Does a reduction in antibiotic consumption always represent a favorable outcome from an intervention program on prescribing practice?

Abstract

In our hospital, a continuous intervention program aimed at optimizing the quality of antibiotic use was introduced by late 1999 and antibiotic consumption was a major outcome for assessment. However, healthcare conditions have been subject to change over the last five years, and a pronounced economic crisis in 2002 affected the availability of antibiotics. Therefore, we hypothesized that the consumption of these drugs could be a suitable indirect marker of the crisis. We performed segmented regression analysis between different periods. Variations in antibiotic consumption during periods corresponding to the four-phase intervention program (from 1999 to the first six months of 2001) were assumed to be 'intervention-induced', while those observed during the crisis period were considered as 'situation-enforced'. Whereas the intervention-induced (desirable) decrease of total antibiotic and carbapenem consumption proved to correlate with a decreased crude mortality rate during the control period prior to the crisis (R2, 0.82 and 0.91, respectively), the crisis-induced (undesirable) decrease in total antibiotic and carbapenem consumption correlated with an increased mortality during this phase (R2, 0.80 and 0.75, respectively). Our results illustrate that a reduction in antibiotic consumption does not always represent a favorable outcome from an intervention program on prescribing practice. Moreover, it may be a sensitive indirect marker of a deficient healthcare condition leading to an increase in in-hospital mortality.