Your doctor wants to schedule a test—but not all medical tests are created equal
By Wendy Haaf
Some medical tests are intended to detect serious diseases in apparently healthy people before symptoms appear—a good idea, but what are the trade-offs? Research suggests that many of us simply don’t know.
In a number of studies that compared participants’ expectations with the actual risks and benefits associated with common screening tests, most people underestimated the magnitude of the risks and overestimated the benefits, often grossly. For instance, in one New Zealand study in which respondents were asked to say how many lives would be saved by screening 5,000 people for colon cancer over 10 years, 94 per cent overestimated the number supported by scientific evidence, with onethird overshooting the mark 200-fold.
Moreover, because medical knowledge evolves as research leads to new insights, some experts believe that the harms of certain screening tests now widely accepted may actually outweigh the benefits.
We decided to delve into the data on four conditions for which either screening isn’t supported by solid evidence or the risk/benefit balance has begun to shift. However, you can apply what we learned about weighing the pros and cons to any screening test. So…what should you know before you go for your next screening test?
No Test Is Perfect
To begin, let’s backtrack a bit and clarify what we mean by a screening test.
“Screening means going after a whole population at risk and applying a test that you hope will give you advance warning of a condition; it’s a way of picking up a condition before it becomes clinically obvious,” explains Dr. Christopher Patterson, professor emeritus in the Division of Geriatric Medicine at McMaster University and chief of geriatric services at Hamilton Health Sciences in Hamilton, ON. For example, if a woman is undergoing a mammogram to look for signs of breast cancer simply because she’s reached an age at which the dis ease becomes more common, that woman is being screened—which is very different from using more specialized tests to determine whether a lump in a woman’s breast means cancer.
Another important point about screening tests: they all have limitations.
“All tests have false positives and false negatives,” says Dr. Ross Upshur, Canada Research Chair in Primary Care Research and a professor in the University of Toronto’s Department of Family & Community Medicine and Dalla Lana School of Public Health. “With a false negative, you think, I don’t have the condition, but you do, and with a false positive, you think you have it but don’t. There’s no such thing as a perfect screening test, and people need to be mindful of that.”
While subsequent testing can often sort out false positives, those are not the only problems that can arise. In some situations, tests may uncover evidence of a disease that might never pose a threat to the person’s health. For example, some cancers grow very slowly, some at a moderate pace, and others very rapidly. Some experts liken these three categories of cancers to turtles, rabbits, and birds. Birds move so quickly that screening and treatment don’t improve the odds for survival. Rabbits, on the other hand, are slow-moving enough that these interventions can make a difference. The trouble is that the very slow turtles get caught in the screening net, too, and we don’t yet have tools that can reliably distinguish between rabbits and turtles.
In the case of some screening tests, that means a sizable number of people are treated for a disease that might never have caused harm. “If you have a disease that you’re not going to die of but you treat it anyway, the outcomes are going to be pretty good,” says Dr. Neil Bell, a professor in the Department of Family Medicine at the University of Alberta and a member of the prostate cancer screening group for the Canadian Task Force on Preventive Health Care (CTFPHC, which weighs scientific evidence behind various screening tests and issues recommendations based on its findings). “This is a concept called overdiagnosis, and prostate cancer is the poster child for overdiagnosis.”
Prostate Cancer Screening
The only screening test we currently have for prostate cancer is a blood test that measures levels of a substance called prostate-specific antigen (PSA), appropriately called the PSA test. But this test won’t detect only prostate cancer—other conditions, such as an infection or benign enlargement of the prostate gland (the latter of which becomes increasingly common with age), can cause elevated PSA levels. In addition, there’s no magic cut-off—no low level of PSA that signifies the absence of prostate cancer; thus, the lower you set this bar, the more false positives you’ll find. “They’ve used various cut-offs; 2.5 [ng/mL of blood] is more or less standard now,” Bell says.
Men who receive that result or higher have to decide whether to have a prostate biopsy, a procedure that involves taking between 12 and 36 samples of tissue from the prostate with a thick needle, via the rectum. While a minor procedure, biopsy itself carries risks. “One or two men per thousand die within a defined period of having the biopsy done—whether as a direct result of it or for some other reason, it’s hard to know,” Bell says. Another two per cent will end up in hospital, usually due to infection.
One large European study screened men aged 55 to 69 for 13 years and found that, using a PSA of 3.0 as a cut-off, 178 of 1,000 patients received false positives (four of whom experienced biopsy-related complications) and 102 developed prostate cancer (33 of whom would never have succumbed to the condition).
Yet while there are other options, including repeat PSAs and periodic PSA plus biopsy, many men will opt for treatment—radiation, or surgery to remove the gland. Of every 1,000 men who receive treatment for prostate cancer, 114 to 214 will experience shortterm complications (such as infection and the need for transfusion), 127 to 442 will experience long-term erectile dysfunction, and 178 will experience urinary incontinence. Another four to five will die from side effects of treatment itself, and five will die from the disease despite screening.
