We found the reliability of coding drug claims on the ODB database to be very high, with errors occurring in only 0.7% (95% CI 0.5% to 0.9%) of over 5000 dispensed medications that we examined. Of the 37 errors, 13 resulted in the subject being dispensed something other than what was prescribed; nine had errors in the instructions to the patient. These 22 errors could have had an impact on the patient. The remaining 15 errors were clerical or had the incorrect physician listed, which would only have an impact on investigators using the data for research purposes. flovent inhaler
At the outset, we hypothesized that the time spent dispensing each medication may affect the coding accuracy. Pharmacists and assistants in high productivity pharmacies have less time to dispense a medication, and that may result in a higher likelihood of a coding error. Thus, one of the original objectives was to evaluate whether the reliability of coding differs according to a pharmacy’s productivity, defined as the annual number of prescriptions per pharmacy divided by the annual cumulative total hours of work done by pharmacists and assistants. However, given the low rate of errors (on average less than one error per pharmacy), it was not possible to detect any such association.