The result sound counterintuitive - a recent Cochrane systematic review of studies looking at the benefits of routine health checks found that there weren't any. How can that be? Screening for medical conditions that can be treated, where treatment is known to improve outcome, must be beneficial, right?  

This result is a little surprising, but not much. There are two broad explanations for the negative outcome of this review: that the studies were not able to detect a real benefit, or that there is no real benefit.  

The studies in the review were not perfect. Many of them were decades old and therefore may not apply to current medical practice. Some outcomes, like hospital admissions and disability, were poorly followed. But collectively the studies were fairly large and they found no difference in overall mortality - so it seems that a routine doctor visit does not make you live longer.  

Assuming this conclusion is close to the truth, how can we make sense of that? It largely comes down to understanding statistics - something of which people generally have a very poor intuitive grasp. The result of this review seems counterintuitive because our statistical intuition is inherently flawed.  

Screening tests of all kind have a certain specificity (the chance of a positive outcome being a true positive) and sensitivity (the chance of a negative outcome being a true negative). Let's say we have a screening test with a 99% sensitivity and specificity, which is very good and better than most tests using in medical diagnosis. Let's further say we are screening for a disease with a prevalence of 1 in 1000. If we test 1000 people then statistically 11 will test positive - 1 true positive and 10 false positive (1% false positive rate with a specificity of 99%). There is also a 1% chance that the true positive will be missed, or on average 1 positive person will be missed for every 100,000 people screened.  

The next question is - what do we do with the positive tests? Do they get treatment, or perhaps a follow up test that is more specific but is also more invasive and contains some risk? It is quite possible that we will do more harm to the 10 false positives collectively than benefit to the 1 true positive by early treatment.  

The point of this is that screening populations with a low risk is inherently problematic, and is very likely to cause more harm than good. Routine health checks is essentially about screening people without symptoms and who appear healthy, a group that has a statistically low incidence of the diseases you are screening for.  

But what about things like high blood pressure, high cholesterol, glaucoma, and diabetes - these are chronic illnesses and it has been clearly established that treating them preventively improves health outcomes. The prevalence of these diseases is also quite high in the general population.

For these types of diseases it is possible that routine screening is not of statistical benefit because most people would get diagnosed even without such screening, diluting the benefit or adding routine screening. For example, people who visit their doctor for some other reason, like the flu, migraines, minor trauma, GI symptoms, or some other common ailment, are going to get their blood pressure checked, routine blood tests, a physical exam, and perhaps other tests and exams. In other words, it's likely that many (perhaps most) people with a chronic illness like high blood pressure will encounter the healthcare system at some point in their life and get screened for their blood pressure.  

Further, there are many reasons for asymptomatic people to suspect that they might have a chronic illness, such as family history. If everyone in your family has high blood pressure, you are likely to seek out screening (which is now no longer routine, but targeted).  

Therefore the general healthy population contains people at low risk for disease, and from which many of those with chronic illness have already been identified, so the prevalence of undiagnosed chronic illness is even less.  

There are other potential confounding factors as well. Perhaps the unhealthy population (those who might benefit from screening) disproportionately contains people who generally do not take good care of themselves and are not compliant with things like an annual health check.  

What, then, do we do with this information? It does not imply that you should stay away from the doctor, or that there is no benefit to testing for or treating common medical conditions like diabetes and high cholesterol. What it means is that we have to think carefully about how we spend our limited health care dollars.  

I do think that the existing studies are not the final word. The fact that they found no benefit from routine health checks does likely mean that any possible benefit is relatively small, but not necessarily zero. There is room for improved studies to look more carefully at how routine checks are done and all relevant outcomes.  

It is likely, however, that an annual routine health check is simply not effective and should be abandoned. In fact, it already largely has been. Most medical organizations stopped recommending a routine visit decades ago. (The US Preventive Services Task Force stopped recommending annual check-ups in 1989.).  

This does not mean doing no screening, however. Perhaps we need to enhance targeted screening programs, those focusing on high-risk groups. An optimal screening algorithm may contain the occasional routine visit, perhaps once a decade for those who have not visited a physician for any other reason in that time.  

An online questionnaire may also be highly cost effective and effective in identifying people that do need targeted screening. Online self-service visits are likely to be increasingly used in the future (at least that's my prediction). People already use the Internet more for health information than any other kind of information. Having a formal health portal that reminds people of healthy lifestyle practices, and screens them for conditions that might need testing or treatment, is at least cheaper than an office visit. Unfortunately, the Internet is currently overwhelmed with medical misinformation, and I can also see cranks and charlatans using this method to drive business.  

All of this is a powerful reminder that we need evidence-based standards in medicine more than ever. Health care systems are straining under rising costs. We need to use the best evidence available to figure out what not to do, because it wastes resources or even does net harm. Science remains our best tool for improving the health care system.


Steven Novella, M.D. is the JREF's Senior Fellow and Director of the JREF’s Science-Based Medicine project.