What it comes down to, Bell says, is that “if you screen, five per 1,000 men will die, and if you don’t screen, six per 1,000 men will die.” In light of these figures, Bell says, “Our recommendation for men 55 to 69 was a weak recommendation against screening. What that means is that doctors should talk to their patients about the good and the bad of screening and inform them.
“In practice, what happens is that most physicians would probably not do anything unless a man asked, ‘What about PSA? Some doctors don’t screen anymore, and some screen everybody.”
Breast Cancer Screening
After the CTFPHC last reviewed the evidence for screening mammography in women aged 55 to 64 at average risk of breast cancer, they issued a weak recommendation in favour of doing the test every two years.
So how did the numbers stack up? Among 100,000 women in this age bracket screened every two years for 11 years: 510 will die of breast cancer, 28,200 will experience a false alarm, 3,700 will undergo a biopsy, 500 will have all or part of a breast removed without having cancer, and 138 will be treated for a discovered cancer.
On the other hand, among 100,000 women aged 50 to 69 who don’t undergo screening over 11 years, 640 will die of breast cancer but the remaining 99,360 won’t, nor will they be exposed to biopsies or unnecessary surgeries or radiation.
Put another way, if you’re a woman between the ages of 50 and 69, your chances of dying of breast cancer are 1 in 196 (0.51 per cent) if you’re screened and 1 in 155 (0.65 per cent) if you’re not screened, a difference of only .14 per cent. (If you’re between 70 and 74, the benefit of screening is slightly greater: among women who are screened, the risk of dying of breast cancer is 1 in 217, or 0.46 per cent, versus 1 in 146, or 0.68 per cent, among women who aren’t screened.)
If you’re 55 to 69 and you’re screened, your odds of having a false positive requiring further screening are 1 in 4; your risk of having a biopsy is 1 in 28, and your odds of having a breast or part of a breast removed unnecessarily are 1 in 200. (As with prostate cancer, some breast cancers, as well as many cases of potentially precancerous lesions called DCIS, which are currently treated as aggressively as breast cancer, wouldn’t cause problems if left alone. Unfortunately, we don’t yet have a reliable way of predicting which cancers and DCIS lesions won’t progress.)
Moreover, “there’s lots of documentation that invitations to screening do not include information about the risk of overdiagnosis,” says Dr. Cornelia Baines, professor emerita in the epidemiology division of the University of Toronto’s Dalla Lana School of Public Health. And more evidence has come to light in recent years that casts doubt on whether screening mammography saves more women than it harms. For one thing, a Canadian study published in 2013 (coauthored by Baines), which analyzed a 25-year follow-up of 90,000 women, found that mammography didn’t reduce deaths from breast cancer.
While not all medical professionals have been persuaded to her point of view, all of this has led Baines to conclude that “screening is doing more harm than good. I think the evidence is very strong that the risk of being overdiagnosed, and overtreated with toxic therapy, is much greater than the likelihood you’re going to escape death from breast cancer.”
“There’s been a big shift over the last several years in terms of the guidelines on when to do bone density testing” to screen for osteoporosis, says Dr. Scott McKay, Chief of Family Medicine at Victoria Hospital, part of Western University’s Schulich School of Medicine & Dentistry in London, ON. (The scientific advisory committee for Osteoporosis Canada issues screening recommendations; however, this body doesn’t break down the risks and benefits of screening in the same way the Canadian Task Force on Preventive Health Care (CTFPHC) does.) “In the past, it was typical that every woman would get a bone density test done once they went through menopause—so, in their 50s. And then we really looked at the number [test results] on the bone density” as a reflection of bone health and fracture risk.
Since new Osteoporosis Canada guidelines were released in 2010, McKay says, “there’s been a shift away from the number itself towards the patient’s overall fracture risk.” This risk depends on factors such as age, weight, a history of fracture, a parent who suffered a hip fracture, smoking status, the use of certain steroid medications, and alcohol intake. (The World Health Organization has devised a tool called FRAX, which calculates 10-year fracture risk by taking all of these risk factors into account.) “We really should be using the medications to treat bone density only in patients who are at the highest risk for fracture because these medications aren’t benign,” McKay says. In addition to increasing the risk for rare complications such as osteonecrosis (bone death) of the jaw and an uncommon type of femur fracture, “they cause a lot of stomach upset and heartburn.”
Furthermore, now that they have been in use for some time, McKay says, experts have come to recognize that these medications “likely have a limited lifespan in terms of how long they’re going to be effective.
