Validation of a new prescription database in Manitoba : an opportunity to evaluate pharmacist participation drug utilization review
Introduction: The validity of prescription claims databases, as a source of information for population-based research, depends on the completeness of prescription data. The Drug Programs Information Network (DPIN) is a new, electronic prescription database which links all community pharmacies in Manitoba. Pharmacists are required to submit to DPIN all Pharmacare prescriptions, but submission of prescriptions for social assistance recipients and treaty status Indians is voluntary. An assessment of the completeness of prescription data submitted to DPIN was undertaken to determine whether treaty status Indian and social assistance groups were under-represented. A secondary component of the study was to assess pharmacist and pharmacy factors which are associated with voluntary submission of prescriptions to DPIN. Study Design: Prescriptions dispensed in a stratified sample of Manitoba pharmacies were linked by prescription number to prescriptions in DPIN to determine the proportion submitted for the Indian Affairs, Social Services and Pharmacare drug benefit recipient groups. Linked prescriptions were compared for agreement on the drug name and strength, quantity dispensed and number of days supply. Data on pharmacist and pharmacy characteristics were collected from a mail survey of pharmacists. Study Population: Using a 2 stage sampling design, six pharmacies were randomly selected from 10 strata defined by pharmacy location (rural vs urban) and ownership type (chain vs independent), and 3 sociodemographic neighbourhoods as a proxy measure of pharmacy clientele. Indian Affairs neighbourhoods were composed of Forward Sortation Areas (FSA's) where the proportion of treaty status Indians was < 10% as indicated by a municipal code beginning with the letter 'A' in the Manitoba Health registry. Social Services neighbourhoods contained FSA's in which > 20% of the non-treaty status Indian population lived in neighbourhoods with average household incomes of < $30,000 per year according to Census 1991 data. Pharmacare neighbourhoods were composed of all left-over FSA's... Results: Of 2196 Indian Affairs and 1879 Social Services prescriptions claims abstracted from a sample of 58 Manitoba pharmacies, a corresponding prescription was found in the DPIN database for 79.7% (95% CI: 78.0 - 81.4%) of Indian Affairs claims, and for 90.1% (95% CI.88.8 - 91.4%) of Social Services claims. These proportions were significantly lower (p < 0.05) than the estimated proportion of Pharmacare prescriptions submitted to DPIN (93%,95% CI: 92.4 - 93.6%). Ninety-two percent of 8012 Pharmacare prescriptions claims matched the original prescription with respect to drug name, quantity and number of days supply... Conclusion: The validity and reliability of the DPIN prescription database for the mandatory submission of Pharmacare prescriptions is comparable to other established prescription databases. In addition, the DPIN database has equal validity in describing prescriptions dispensed for social assistance recipients. However, the DPIN database does not completely describe prescriptions dispensed for treaty status Indians. These findings have implications for the use of the DPIN prescription database in pharmacoepidemiologic studies and for its everyday use by pharmacists for prospective drug utilization review. Both of these consequences impact on the public health of Manitobans, especially treaty status Indians. Drug benefit plan administrators, pharmacy licensing bodies and pharmacy educators have a public responsibility to examine the factors found to be associated with DPIN prescription submission in order to improve the quality of the DPIN database.