NEW YORK — Children in Lebanon, N.H., are more than twice as likely to have their tonsils removed as those in Bangor, Maine. Kids in Lewiston, Maine, are 50 percent more likely to have a CT scan of their head than are kids in Portland, Maine, or Lebanon and Burlington, Vt.

Lebanon’s children don’t have especially infection-prone tonsils, and Lewiston’s don’t fall on their heads more than kids elsewhere do. Instead, according to a report released on Wednesday, the glaring variation means that in some cases “children are not receiving enough good care,” said pediatrician and health policy analyst Dr. David Goodman of the Geisel School of Medicine at Dartmouth, who led the study for the Dartmouth Atlas Project. But in other cases they “may be receiving unnecessary care that is harmful.”

Similar research from the project has shown that the rate of medical procedures performed on older Americans covered by Medicare varies enormously depending on where they live.

The new report, which focused on northern New England, is the first to show that geographic variability exists in children’s health care, too, raising questions about why tens of thousands of kids are not receiving recommended care such as screening for lead poisoning and why tens of thousands of others are subjected to potentially unneeded treatments such as CTs for stomachaches.

The Dartmouth Atlas was able to examine geographic variations in the medical care provided to children in Maine, New Hampshire and Vermont from 2007 to 2010 because those states are among the few that collect data on all health insurance claims, which the researchers analyzed.

“This suggests that there is a significant amount of overuse of medical services in some areas,” said Dr. Vikas Saini, a cardiologist and president of the Lown Institute, a health care think tank in Boston. “Especially because unneeded care can expose children to harmful side effects, this is very troubling.”

The idea that Americans undergo millions of unneeded procedures every year has become more widely recognized as a result of the Choosing Wisely campaign, in which medical specialty societies have for the past two years identified procedures that should be avoided or questioned. The campaign has been adopted by a number of health care groups in Maine.

That whether a child undergoes a procedure depends in part on where he or she lives has been known since the 1970s, when researchers led by Dr. John Wennberg of Geisel found that rates of tonsillectomy are 60 percent in some places and less than 20 percent in others. The difference, they found, could not be explained by underlying medical conditions among the children, and it was instead more likely due to the local medical culture: Some doctors believe in removing tonsils and others don’t.

Apart from the tonsillectomy work, research on variations in the health care children receive has lagged that on adults.

In the new study, the researchers found regional variations in rates of hospitalizations, common surgeries, imaging, prescriptions and office visits such as for well-child care and middle-ear infections.

Children in St. Albans and Bennington, Vt., had triple the number of annual office visits (3.6) as kids in Houlton, Maine (1.2). The presence of large medical centers, children’s hospitals and, therefore, pediatricians and specialists hardly mattered: Burlington and Bangor both have such centers, but the rate of office visits was more than 50 percent higher in Burlington.

In areas with fewer pediatricians, children went to emergency rooms more often, but what happened to them there varied even more.

Some of the greatest variation occurred in “preference-sensitive” care. These are procedures, such as placing tubes in the ears, that have unclear benefits but which parents or physicians may opt for anyway. There is no consensus on which children will get fewer ear infections or less hearing loss thanks to tubes, for instance.

As a result, preference ruled: Middlebury, Vt., has more than four times the rate of tube insertions (15.2 per 1,000 children) as Bangor (3.4). Littleton, N.H., has more than four times the rate of tonsillectomies (10.9 per 1,000 children), which are also of questionable benefit, as Bangor (2.7).

Differences in how many kids get ear, nose and throat disease are unlikely to explain these differences, said Lown’s Saini.

Almost nothing is known about the effectiveness of ear tubes for children under age two, but a substantial number of toddlers are undergoing the procedure, Goodman said in an interview with the Bangor Daily News.

“For parents, the takeaway is ask lots of questions about the care that your doctor is recommending, particularly for surgical procedures like [ear, nose and throat] procedures, and imaging studies and for medications,” Goodman said.

Particularly worrisome to experts is the high rate of CT scans in some places. CT delivers radiation equivalent to 200 to 400 chest X-rays, increasing the risk of cancer decades later. Children in Presque Isle, Maine, had 19.7 head CTs per 1,000, compared to 8.4 in Burlington. The American Academy of Pediatrics says head CTs are “not necessary” for the immediate evaluation of head injuries and pose “considerable danger to children.”

Bennington had more than triple the rate of abdominal or chest CTs (15.4) as Machias, Maine (4). The variation did not depend on whether a major medical center is nearby: The rate in Bangor (11.7) was more than double that in Lebanon (4.7), though both have large hospitals.

Use of medication for attention deficit hyperactivity disorder (ADHD) varied by a factor of 2, with 2.9 percent of Fort Kent, Maine’s, children on the drugs but 8.1 percent of Ellsworth’s. There was a comparable gap in use of antidepressants and antipsychotics, whose use in children has been widely criticized.

More doctors in an area did not necessarily increase the rate of questionable procedures, while fewer doctors did not reduce it. Instead, overuse is likely a result of the prevailing “medical culture,” said Saini: Doctors who walk the same hospital corridors and socialize through the same professional and other organizations trade anecdotes about what works.

“Most of what we do in medicine doesn’t have empirical evidence” for whether it works and for whom, said Saini. “Instead, it’s driven by anecdotal evidence and professional opinion,” which doctors who practice in the same area are likely to hear about and be influenced by, especially early in their careers.

Although the Dartmouth researchers underlined the overuse of some medical procedures in some places, they also found worrisome underuse of proven therapies. Practice guidelines call for strep throat to be treated with antibiotics. More than 90 percent of such cases were properly treated in Exeter and Derry, N.H., while only 41 percent to 47 percent were in Calais, Presque Isle and Houlton, Maine.

In some areas where Maine scored relatively poorly, such as lead screening for children under age two, improvements have been made since the data was collected a few years ago, according to Amy Belisle, director of child health quality at Maine Quality Counts.

The report allows health care providers to gauge their performance in a new way, the first step in learning how to better care for children, she said.

“That’s the power of this report, it gives us the data and a starting place on where we should be looking next around children’s health-care quality,” Belisle said.

The Dartmouth Atlas Project published the data at

Bangor Daily News health editor Jackie Farwell contributed to this report.