Abstract: Comparing prescribing v. dispensing data can gain a more comprehensive insight into antibiotic use. The study performed linkage matching on variables including patient ID, time of record, and antibiotic classification. Antibiotics exposures were defined using script and episode approaches. Using patient ID, antibiotics spectrum, class and sub-classification, 7-lag days and the script approach, 91.2% of prescribing (N=257,779) and 95.1% of dispensing (N=238,526) from IDS sites and 61.4% of prescribing (N=161,305) and 78.2% of dispensing (N=122,770) from non-IDS sites can be linked.
Learning Objective 1: To learn about methods used to link dipsensing with prescribing data in the PCORnet Antibiotics and Childhood Growth Study
Learning Objective 2 (Optional): To learn about the difference in linkage rate between broad and narrow spectrum antibiotics among the pediatric population.
Learning Objective 3 (Optional): To learn about clinical and demographic characteristics affecting linkage rate in the PCORnet Antibiotics and Childhood Growth Study
Pi-i Lin (Presenter)
Harvard Pilgrim Health Care Institute
Sheryl Rifas-Shiman, Harvard Pilgrim Health Care Institute
Charles Bailey, Children’s Hospital of Philadelphia
Janne Boone-Heinonen, Oregon Health & Science University
Matthew Daley, Kaiser Permanente Colorado
Christopher Forrest, Children’s Hospital of Philadelphia
Casie Horgan, Harvard Pilgrim Health Care Institute
Darren Toh, Harvard Pilgrim Health Care Institute
Jessica Young, Harvard Pilgrim Health Care Institute
Jason Block, Harvard Pilgrim Health Care Institute