974. External Validation of Estimates of Antibacterial Dispensing in the IMS[jc1]  Brogan Xponent® Database in a Canadian Province
Session: Poster Abstract Session: Stewardship: Epidemiology of Antibiotic Use
Friday, October 4, 2013
Room: The Moscone Center: Poster Hall C
Posters
  • Dalton. External Validation of Estimates of Antibacterial Dispensing in the IMS Brogan Xponent® Database in a Canadian Province.pdf (348.1 kB)
  • Background: IMS Brogan (IMS) monitors ambulatory antibacterial (AB) consumption using the Xponent®database, which samples ~60-70% of retail pharmacy dispensing records and estimates total use using a proprietary method.  Many studies use IMS data for monitoring the effect of antimicrobial stewardship (AS) activities and identifying potetntial AS targets. Since 09-2007 Alberta pharmacies have submitted dispensing records to a provincial database; the Pharmacy Information Network (PIN).  Our objective was to validate the IMS Brogan estimates for the province of Alberta using data from the PIN.

    Methods: Guidelines for reporting reliability and agreement studies (GRRAS) were followed. We obtained dispensing data from IMS and PIN for systemic AB [ATC J01] for 8 consecutive quarters from 07-2010 to 06-2012. Number of scripts were provided for all systemic ABs in 5 provincial health zones grouped by age (0-17, 18-64 and ≥65).  Population estimates for each zone and each age group as of June 30 (mid-year) of 2010, 2011 and 2012 were derived from the Alberta Health Care Insurance Plan and were used to calculate the annual rate of prescriptions per 1000 inhabitant-day (PrID) using data from IMS and PIN.   Agreement between measures was assessed using the Bland-Altman method and the St Laurent gold standard correlation coefficient.

    Results: Annual rates of total antibacterial use from PIN were 1.49-1.64 PrID compared to IMS estimated rates of 1.59-1.79 PrID, an overestimation of 6.4-11.3%.  Analysis by class of AB at the zone/age group/year level by Bland-Altman plots found  a mean difference (PIN-IMS) of -0.0118 PrID and 95% limits of agreement of -0.12 to 0.097.  St Laurent correlation was 0.68 (95% CI 0.49-0.86) for total (all AB) PrID at the zone level for each year.

    Conclusion: We found moderate-good agreement between PrID for AB derived from the IMS vs PIN databases, dependent on the level of analysis (eg. total AB vs AB use by class). Our investigation, which is the one of the first to assess the validity of the IMS database suggests it is generally reliable with some potential for overestimation, which may not be relevant to AS initiatives.  Our conclusions would be strengthened by replication using similar databases from other jurisdictions.

    Bruce Dalton, PharmD1, Lawrence Svenson2, Lauren Bresee, BScPharm, ACPR, MSc, PhD1, Deana Sabuda, BSP1, Bayan Missaghi, MD3,4, Oscar E. Larios, MD5, Kimberley Simmonds, MSc6 and John Conly, MD3, (1)Pharmacy, Alberta Health Services, Calgary Zone, Calgary, AB, Canada, (2)School of Public Health, University of Alberta, Edmonton, AB, Canada, (3)University of Calgary, Calgary, AB, Canada, (4)Alberta Health Services, Calgary Zone, Calgary, AB, Canada, (5)Division of Infectious Diseases, University of Calgary, Calgary, AB, Canada, (6)Alberta Health, Edmonton, AB, Canada

    Disclosures:

    B. Dalton, None

    L. Svenson, None

    L. Bresee, None

    D. Sabuda, None

    B. Missaghi, None

    O. E. Larios, None

    K. Simmonds, None

    J. Conly, None

    Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.