the U.S. Preventative Services Task Force, while finding that screening does detect some cases of disease, was not able to say that screening held any advantage over simply discovering a prostate cancer later.  Other organizations have slightly more aggressive recommendations, but the sum of evidence is murky.  We can detect prostate cancers early, but since prostate cancers tend to behave unpredictably, its hard to tell whether we do any good by detecting it early.  We may actually do harm, by subjecting men to invasive testing and treatment that may not leave them any healthier were they not tested in the first place.

Men should be counselled about the uncertainty of the outcomes of prostate cancer screening, and can be offered it at various times depending on their risk of having the disease.  This is an imperfect method of screening, but it’s what we’ve got.   My read of the literature tells me that we’re going to get better at this over the next few years.

Some cancers aren’t worth screening for at all.  Lung cancers are common as cancers go, but we have not yet found a way to screen for it.  Chest X-rays are not, it turns out, a great way to find early lung cancers.  CT scans can, but these tests are more expensive and involve higher doses of radiation.  One way to reduce the risks of screening and to improve the tests ability to detect real cancers and avoid false positives is to screen those at highest risk, for example smokers.  This is being investigated, but the results have been disappointing so far.

We do much better with screening for cervical cancer, breast cancer, and colon cancer, but the data are not iron-clad.

Screening is a complicated area, but one with great potential rewards.  But there are lots of ways to get it wrong.   Remember that we screen for diseases so that we can prevent suffering and premature death.  If a patient already has an advanced cancer, one likely to kill them, there’s little point in screening for other cancers that may not even affect them for many years.

And yet, we do.

A recent study in the Journal of the American Medical Association took a lot at a large group of patients with advanced cancer and found that, while they were not being screened at the same rate as healthier patients, they were still being screened at pretty high rates.

This is several different kinds of horrible, but there are some subtleties.  If, for example, you have stage IV lung cancer, this is probably what you will die of.  And it won’t take very long.  Finding early breast, prostate, colon, or other cancers involves going through unnecessary procedures and provides no benefit.  You will  not  get the chance to die of an early-detected colon cancer.

But many patients do not like to hear that.  After being diagnosed with an advanced cancer, you might not enjoy reminders of your mortality, and being told by your doctor that you don’t need a test because you’re going to die of something else can be pretty shocking. But one of our jobs as physicians is to provide realistic guidance and to help guard our patients from unnecessary harm.   If you can’t give bad news in a useful way, you shouldn’t be a doctor.


Sima, C., Panageas, K., & Schrag, D. (2010). Cancer Screening Among Patients With Advanced Cancer JAMA: The Journal of the American Medical Association, 304 (14), 1584-1591 DOI: 10.1001/jama.2010.1449


  1. How many advanced-stage cancer patients ask for screening for other cancers? There’s a difference between not offering a screening test you think is irrelevant and telling the patient, “Normally I would suggest breast cancer screening for a woman of your age, but I’m not gonna, because you’re likely to die before it’s relevant.”

    This is a great post. I’d be interested in your take on using advanced testing procedures but getting worse outcomes. I’m thinking specifically of using MRIs to look at people with lower back pain. I seem to remember reading something about dramatically increased use of invasive procedures but no improvement (or worse – increased harm), but I don’t remember the details. Do I have it right, and are there other examples of this?

  2. Lots.

  3. Jonah Lehrer wrote about the MRIs-in-back-pain story in his book The Decisive Moment. IIRC, before MRI, no-one knew what caused most cases of back pain, and the prescription in such cases was rest, which worked pretty well.

    Then MRIs came along and people started finding loads of “abnormalities” and recommending interventions to fix them. But in most cases it turned out that the abnormalities weren’t causing the pain – they are extremely common – and the interventions weren’t useful.

    But it was hard for people to accept that because it always seems better to know more rather than less.

    • Ahh, Jonah Leher – that makes sense. Must have read it in his blog, in his books or heard it on radiolab.

  4. Daniel J. Andrews

     /  October 15, 2010

    I think digital prostate exams cause prostate cancer. There’s a reason why the prostate is tucked out of the way. Don’t poke it, and it’ll be fine. 🙂

    Just joking. Physical exams are on my mind. I had my exam and blood tests two weeks ago. I do wonder just how much info about the health of the prostate you can determine by poking it though. If you want to write an article about that, PalMD…..


    Rectal exam is “forthcoming”

%d bloggers like this: