Reply to Characteristics of Pediatric Chemotherapy Medication Errors in a National Error Reporting Database

The letter by Dickens et al. regarding our article1 falls into 2 common patient safety pitfalls associated with voluntary error reporting systems by: 1) erroneously assuming that voluntary error reporting systems provide complete pictures of the scope of errors and 2) misinterpreting error reporting count data as rates and trying to compare the data with other rates.1

Our article does not state that only 310 pediatric chemotherapy errors occurred in the MEDMARX institutions between 1999 and 2004, but that only 310 errors were reported within this time frame. The actual universe of all pediatric chemotherapy errors in the MEDMARX institutions is unknown because no medical error reporting system in the U.S. is a mandatory reporting system. The voluntary nature of error reporting by definition makes the number of reports a subjective sample of the true and unknown universe of errors.

MEDMARX error count data are not valid as a rate because the data are only a subjective sample of the universe of numerator data. A valid rate should have uniformly defined and captured denominator data and numerator data. Comparing the percentage of errors reaching the patient in our sample (85% derived as the number of reported errors reaching the patient divided by the number of reported errors) with the sample provided by Dickens et al. (0.027% derived as the number of reported errors reaching the patient divided by the doses of chemotherapy dispensed) is scientifically invalid and illogical. It is the epitome of comparing apples and oranges. Practical experience and the patient safety literature both attest to the fact that the majority of medication doses given in this country are administered without error and, of the universe of errors, a fair percentage do not reach the patient.

Our aim was simply to begin to shine a light on pediatric chemotherapy errors. We know children are at much greater risk for medication errors compared with adults, that children have less resilience to medication errors, and that chemotherapy medications are some of the most dangerous medications used by the medical community.2 The pediatric oncology community needs to acknowledge that pediatric chemotherapy mistakes occur and band together to remedy them. Our patients would expect no less.

Marlene R. Miller MD, MSc*, * Quality and Safety Initiatives, Johns Hopkins Children's Center, Department of Pediatrics, Johns Hopkins University, School of Medicine, Johns Hopkins Bloomberg, School of Public Health, Department of Health Policy and Management, Baltimore, Maryland.