“We know they’re effective for five or 10 years, but we really want people to have that five to 10 years when they’re at highest risk. If we treat them from 55 to 65, we may have missed the window because that patient may have needed to be on the medication from 65 to 75 or 75 to 80.”
Consequently, “the new recommendations are to do bone density testing when you have one major or two minor risk factors for osteoporosis,” McKay says. “For most people, turning age 65 is the first major risk factor they hit.
“Then you get into, How often do you repeat the test? In the past, we tended to repeat it every couple of years. If you have a normal bone density, you probably don’t need to have the test repeated for five years. Bones don’t change that quickly, and the recommen dations for keeping your bones strong when you’re 65—weight-bearing exercise, vitamin D, and good dietary calcium intake—are things people really should be doing anyway.”
Cognitive Impairment and Dementia Screening
“Ten criteria for screening for cognitive impairment were originally set out in 1968 by the World Health Organization,” says McMaster’s Dr. Christopher Patterson. “Those criteria include things such as the condition being an important health problem, which clearly cognitive deficits are. Another is that there should be an accepted, effective treatment for the disease. Well, there aren’t, really, at the moment. There are some rather unsatisfactory drugs, however.
“There should be facilities for diagnosis and treatment, which there are in some parts of the country—but pretty thin in others. There should also be a suitable test for detecting the condition, but there really isn’t,” Patterson says. “The common tests proposed for screening include things such as a mini mental-state examination, the Montreal cognitive assessment, and the minicog [all questionnaires]. They all have a number of false positives and false negatives. For example, a mini mentalstate examination will miss a diagnosis of dementia in about one person in five and will incorrectly label someone as having a problem in about one person in five.”
Another of the criteria for screening is that “the natural history of the condition, including the development from latent to declared disease, should be adequately understood,” Patterson says. “We know that in many people with Alzheimer’s disease, there’s a period beforehand during which they have memory deficits that aren’t interfering with their daily life—it’s a condition called mild cognitive impairment. Each year, about five to 10 per cent of those with mild cognitive impairment develop dementia, but 50 per cent never do—some of them even get better.”
That means that while there’s a good tool for detecting mild cognitive impairment— the Montreal cognitive assessment— in many people deemed to have the condition, it would never progress to dementia. “Those that aren’t going to experience progression are being unnecessarily worried.” And in December 2015, following a review of international studies, the CTFPHC recommended against screening for mild cognitive impairment in seniors in the absence of symptoms such as memory loss, saying there was no evidence that screening offered any benefit.
Furthermore, simply telling people that they have a condition can have adverse effects.
“One of the best studies to look at this phenomenon, which is called labelling, was done in 1978,” Patterson says. A team of researchers measured blood pressure in 5,400 Hamilton, ON, steelworkers, ultimately diagnosing 245 of them with high blood pressure. These men were randomly assigned either to be told they had the condition or not to be told. “The people who were told they had high blood pressure started getting headaches, and their absenteeism increased by 80 per cent in the next year,” Patterson says. “Simply telling someone he or she has a condition changes his or her behaviour. If you medically label someone incorrectly— in other words, if a test produces a false positive—you’re creating problems that you shouldn’t.”
Perhaps it’s not surprising, then, Patterson says, that the US Preventive Services Task Force, “the most rigorous authority for assessing preventive health care measures, concluded that there is insufficient evidence to recommend—or to recommend against—screening for cognitive impairment.”
Things to Consider
No matter how much evidence there is to support the use of a particular screening test, you have to make the decision for yourself based on the best information available, as well as on factors such as your overall health, how long you can expect to live, how treatment could affect your quality of life, and whether you’d opt for treatment at all. After all, if you’ve been through cancer treatment twice and decided you’d forgo it a third time, you might ask yourself why you’d bother getting screened. “The most important thing is to reflect on your own values and preferences,” says the University of Toronto’s Dr. Upshur.
Another important consideration is age.
“Why do we continue screening into the eighth and ninth decade when there’s no evidence of benefit? Nobody has studied these populations and you can’t extrapolate the benefits from a younger population. The point of screening is to reduce mortality, and doing so gets more and more difficult to prove as you get older,” Upshur says. “There’s good reason to be skeptical about the benefit of some screening tests in older adults. And there are a lot of things older adults can do—such as eating well, exercising, and staying socially active—that are much more influential in enhancing their health and well-being than going for a screening test. Social isolation for older adults is like smoking: it really cuts into well-being—and life expectancy.”
QUESTIONS TO ASK: How accurate is this test? What are the rates for false positives/false negatives? What are the risks of having this test? What’s the chance I could be overtreated for a condition—that the condition ultimately wouldn’t have harmed my health anyway? If I get a positive result, what happens next? What happens if I decide not to undergo the test? Resources Choosingwiselycanada.org Canadiantaskforce.ca Healthydebate.ca taskforce.org