Overdiagnosis Problematic in Pediatric Medicine

TORONTO — In pediatric medicine, unnecessary testing can be eliminated without compromising outcomes if every physician simply asks how the test will benefit the patient, new data show.

Overtesting and overdiagnosis is prevalent in pediatrics, said Eric Coon, MD, from the University of Utah School of Medicine Primary Children's Hospital in Salt Lake City.

And the increasing incidence of disease might be directly related to overdiagnosis, he said during a special session here at the Pediatric Academic Societies 2018 Meeting.

An increase in the incidence of a disease without any change in morbidity or mortality from that disease is a sign that the abnormalities being detected are not that severe because they are not affecting patient outcomes, he explained.

For example, Kawasaki disease has long been treated with intravenous immunoglobulin to prevent the development of coronary artery abnormalities that can progress to adverse outcomes, such as thrombosis.

The most recent guidelines on the management of Kawasaki disease call for early and repetitive echocardiography — at diagnosis and 2 and 8 weeks after diagnosis, or more frequently if a coronary artery abnormality is detected — and the use of the z-score to lower the threshold at which physicians label a coronary artery "abnormal".

Kawasaki Disease

For their study, Coon and his colleagues set out to determine whether the practice of aggressively identifying coronary artery abnormalities has led to the overdiagnosis of those abnormalities in children with Kawasaki disease.

The team assessed 342 children with Kawasaki disease who experienced an adverse cardiac outcome. Over the 15-year study period, the rate of coronary artery abnormalities per 1000 Kawasaki patients doubled.

During the same period, "adverse outcomes related to coronary artery abnormalities remained very stable," Coon reported. "Together, these two trends fit the pattern of overdiagnosis."

This increase in the detection of coronary artery abnormalities was primarily driven by an increase in nonsevere coronary artery abnormalities, "further supporting the concept of overdiagnosis," he said.

Reversing the trend toward overdiagnosis will reduce the need for more frequent follow-up and more testing — both of which can be costly — and will reduce the stress parents can feel when they are told that their child has an abnormality in a vessel that is critical to the heart.

"As providers, we may think the abnormality is fairly benign," Coon explained. But studies have shown that parents can develop a persistent belief in their child's vulnerability despite full recovery from an illness.

This belief — called the vulnerable child syndrome — often results in parents restricting children from participation in physical activities later in life because of a "heart problem," he pointed out.

Overdiagnosis can also happen when a child with isolated head trauma presents to the emergency department.

Head Trauma

The physician must decide whether or not the child should undergo a CT scan "to find a fracture or bleed, particularly a slow bleed that, if missed, could extend to catastrophic consequences," Coon explained.

But the use of CT scans has been on the decline because of concerns that radiation overexposure can contribute to malignancies.

So Coon and his colleagues examined whether the decrease in CT scans was accompanied by a decrease in the detection of abnormalities and, if so, whether patient outcomes were affected.

The team assessed the records of 300,000 children treated for isolated head trauma at 34 children's hospitals in the United States from 2003 to 2015. The use of imaging, including CT scans, peaked at about 40% in 2008, but declined to 25% in 2015.

The incidence of skull fractures and bleeds both declined during the study period, especially after 2008. These declines were accompanied by a decrease in hospitalization rates and neurosurgery, again largely after 2008.

Rates of revisits to the hospital in the week after the index event were exceedingly low during the study period. And mortality and persistent neurologic impairment were very rare outcomes in these children.

"In other words, decreased imaging was accompanied by decreased detection of abnormalities and decreased intervention without measurable harm to the patient," Coon reported.

Although the use of bicycle helmets — and perhaps seatbelts and even the heightened awareness of concussion — could explain why children with less-severe head injuries were seen over time in the emergency department, "the trends we found were consistent across age groups, so increased use of bike helmets should not affect children under the age of 2 years," Coon observed.

"The implication here is that we can safely do less while decreasing radiation exposure and reduce overdiagnosis," he concluded.

Drivers of Overdiagnosis

Patient and family pressure to "do something" if a child is perceived to be ill is often cited as a common driver of overdiagnosis, as is peer pressure, or at least the fear of being judged incompetent if caught doing nothing.

Malpractice concerns are often thought to be a key driver of overtesting and overtreatment, although physicians at the meeting were divided on how influential the fear of litigation is in shaping physician behavior.

Other less-recognized influences of physician behavior might be as simple as being presented with an abnormal result.

A study on the use of vesicoureteral reflux imaging and prophylactic antibiotics was conducted by a team led by Alan Schroeder, MD, from the Stanford University School of Medicine in California.

