Background: Pharmacy practice guidelines promote the role of community pharmacies in self-monitoring of blood glucose. However, variations between Dutch pharmacies exist in the proportion of patients to whom test strips are dispensed. Objective: Toassess whether variations between community pharmacies in dispensing of blood glucose test strips can be explained by differences in patient characteristics and the region in which the pharmacy is located. Setting: PHARMO-Record Linkage System containingdrug dispensing histories from 40 community pharmacies of about 450,000 patients in the Netherlands. Method: We performed a retrospective cohort study. Data on prescription of all drugs and medical aids between 1991 and 2001 were extracted for all new users of antidiabetic drugs. Patients were excluded if the dispensing history did not allow classification of the type of diabetes or if the dispensing pharmacy could not be determined. The data were analysed using a Cox proportional hazard model. Main outcome measure: Time to first test strips dispensed. Results: We identified 8,233 starters of antidiabetic drugs. During a median follow-up of 2.1 years, 20% of the patients were dispensed test strips at least once. Community pharmacy was significantly associated with the dispensing of test strips after adjustment for patient characteristics. This association was less apparent when stratified for the geographical location of the pharmacy. Conclusion: Community pharmacy is an independent determinant of the start of use of test strips. Differences in dispensing of test strips between pharmacies are dependent on geographical region. This suggests that implementing practice guidelines for diabetes care in community pharmacy requires different approaches indifferent regions.