A simple blood test may help young infants with fever avoid unnecessary testing and treatment, and may help emergency clinicians spot serious bacterial infections (SBIs) like urinary tract and bloodstream infections, finds a new study. Led by Richard Bachur, MD, of Children's Hospital Boston's Division of Emergency Medicine, in collaboration with George Washington University, the study shows that a diagnostic marker called procalcitonin can identify infants at high risk for SBIs, while helping rule out these infections in the rest.
Fevers commonly bring infants to emergency rooms, where physical examinations often cannot identify the cause of the fever. While most fevers turn out to be minor and self-limiting, a few infants who do get them harbor serious infections. The risk is most significant in infants under 3 months of age. Many of these babies currently undergo extensive evaluations that often include blood and urine tests, lumbar puncture, hospitalization and a course of antibiotics when they present with a fever of unknown origin.
When Bachur and his colleagues used the procalcitonin test in 234 infants with fever, all under the age of 3 months, it not only detected most SBIs, but reliably indicated when infants were at low risk. In fact, this single test, which is now commercially available, performed nearly as well as the current strategies involving numerous clinical and lab tests. Bachur hopes it will eventually become a standard tool.
The study, supported by the Frederick Lovejoy, Jr., MD, Resident Research Fund and the American Academy of Pediatrics Resident Research Grant, was published in the October issue of Pediatrics.
More than 5,000 children were separated from their families during Hurricane Katrina, some for up to 18 months. To avoid such circumstances in future disaster situations, Michael Shannon, MD, MPH, and Sarita Chung, MD, both of Emergency Medicine, and Stephen Monteiro, MS, EMT-P, Children's emergency management coordinator, are creating Reunite, an automated system to help parents identify missing children. Partnering with the University of California-Berkeley and the International Computer Science Institute, they recently won a three-year grant from the government's Emergency Medical Services for Children program.
Children, especially young children, present special challenges during natural disasters or terrorist attacks, yet they've been overlooked in disaster planning. "Even before Katrina, we realized the challenges of identifying kids who can't identify themselves," says Chung. In large-scale disasters, the challenges are heightened. "In pandemic flu modeling, we expect entire families to become ill," says Monteiro. "There may be a sole surviving child who must be identified by distant relatives."
Methods like fingerprinting and retinal scanning may not work well in infants and toddlers. "In children under 5, that would probably be impossible," says Chung. "In our disaster drills, we've learned that the best mechanism is photographs."
The idea is simple: Take a digital photo of every child who comes into emergency departments (EDs) during a disaster and, through computer software, automatically catalog the child's physical features: eye, skin and hair color, approximate age and clothing. Instead of having to search every ED in a city for a child, families could go to one location, give a description and view a database of children meeting that profile. Images could be tweaked to eliminate distressing lacerations and bruises; faces could even be reconstructed.
But the plan has its challenges. Current face recognition methods perform surprisingly poorly, especially when lighting is bad and the children being photographed are screaming or have their eyes closed. "We need an algorithm that's able to deal with different field conditions," says Chung.
Reunite could potentially help families find not just children, but anyone who cannot identify themselves due to severe injury, disorientation or dementia, Chung adds. The team hopes to have a prototype for testing by next September.