Prevalence and clinical significance of medication discrepancies at pediatric hospital admission.

Objective: To quantify admission medication discrepancies in a tertiary-care, general pediatric population, to describe their clinical importance and associated factors, and to assess a screening approach to pharmacist involvement.

Methods: A total of 272 patients were studied prospectively at hospital admission. The study pharmacist performed a medication history and compared it to physicians' admission medication orders. Discrepancies between the 2 were coded as intentional but undocumented or unintentional. Unintentional discrepancies were rated for potential to cause harm by 3 physicians. Additional data collected included patients' reason for admission and presence of chronic conditions, whether physicians used a medication reconciliation form, the characteristics of patients' home medication regimen, and the time required to perform a pharmacist history and reconciliation. Interrater reliability and associations between baseline characteristics and discrepancy rates were explored.

Results: Eighty patients (30%) had at least one undocumented intentional discrepancy (range, 0-7). At least one unintentional discrepancy (range, 0-9) was found in 59 patients (22%). Of the unintentional discrepancies, 23% had moderate and 6% had severe potential to cause discomfort or deterioration. Ratings were similar among the 3 physicians. Characteristics associated with higher risk of clinically important discrepancies were: use of the medication reconciliation form, > or =4 prescription medications, and antiepileptic drug use. Logistic regression revealed that only the variable > or =4 medications was independently associated with clinically important discrepancies.

Conclusions: Admission medication errors are common in this tertiary-care, general pediatric population, and nearly a third represent potential adverse events. The use of a medication reconciliation form by physicians without pharmacist involvement does not appear to reduce errors. A cutoff of > or =4 prescription medications is highly sensitive for identifying patients at risk of clinically important discrepancies.