The incidence of stroke among American children could be two to four times greater than has commonly been estimated, a new study says.
Pediatric strokes are rare; even the new estimate puts its incidence at only 2.4 strokes per 100,000 person-years (person-years represent the number of years children were studied). But the report, published in the Sept. 17 online issue of Stroke, also cites five previous studies in which the estimated incidence ranged from .54 to 1.2 per 100,000 children per year.
“All those studies relied on billing coding to identify patients with stroke,” explained study author Dr. Heather J. Fullerton, director of the Pediatric Stroke and Cerebrovascular Disease Center at the University of California, San Francisco, Children’s Hospital. “Our study looked first at billing, and then at radiology reports.”
Fullerton and her colleagues looked at data on 2.3 million children up to the age of 19 who were enrolled in the Kaiser Permanente managed-care plan in northern California from 1993 to 2003. They searched for stroke cases listed by diagnostic code for billing purposes and also for reports indicating strokes in radiological studies, including computed tomography and MRI. The radiology reports yielded a higher incidence of stroke.
Estimates based on billing are not reliable for several reasons, Fullerton said. “Children who have strokes often are ill for other reasons, such as meningitis or congenital heart disease, so they are coded for that,” she said. “Also, coders apply the stroke diagnoses less often for children because they are perceived as rare events. Or they get nonspecific diagnoses.”
The concept that children simply don’t have strokes is widely prevalent, Fullerton noted. “Even among insurance companies, the impression seems to be that strokes don’t happen to children,” she said. “I have received calls from insurance companies questioning a diagnosis of stroke in a child.”
Dr. E. Steve Roach, director of pediatric neurology at Nationwide Children’s Hospital in Columbus, Ohio, who said, “I’ve been studying stroke in kids for 20 years,” added that “the new study confirms what I’ve been saying all that time. It’s just way, way underdiagnosed.”
Roach said he led two studies, one reported a decade ago, the other last year, that came to the same conclusion. “When we did a search of discharge records, a lot of them just did not show up,” he said.
“Most people, and even most doctors, can’t accept the idea that children have strokes,” Roach said. “It’s just a bias on the part of physicians that stroke just doesn’t occur in children.”
This new comparison of the two methods of identifying a pediatric stroke found that radiological evidence was much more sensitive than the billing code. The radiology method was far more sensitive (83 percent) than the billing code method (39 percent.)
The difference was even greater for strokes that occurred at the time of childbirth, with a sensitivity of 12 percent using the billing codes and 87 percent using radiological records.
But the study came with a word of caution. Because it was based on data from one health-care organization, it’s not clear whether the difference in incidence between billing codes and radiological findings will be found elsewhere, the report noted.
More studies in different populations are needed to firm up the finding, Fullerton said.
Failure to identify a childhood stroke properly might not be an issue for the child’s family, Fullerton said. “It is more an issue for those of us who investigate strokes,” she said. “When we try to identify pediatric strokes, it is important that we recognize that we could be missing a large chunk of cases.”