Just last week I saw a full-page ad in the newspaper for a series of medical screening tests—EKG, echocardiogram, aortic ultrasound, and the like (all interpreted by “certified doctors”)—that a person could arrange for oneself or a loved one. We newspaper readers were told we could arrange to have these tests performed without a doctor’s order as long as we’re willing to pay for them ourselves. We could choose à la carte from a menu of screening tests or order a group of them as a package (Do you have a relative who died from a brain aneurysm? Purchase peace of mind with the vascular package, including EKG, carotid ultrasound, echocardiogram and abdominal ultrasound!). They even had testimonials from lucky survivors whose lives were saved by early, proactive testing.
The purveyors of these roving imaging clinics are looking for people with no particular symptoms but who are anxious about the possibility of occult disease lurking inside their bodies—the “ticking time bomb.” Insurance companies don’t typically sport for most of these tests unless your doctor thinks you have symptoms to warrant them.
Should you sign up for testing? Seems like a bargain, don’t you think? A few hundred bucks and you get either peace of mind or early detection that could save your life. Before you call the 1-800 number let’s look more closely at one of these tests.
The carotid Doppler ultrasound (CDUS) is a study that is meant to detect blockage in the arteries of the neck that provide blood flow to the brain. High-grade carotid artery stenosis (CAS) is a risk factor for stroke and we know that CDUS is a critical part of the evaluation in patients who’ve suffered a stroke or transient ischemic attack (TIA, or “mini-stroke”).
But what about people like you and me who have no weakness, numbness or slurred speech? To further tease out whether this is a useful test for those of us who are asymptomatic I’ll need to guide you through the complex web of population statistics, and to do that without inducing irreversible narcolepsy I’ll need to use a simple analogy.
Let’s say you’re the ruler of a country with a small population, a thousand of whom are over the age of 65. This group of geriatric citizens is very lucky in that they are not burdened with more than the average collection of health woes: some with hypertension, others with diabetes, a few smokers. None of them have had a stroke or any symptoms suggestive of carotid vascular disease.
Your country struck it rich with the discovery of some valuable mineral that the world just can’t live without. You are now so wealthy that your citizens don’t pay taxes and all their municipal needs are provided painlessly. Needless to say, your nation has also constructed the most comprehensive healthcare plan imaginable with state funding for every procedure, test and therapy that modern medicine can devise.
You, as the benevolent ruler that you are, want no one to suffer a stroke, and when you hear that you can test the carotid arteries with CDUS you jump at the opportunity (all costs be damned!). You beam with pride as you announce that all one thousand citizens will undergo CDUS to screen for possible carotid vascular narrowing.
How many lives do you save through your magnanimous efforts?
I did a little research on the subject and have pulled some fairly reliable data on which I base my estimates. Out of a thousand asymptomatic people over the age of 65 in the United States only 1% (ten patients) will have CAS that is significant (defined as narrowing of greater than 70% vessel diameter). This sets the stage for our experiment in disease screening—we will be looking for 10 patients with disease among 1000 people we screen.
Now, using the wizardry of my higher math skills (adding, subtracting, pushing buttons on a calculator) I conclude that CDUS on these thousand people will produce a negative (normal) result in 892 patients and a positive (abnormal) result in the other 108*.
“Wait!” you say, “there are only 10 people in this population with carotid blockage. How can we have an abnormal test in 108 people?”
It’s fairly well established that CDUS is not a perfect test. It tends to over-diagnose some people and under-diagnose others. A false positive result arises when the CDUS suggests blockage that isn’t really there, and a false negative means that the CDUS failed to detect blockage in an individual with significant narrowing. We quantify the diagnostic accuracy of any test by using the concepts of sensitivity and specificity (please see a previous post for more fun with sensitivity and specificity, concepts which have bedeviled medical students for ages) and can translate them into more useful parameters by incorporating what we know about the prevalence of CAS in the population.
In our group, where patients with CAS make up only 1%, the CDUS will produce far more false positives (99, to be exact) than true positives (9). And, since CDUS is such an imperfect modality, all 108 of our patients with abnormal studies will need further testing to determine who really has meaningful CAS. The next test would be either an MRI or invasive carotid angiography.
Carotid angiography is the “gold standard” study that provides the definitive answer. The problem with this method is that it is invasive and fraught with about a 1% risk of stroke (as well as other less dangerous problems). If we do carotid angiography on all 108 patients we will leave one of them with a potentially debilitating stroke.
MRI is nice because it is noninvasive. Its only drawback is that it, like CDUS, can produce false positive results, albeit less frequently. Using MRI in our 108 patients will result in some of them going on to surgery who don’t really need it.
Once we’ve narrowed down our 108 CDUS-positive patients to only the 9 who have real CAS (not 10, by the way, because we missed one with true CAS), we have to ask ourselves “What then?” Do we send them all to surgery? Carotid endarterectomy (CEA) is a good, time-tested procedure, but even in the best of scenarios it is accompanied by significant stroke in 3% of patients. The benefit of CEA in patients with asymptomatic CAS is not clearly proven and is reserved mainly for younger, healthier candidates with good life expectancy (see American Heart Association guidelines).
So, in summary, for our 1000 asymptomatic test subjects we will have correctly detected CAS in 9 of the 10 with this problem, but at a cost of 1 disabling stroke (presuming the use of angiography as the follow-up test). We will have failed to detect true CAS in only one individual. Those going on to CEA surgery will face a 3% risk of stroke at the time of the procedure.
“But,” you counter, “we’ve prevented strokes in 9 patients. Surely that’s worth something.” Not true. A patient over the age of 65 with high-grade CAS but who has no symptoms still has only an 11% risk of stroke over the next 5 years. Eleven percent of 9 is exactly one—you prevented one stroke over 5 years and you caused one through your testing. Net score: zero.
As you can see, the utility of screening for CAS in asymptomatic people is far from proven. The U.S. Preventive Services Task Force—which recommends against such screening—calculates that we would need to screen 8696 people to prevent one disabling stroke. And although you, as the ruler of a fabulously wealthy nation, wouldn’t care about the financial burden of such a screening effort, you’d have to recognize that the cost of such screening and follow-up testing would be astronomical.
Now back to real life. Do you get yourself screened or not? Still hard to say. It’s one thing to talk in terms of “populations” and another to discuss you personally. If you’re at high risk for CAS (smoker, high blood pressure, high cholesterol, etc.) your chance of CAS and stroke might be considerably higher than the population average of 1%. In that case it may be worth your money and time to get screened (it goes without saying that you need to aggressively treat your risk factors as well).
If, on the other hand, you’re a healthy but anxious middle-aged adult with few risk factors, and don’t happen to be the ruler of a small, independently wealthy nation, you could probably find better ways to spend your money.
*Sensitivity and specificity for CDUS is documented in many studies to be right around 90%. For those of you who question the results I’ve come up with I encourage you to dust off your medical statistics textbooks and run the numbers for yourself. The high number of false positives is a weakness of all screening tests when applied to a population with low disease prevalence.