Accuracy of a Prescription Claims Database: DISCUSSION

15 Nov
2010

More than 70% of the PharmaNet profiles reviewed for this study contained some inaccurate or misleading information about prescription medications currently consumed by the patients interviewed. These results suggest that sole reliance on PharmaNet profiles for medication histories could result in a high incidence of prescribing errors. Conversely, since it was relatively uncommon for a patient to be taking a medication that did not appear anywhere on the 14-month profile, PharmaNet provides an excellent starting point for medication reconciliation.

A recent review evaluating research on medication history-taking concluded that 60% to 67% of medication histories obtained during the hospital admission process contained at least one error, and that 11% to 59% of these errors were clinically important. The authors of the study suggested that prescription databases could be used to improve the accuracy of medication information obtained during the hospital admission process, and they specifically mentioned PharmaNet as an example of such a system. Not surprisingly, the cumulative evidence suggests that reviewing with patients information from a prescription database (e.g., PharmaNet) provides more accurate medication history than using either an interview or PharmaNet profiles alone. What may be surprising to some clinicians is the degree of inaccuracy of the information in PharmaNet. These results are particularly pertinent given that anecdotal reports of the medication reconciliation processes carried out in Canadian hospitals (from those participating in the Safer Healthcare Now! initiative) have revealed that sole use of PharmaNet profiles for prescription medication information is common. Even when a medication history is conducted, clinicians often streamline the process by accepting the PharmaNet profile as the list of current prescription medications and focusing only on nonprescription products during the patient interview. levitra professional

These results indicate that asking specific questions while reviewing PharmaNet data with patients is a relatively efficient approach for compiling an accurate medication history. The majority of discrepancies identified in the current study were the result of overdue refill dates, so these inconsistencies should be clarified with patients, regardless of how long overdue the prescription appears. Although the names of most active medications were usually listed somewhere on PharmaNet, at least some doses were inaccurate for approximately one-third of patients; thus, doses should be queried routinely. Furthermore, specific drug classes should be targeted, since patients more commonly deviated from the original prescription directions for medications with rapid effects on symptoms (e.g., diuretics) and those with a high incidence of adverse effects (e.g., ^-blockers). Patients should also be asked whether they have been given drug samples at the physician’s office, since this was the most common explanation for use of medications not listed in PharmaNet. Where reasons for discrepancies were reported, adverse effects were most often implicated; therefore, patients reporting adverse effects during the interview should be questioned further to determine whether the adverse effects have resulted in deviations from prescribed instructions. Use of memory aids was associated with a lower rate of discrepancies, probably through facilitation of adherence to prescribed dosing instructions. During the medication interview, patients should be asked about the use of memory aids and should be encouraged to use such aids.

top