151 Routine data registries as a basis to analyse and improve the quality of antimicrobial prescription in Primary Care 6 Discussion A primary goal of this study was to combine and to use two large registries to identify and determine the number of antimicrobial prescriptions in primary care and the determinants of appropriateness in prescription. Antimicrobial prescriptions were subsequently defined as appropriate or inappropriate following guidelines, and linked with potential determinants of appropriateness. By combining data from two large registries (ELAN and SN) at an individual patient level, we were able to explore associations of several determinants with appropriateness that are not registered in an EMR. Our principal findings were: 1) the highest rate of antimicrobial overprescribing, in both number and proportion, was for RTIs, 2) most prescriptions of macrolides did not correspond with the 1st and 2nd choice in guidelines, and 3) determinants including female gender, age 5 years and older, migration background (Turkish, Surinamese, Dutch-Caribbean), and a large primary care practice size were all associated with antimicrobial overprescribing. Large registries A major strength of our study was that we were able to identify potential determinants of antimicrobial prescription in the context of the patient by combining routine healthcare data with individual socioeconomic - and context data from SN. The use of routine healthcare data for medical research has many advantages, as it provides relatively easy access to rich, ecologically valid, longitudinal data from large populations (67). In other words, it potentially more accurately reflects daily practice in accordance with our aim of understanding patterns of daily antimicrobial prescribing in primary care (17). Combining primary care EMR data with data from SN allowed us to explore novel associations such as migration background, household income and number of parents per household, data that are not routinely recorded in an EMR. A potential downside of routinely collected healthcare data is the risk of missing data. The data were not systemically recorded for research but for healthcare purposes, for which data are recorded only when relevant for the treatment of patients in the eyes of the provider or practice staff. ICPC codes for antimicrobial prescriptions were sometimes missing or a registered ICPC code was not related to the infection. We were also unable to verify registered diagnoses in this large dataset, which may have led to registration bias, with either under- and over-registration. To better gauge this risk, we compared our study with two prospective Dutch studies on appropriateness of antimicrobial prescribing for RTIs, as prospective data collection is less prone to incorrectly registered or missing data. Both studies had a comparable proportion,
RkJQdWJsaXNoZXIy MTk4NDMw