A new method for identifying antibiotic-treated infections using automated pharmacy records

Area of Science:

• Infectious Disease Epidemiology
• Pharmacoepidemiology

Background:

• Ascertaining community infections for research is often limited by the high cost of surveillance.
• Existing methods for infection detection in large populations can be resource-intensive.

Purpose of the Study:

• To evaluate the utility of antibiotic prescription fills in automated pharmacy records for identifying treated infections in postmenopausal women.
• To compare the accuracy of a pharmacy-based method with traditional outpatient medical record reviews for infection ascertainment.

Main Methods:

• Utilized automated pharmacy records from Group Health Cooperative of Puget Sound.
• Excluded antimicrobial prophylaxis and chronic antibiotic use.
• Tested four intervals (30, 45, 60, 90 days) between antibiotic fills to detect new infections.
• Validated findings against outpatient medical record reviews in 150 women.

Main Results:

• The pharmacy method demonstrated high sensitivity (80-88%) in detecting new infections.
• 75% of women with no infection in chart reviews were also identified as infection-free by the pharmacy method.
• The 60-day interval showed the strongest agreement (kappa = 0.55) with chart reviews for infection counts over 2 years.

Conclusions:

• Automated pharmacy records provide a feasible and cost-effective approach for infection ascertainment in epidemiologic research.
• This method offers a valuable alternative to traditional surveillance, particularly for studying infections in community-dwelling populations.
• The pharmacy-based approach shows good concordance with medical record reviews, supporting its use in research settings.