The aim of this study was to measure the validity of the prescriber information recorded in the Danish National Prescription Registry (DNPR). The prescriber information recorded in the pharmacies' electronic dispensing system was considered to represent the prescriber information recorded in the DNPR. Further, the problem of validity of the prescriber information pertains only to non-electronic prescriptions, as these are manually entered into the dispensing system. The recorded prescriber information was thus validated against information from a total of 2000 non-electronic prescriptions at five Danish community pharmacies. The validity of the recorded prescriber information was measured at the level of the individual prescriber and the prescriber type, respectively. The proportion of non-electronic prescriptions with incorrect registrations was 22.4% (95% confidence interval (CI): 20.6-24.3) when considering individual prescriber identifiers and 17.8% (95% CI: 16.1-19.5) when considering prescriber type. When excluding prescriptions specifically registered as 'missing prescriber identifier', the proportions decreased to 9.5% (95% CI: 8.2-11.0) and 4.1% (95% CI: 3.2-5.1), respectively. The positive predictive values for the classification of prescriber types were in the range of 94.0-99.2%, while the sensitivity ranged between 64.6% and 91.8%. With a maximum of 14% non-electronic prescriptions of all prescriptions in the DNPR in 2015, this corresponds to correct classification of prescriber types in the DNPR of at least 97.5%. In conclusion, the prescriber information in the DNPR was found to be valid, especially in recent years. Researchers should be aware of the low sensitivity towards prescriptions from private practicing specialists.
© 2016 Nordic Association for the Publication of BCPT (former Nordic Pharmacological Society).