Physicians were given a clinical vignette, asked whether or not imaging should be ordered, given either a normal result or a result indicating that the patient had grade 2 reflux, and then asked whether antibiotics should be prescribed (Hosp Pediatr. 2018;8:21-27).

The findings show that "we have a hard time not responding to abnormalities from tests, even if it is a test we wouldn't have ordered in the first place," he added.

Table. Physicians Who Would Prescribe Prophylactic Antibiotics

Results Provided	Imaging Ordered, %	Imaging Not Ordered, %
Normal	6	0
Grade 2 reflux	56	39

Guidelines themselves can sometimes drive what, on the surface, appears to be excessive testing.

For example, recommendations for dyslipidemia from the National Heart, Lung, and Blood Institute and the American Academy of Pediatrics suggest that every child 9 to 11 years of age be screened for abnormal lipids. They also recommend that pediatricians do at least two fasting lipid panels, beginning at 2 years.

Although experts here agreed that hardly any pediatricians follow these recommendations, they still exist.

An assessment of these guidelines was presented by Thomas Newman, MD, from the University of California, San Francisco.

If adult lipid guidelines were applied to young patients, practitioners would end up treating 78,000 patients 17 to 20 years of age; however, if pediatric treatment guidelines were followed, 483,000 patients 17 to 20 years would be treated (0.4% vs 2.5%).

And the risk–benefit ratio of treating children and adolescents with dyslipidemia is quite different than it is in adults.

"If you treat 100,000 high-risk adults and lower their risk of a cardiovascular event by 30% with statin therapy, you will prevent 3000 cardiovascular events but you will cause 100 to 200 excess cases of diabetes, which is a pretty good trade," Newman said.

"But in children, there is no heart disease to prevent," he said. "If you prevent even one cardiovascular event in children and adolescents, you can cause as many as 400 to 500 cases of diabetes."

Detection of an abnormality that does not benefit the patient is another definition of overdiagnosis, Newman explained.

The right question is, 'How will this test provide net benefit to my patient?'

Changing the mindset of residents in training might be a good place to start to reverse the drive to overtest and overtreat.

Residents are often asked to go through a diagnostic-dilemma exercise, in which they are presented with the most esoteric, most random, most fascinating disease their mentors can come up with, Schroeder explained.

"We go around the room and create this tremendous list of diseases that many of us have never seen before, and then the laundry list of tests starts. We don't give a lot of thought about how that test will help patients," he said.

Teachers should move away from asking what the patient has, he suggested, and simply ask, "How can the test help this particular patient?"

Many think that an evidence-based approach to testing is to ask whether test results will change management, "but I don't think that is the right question," Schroeder said. "We've shown that test results can change management without benefiting patients."

"The right question is, 'How will this test provide net benefit to my patient?'" he pointed out. "This is one way to mitigate some of our concerns about overdiagnosis."

We have to learn to be comfortable with uncertainty, focus more on value and less on cost, and address clinician fears about underuse.

Physicians should engage patients and the public, wherever possible, by promoting campaigns such as Choosing Wisely, where doing less is among the key goals, said Virginia Moyer, MD, from the Baylor College of Medicine in Houston.

And physicians should be wary of expanding disease definitions, such as the recent change in the definition of hypertension, which increases the number of adults who will now be considered hypertensive, she cautioned.

"Most quality measures we use focus on underdiagnosis, but we need to think about quality measures that focus on overdiagnosis," said Moyer. "And we need to measure errors of commission, not only errors of omission."

Because the United States is one of the few countries in the world to allow the direct marketing of pharmaceutical and ancillary products to consumers, physicians need to warn patients against egregious enterprises, such as Life Line Screening, offered by companies with no medical expertise, she explained.

Tests are becoming increasingly sensitive and, as such, "if we apply the old guidelines to the new test, we will be misinterpreting test results," she stressed.

"We also have to critically evaluate standard practice," such as the practice of prescribing long courses of antibiotics when much shorter courses will do, she added.

We are taught to first do no harm, be we actually think, "first do something," Moyer said. "We have to learn to be comfortable with uncertainty, focus more on value and less on cost, and address clinician fears about underuse."

Coon, Schroeder, Newman, and Moyer have disclosed no relevant financial relationships.

Pediatric Academic Societies (PAS) 2018 Meeting. Presented May 7, 2018.

Follow Medscape Pediatrics on Twitter @MedscapePeds

Comments

Commenting is limited to medical professionals. To comment please Log-in.

Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.