1003 — Clinical Decision Support at the Time of an e-Prescription Can Sustainably Decrease Unwarranted Use of Antibiotics for Acute Respiratory Infections
DeLisle S (VA Maryland Health Care System and University of Maryland), Rattinger GB
(VA Maryland Health Care System and University of Maryland), Mullins CD
(University of Maryland), Onukwugha E
(University of Maryland), Zuckerman IH
(University of Maryland), Walker LD
(University of Maryland), South B
(VA Salt Lake City Health Care System and University of Utah), Samore M
(VA Salt Lake City Health Care System and University of Utah), Gundlapalli A
(VA Salt Lake City Health Care System and University of Utah)
Because exposure to antibiotics (ABX) promotes the development of bacterial resistance, these agents should not be used to treat conditions for which they are ineffective, including many uncomplicated acute respiratory infections (ARI). We asked if interposing clinical decision support (CDS) software at the time of electronic order entry could adjust ABX utilization toward consensus guidelines for ARI.
The CDS was introduced in 1/2003 at the Maryland VA site (Intervention) but not at the Utah VA site (Control). The CDS targeted two ABX, gatifloxacin and azithromycin (Targeted ABX). Access to all other ABX (Other ABX) was unrestricted. To derive an enriched, population-based sample, we applied a previously validated ARI surveillance algorithm [(one of 197 "respiratory" ICD-9 diagnostic code OR New Cough Remedies OR Temp >= 38ºC) AND (text analysis of the clinical note for non-negated ARI symptoms)] to EMR entries surrounding all outpatient visits (n = 4.1 million) during our study period (1/2002-12/2006). Flagged records (n = 7000) were then manually abstracted on the day of each index visit for all information required to establish guideline-defined ARI diagnoses and ABX treatment.
3,831 unique patients satisfied the consensus case definitions for pneumonias (537), bronchitis (2,931), sinusitis (717), and non-specific acute respiratory infections (145). For the Targeted ABX, the proportion of unwarranted prescriptions for ARI decreased from 22% to 3%, pre (2002) vs. post-intervention period (2003-2006, p < 0.0001). This proportion did not change for the Other ABX at the Intervention site (30% pre vs. 31% post) or for both the Targeted (16% vs. 20%) and the Other ABX (22% vs. 27%) at the Control site. The CDS did not impede utilization of antibiotics when they were indicated, including all cases with pneumonia. Azithromycin and gatifloxacin remained the antibiotics most often prescribed when indicated for ARI.
CDS interposed during the e-prescription process nearly extinguished unwarranted use of targeted ABX for the treatment of ARI for four years, and did so without reducing indicated use or shunting misutilization toward unrestricted agents.
CDS interposition at the time of e-prescription represents an effective, scalable approach to promote evidence-based drug utilization.