DENVER – A high degree of unwarranted practice variation exists among pediatricians and family physicians with regard to screening children for obstructive sleep apnea in accord with current evidence-based practice guidelines, Sarah M. Honaker, Ph.D., said at the annual meeting of the Associated Professional Sleep Societies.
The American Academy of Pediatrics and American Academy of Sleep Medicine recommend that children with frequent snoring be referred for a sleep study or to a sleep specialist or otolaryngologist because of the damaging consequences of living with untreated obstructive sleep apnea (OSA). But Dr. Honaker’s study of 8,135 children aged 1-12 years seen at five university-affiliated urban primary care pediatric clinics demonstrated that the identification rate of suspected OSA was abysmally low, and physician practice on this score varied enormously.
“There is a critical need to develop enhanced implementation strategies to support primary care providers’ adherence to evidence-based practice for pediatric OSA,” declared Dr. Honaker of Indiana University in Indianapolis.
She and her coinvestigators have developed a computer decision support system designed to do just that. Known as CHICA, for Child Health Improvement through Computer Automation, the system works like this: While in the clinic waiting room, the parent is handed a computer tablet and asked to complete 20 yes/no items designed to identify priority areas for the primary care provider to address during the current visit. The items are individually tailored based upon the child’s age, past medical history, and previous responses.
One item asks if the child snores three or more nights per week, a well-established indicator of increased risk for OSA. If the answer is yes, CHICA instantly sends a prompt to the child’s electronic medical record noting that the parent reports the child is a frequent snorer and this might indicate OSA.
The primary care provider can either ignore the prompt or respond during or after the visit in one of three ways: “I am concerned about OSA,” “I do not suspect OSA,” or “The parent in the examining room denies hearing frequent snoring.”
The mean age of the patients in Dr. Honaker’s study was 5.8 years, and 83% of the 8,135 children had Medicaid coverage. Parental cooperation with CHICA was high: 98.5% of parents addressed the snoring question. They reported that 28.5% of the children snored at least 3 nights per week, generating a total of 1,094 CHICA prompts to the primary care providers. Forty-four percent of providers didn’t respond to the prompt, which Dr. Honaker said is a typical rate for this sort of computerized assist intervention. Of those who did respond, only 15.9% indicated they suspected OSA. Sixty-three percent declared they didn’t suspect OSA, and the remainder said the parent in the examining room didn’t report frequent snoring.
Although responding providers indicated they suspected OSA in 15.9% of the children, that’s a low figure for frequent snorers. Moreover, 31% of children who got the CHICA frequent snoring prompt were overweight or obese, and 17% had attention-deficit hyperactivity disorder symptoms, both known risk factors for OSA. Some of the kids had both risk factors, but 39% had at least one in addition to their frequent snoring, Dr. Honaker noted.
The investigators carried out multivariate logistic regression analyses of child, provider, and clinic characteristics in search of predictors associated with physician concern that a child might have OSA. It turned out that none of the provider characteristics had any bearing on this endpoint. It didn’t matter if a physician’s specialty was pediatrics or family medicine. Providers had been in practice for 3-42 years, with a mean of 15.1 years, but years in practice weren’t associated with a physician’s rate of identification of possible pediatric OSA. Individual provider rates varied enormously, though. Some physicians never suspected OSA, others did so in nearly 50% of children flagged by the CHICA prompt.
“It’s not clear that type of training plays a large role in this area,” she commented.
The only relevant patient factor was age: children aged 1-2.5 years were 73% less likely to generate physician suspicion of OSA.
“Surprisingly, none of the patient health factors were predictive. So having ADHD symptoms or being overweight or obese did not make it more likely that a child would elicit concern for OSA,” Dr. Honaker observed.
However, which of the five clinics the child attended turned out to make a big difference. Rates of suspected OSA in children with a CHICA snoring prompt ranged from a low of 5% at one clinic to 27% at another.