Better in their words…

By Adam Cifu, MD, Feb. 28,  2023, Sensible Medicine on Substack

Source: Characteristics of a Successful Screening Test

A screening test is meant to find a disease in a person who has no symptoms. The intent of screening is to improve outcomes for the person, improved as compared to the outcome he would have had if you waited for him to develop symptoms.

I’ve been giving talks about screening tests for decades. My lectures have evolved as our data has matured and as I have gained experience as a general internist. However, some version of this slide has always been included.

Five bullet points. Endless complexity. Here is an attempt to unpack this slide.

As a general internist, screening is a big part of what we are expected to do. I am judged by what percent of my eligible patients have had colon cancer, breast cancer, and cervical cancer screening. The idea that screening helps is baked into our idioms: “An ounce of prevention is worth a pound of cure.” Over the time that I have given these talks I have slowly, but definitely, come to the opinion that physicians are much more effective when we treat diseases that cause symptoms and shorten life rather than when we work to find asymptomatic diseases. However, I am by no means a screening nihilist.

1.      Do no harm

Screening tests are meant to help people. They are supposed to diagnose a disease in an early stage when it can be treated and cured. Better to find that breast cancer before it becomes a lump or a coronary artery blockage before it causes angina (or worse, a heart attack or sudden cardiac death). So goes the reasoning at least. The problem is that finding disease early does not always help. In many cases, we have found that diagnosing a disease in an asymptomatic person provides no benefit. This is true for ovarian cancer. It is true for coronary artery disease. It is true for pancreatic cancer. It will be true for some diseases that we presently screen for as our treatments get better.

Sometimes, screening not only does not yield benefits, it causes harm. Our screening tests may yield false positives. We tell people that they may have disease when they do not. About 60% of woman who get mammograms for 10 years will have a false positive mammogram. A false positive leads to anxiety, further testing (which is often expensive and invasive) and has the effect of making a person less likely to continue screening. (“I never want to go through that again!”)

Possibly more harmful than the false positive is overdiagnosis. This occurs when we find a cancer, and treat it, but it is a cancer that would have never harmed the patient. About 6% of woman who get mammograms for 10 years will experience overdiagnosis. What is the harm of overdiagnosis? We tell people they have cancer, we treat the cancer, we “cure” the cancer, we turn the person into a “survivor.” They are now a patient (rather than a person) and a relentless advocate of screening.

2.      The burden of disease must be sufficient to warrant investigation

This point is obvious. We only screen for diseases that are common enough and severe enough to warrant the cost and effort. We screen for colon cancer but not pre-symptomatic onychomycosis (toenail fungus). However, this point is more complicated than it may seem. Who determines what a “sufficient burden” is? How much should our society spend to relieve that burden. (See #5).

3.      Screening must detect disease in the pre-clinical stage

A screening test is not a screening test if it only detects disease in people with symptoms. A test used in a person with symptoms is a diagnostic test. (Most of our screening tests are also used as diagnostic tests, colonoscopy for rectal bleeding, mammogram for a breast mass, chest CT for cough). There are diseases that just don’t have a detectable, preclinical stage. As our technology advances however, this list may get shorter.

4.      Screening tests must be tolerable and sufficiently accurate

If number 3 is specific to a disease, number 4 is specific to a test. Having a PSA test as a screening test for prostate cancer is acceptable to most people (it is just a blood draw) but a colonoscopy or mammogram is intolerable to many.

Then the test must be accurate. A test must be sensitive (likely to be positive in people with disease) and specific (likely to be negative in those without disease). Poor sensitivity leads to false negatives – an unreliable test – and poor specificity leads to false positives – greater cost and harm. Because we screen low risk populations, false positives are a particularly troublesome issue.

5.      Early detection must improve outcomes when compared to detecting the disease when it would naturally present clinically

This is what it all comes down to. The goal of screening is not to find disease but to help people. The fact is, we almost never know if a screening test is beneficial. It takes a randomized trial with long follow up to see if screening improved overall (or even disease specific) mortality. If these studies are not done well, many of the findings can be compromised by lead time and length time bias. Even a diagnostic test that lowers mortality would also, certainly, cause off target harm (see #1). After a positive trial, we would still need a societal discussion about whether the costs (monetary and otherwise) are worth it. We tend to not be very good at these conversations.

I try to keep a balanced view of screening. I am pretty sure we save hundreds of thousands of people a year with screening tests. But, we falsely reassure some people by telling them that they do not have disease. We put people through diagnostic testing after false positive tests that they would not have needed had they not been screened. And, in some people, we diagnosis and treat a disease that would never have caused them harm.

Share