Abstract
Introduction: Pediatric inpatient populations, including neonates, are at increased risk for clinically significant and/or potentially life-threatening medical errors. Computerized physician order entry (CPOE) holds the promise of reducing medication errors. Special attention must be paid to the match between the specialized needs of the population and the detailed capabilities provided by a CPOE system. As we transitioned from a system designed and developed by a team of pediatric/neonatal sub-specialists for the Children's National Medical Center (CNMC) NICU (TPN 2000) to a more general and adult medicine-based CPOE system (Abacus system, Baxa Corporation, CO), we compared the number and severity of medical errors between the two systems.
Method: In Phase I, we reviewed two hundred randomly selected daily TPN orders using the TPN 2000 system, for the number and severity of errors. Specific variables included calcium to phosphorus ratio, osmolality, total calories and dextrose concentration in relation to types of vascular access. Parameters of expected normality were established. Errors were identified and categorized into clinically significant, serious and potentially life-threatening. After 6 months of training on the Abacus system, Phase II began with a review of an additional 200 randomly selected TPN orders from the new system.
Results: The most common error was the calcium to phosphorus ratio which accounted for a 4% increase in the number of potentially life-threatening errors (11% errors in TPN 2000 vs. 18% errors in Abacus). Error rates for osmolality were 8.5% during Phase I vs. 0% in phase II. Total calories and dextrose rate errors were not statistically different. Osmolality and dextrose concentration in relation to access showed a 100% improvement in errors in both categories with Abacus vs. the TPN 2000 system.
Conclusion: CPOE has the potential to reduce medical errors in the neonatal population. However, a 4% increase in potentially life-threatening errors in calcium to phosphorus ratio seen in our study may support our contention that the sub specialist physicians' input is absolutely essential to the Health Informaticians when designing a new CPOE, modifying an existing one or evaluating and certifying a commercially available or custom developed CPOE system.