Screening Tool Identified Abuse Risk in Kids With Just One Bruise

A screening tool was effective in identifying abuse among young children with a single bruise in a pediatric emergency department, a secondary analysis of a validation study showed.

Of 349 kids with a single bruise, the TEN-4-FACESp bruising clinical decision rule (BCDR) was positive for 22 of 27 cases classified as abuse and 40 of 322 classified as an accident, thus performing with 81.5% sensitivity and 87.6% specificity, reported Mary Clyde Pierce, MD, of the Ann & Robert H. Lurie Children’s Hospital of Chicago, and colleagues in Pediatrics.

“Since mobile children often sustain bruises during everyday play and accidental events, the ability to differentiate abusive from accidental bruising is of high importance to prevent unnecessary investigations,” the authors wrote.

“Negative BCDR results must be interpreted with caution given the higher rate of false negatives in this analysis compared with the validation study,” they added.

In a previous validation study, the TEN-4-FACESp BCDR — a validated screening tool using information about a child’s bruised body region, age, and pattern of bruising to predict abuse in children younger than 4 years — was found to identify abuse with 95.6% sensitivity and 87.1% specificity, Pierce and team noted.

In a commentary accompanying the study, Tagrid Ruiz-Maldonado, MD, MS, of the University of Utah and Primary Children’s Hospital in Salt Lake City, and Suzanne Haney, MD, MS, of the University of Nebraska Medical Center and Children’s Nebraska in Omaha, wrote that the TEN-4-FACESp BCDR is an “effective risk-prediction tool that has revolutionized medical providers’ approach and understanding of seemingly minor injuries in childhood,” but it is “not a diagnostic tool.”

“Child abuse pediatricians are pivotal in providing input that clinical decision support tools are not built to provide, and we encourage providers to engage in conversation and establish relationships with child abuse pediatricians in their areas,” they added.

Pierce told MedPage Today in an email that the secondary analysis was “driven by one of the most common questions we get from clinicians,” which is, “what if there is only one bruise present?” This makes them question whether there is still concern about abuse, she said.

“We actually expected that if there was just one bruise in the … predictive readings for abuse, that [the tool] would not work as well — as in it would be less specific,” Pierce said. “We thought it would be less specific, but it turns out it was actually (slightly) more specific.”

However, it was “less sensitive for picking up abuse,” meaning “abuse is more likely to be missed when there is only one bruise,” she added.

In this analysis, Pierce and colleagues identified five key differences between abusive and accidental injuries.

Children with abusive injuries were younger (median age 1.1 vs 1.8 years, P=0.005), and less likely to present with an injury complaint (18.5% vs 39.8%, P<0.001).

Additionally, they were more likely to have a bruise in a BCDR-positive region. Among patients with abusive injuries, the most common single bruise region was the ear (n=6), followed by the orbital rim (n=4), and eyelid (n=4), while the most common region among patients with accidental injuries was the lower leg (n=113), followed by the forehead (n=76), and knee (n=33).

Furthermore, children with abusive injuries had a lower Glasgow Coma Score (GCS), with six kids presenting with a GCS less than the highest score of 15 (which indicates an individual is fully awake and alert); five cases were classified as abuse, and one was classified as an accident.

Finally, compared with children with abusive injuries, those with accidental injuries were more likely to have no psychosocial risk factors, including prior Child Protective Services involvement, domestic violence, or mental health concerns, among others (72.4% vs 29.6%, P<0.001). More than a third of kids with abusive injuries had two or more psychosocial risk factors compared with 11.2% of those with accidental injuries.

Data from the study “argue for a careful and comprehensive assessment” of the child and injuries, as well as the environment and family situation, Howard Dubowitz, MD, MS, of the University of Maryland School of Medicine in Baltimore, told MedPage Today.

“This is a valuable illustration of all that could and should be considered in making what is a very difficult but important determination as to the likelihood that a child has been abused,” said Dubowitz, who was not involved in the study.

In other findings, three of the five children who had cases misclassified as abuse by the tool were older than 1 year. Two presented with medical chief complaints and no history of trauma, and one was referred for sibling abuse evaluation. All three had long bone fractures.

Of the two patients younger than 1 year, one presented with a medical chief complaint but no known trauma history, and the other presented with neurological symptoms after a reported fall. Both were found to have cranial and/or intracranial injuries, and one case was fatal.

Single bruise locations in the five misclassified cases were the forehead (n=2), upper arm (n=1), finger (n=1), and knee (n=1).

For this secondary analysis, the researchers included 349 patients from the BCDR validation study whose only skin finding was a single bruise (including petechiae, subconjunctival hemorrhage, or frenulum injury). Cases were previously classified as abuse, accident, or indeterminate by an expert panel. Mean age was 1.8 years, and 56% were boys.

Because the participants were enrolled from a pediatric emergency department setting, an “enriched sampling of patients with potential abuse could contribute to a higher prevalence of bruising than would be found in primary care clinics or other clinical environments,” Pierce and colleagues noted.

Please enable JavaScript to view the

comments powered by Disqus.