Measuring Multiple Dimensions of Cumulative Antibiotic Exposure: The Antibiotic Exposure Index

Using a single measure to characterize longitudinal antibiotic exposure (AE) could underestimate variability in antibiotic use and limit understanding of the impact on clinical and microbiologic outcomes. We propose the use of the antibiotic exposure index (AEI), which combines multiple dimensions of longitudinal AE, and demonstrate its potential applicability in two patient populations.

Study investigators had regular meetings to suggest, revise and finalize the AEI, which combined multiple dimensions of AE. The AEI was conceptualized to incorporate multiple characteristics of cumulative antibiotic exposure while retaining the ability to distinguish patients with different durations of therapy. Three dimensions of AE were selected for inclusion in AEI based on review of existing literature: (i) duration (d), the cumulative total days of therapy with a specific antibiotic, (ii) continuity (c), the number of courses in which a particular antibiotic was given, and (iii) regularity (r), the standard deviation of antibiotic-free intervals measured in days. To satisfy a priori criteria, AEI was defined as d + (1 + log_e(x/1 + x)), where x = c + r. The use of AEI was demonstrated using piperacillin/tazobactam (P/T) treatment data for 2013–2014 from two patient populations in different hospitals.

There were 12,683 and 10,839 unique patients who received P/T during the study period (populations A&B). Using cumulative duration of AE alone as a classifier, patients in population A were categorized into 78 distinct groups (range 1–192 days) and patients in population B were categorized into 66 groups (range 1–106 days). Using the number of antibiotic courses that included P/T, population A was categorized into 18 groups (range 1–22 courses) and population B into 14 groups (range 1–19 courses). The AEI was significantly associated with all three dimensions: duration, days of therapy, and regularity (P < 0.01). AEI calculations increased the number of exposure groups to 1307 in population A and 922 in population B (Figure 1).

The AEI expanded our ability to simultaneously capture multiple dimensions of longitudinal antibiotic exposure. Further work will link this measure to patient outcomes.

Open in new tabDownload slide

All authors: No reported disclosures.