Surveillance for Adverse Drug Events in Ambulatory Pediatrics

In this project, the team set out to build and evaluate an automated system to identify adverse drug events (ADEs) in the ambulatory setting for pediatric patients diagnosed with cancer, sickle cell anemia, and cystic fibrosis.

The main objectives of the project were to:

• Implement an automated surveillance system for measuring the incidence of ADEs occurring in the outpatient setting (including the emergency department) in pediatric patients with specific chronic diseases that result in the need for emergency department care or admission to the St. Louis Children’s Hospital.
• Use the automated surveillance system for measuring the incidence of ADEs occurring in these patient populations during transitions in care from outpatient to inpatient settings, e.g., originating during the admission process.
• Use the automated surveillance system to measure the incidence of ADEs in the target pediatric populations within four weeks of discharge.
• Evaluate the performance of the event detection system as employed in the three above aims.

The project team successfully developed and implemented an automated system for measuring the rate of ADEs in this patient population. The team compared the rate of ADE detection with the automated system to a traditional manual chart review. In 1,983 patients the automated system detected 156 ADEs with a positive predictive value (PPV) of 21.5 percent. A random review of 392 patient charts was completed by a pharmacist. The sensitivity of the automated system was 43 percent with a PPV of 16 percent; while the sensitivity of the chart review was 86 percent with a PPV of 59 percent. A total of 66 hours of pharmacist time was needed for the automated method, versus 565 hours for the chart review.

A review of the ADEs missed by the automated system revealed that nearly half of the ADEs from the manual review were based on evidence in documentation that was not available to the automated system, while the other half involved rashes which did not have rules for detection in the automated system. Of note, in comparing the manual to the automated system, there was no difference in the level of harm caused by the ADEs. While the manual review by the pharmacist outperformed the automated system, this method is highly resource-intensive and took almost 8 times longer to perform than the automated system. Both manual chart reviews and automated systems are better at identifying ADEs than voluntary reporting. Therefore, in an environment in which manual chart reviews are not feasible, an automated system is a reasonable alternative, being both feasible and efficient in detecting ADEs in this patient population.