Effect of Utilization Policies
All of the 9 surveys that were mailed (one to each DHA) were returned, for a 100% response rate. The types of policies in place and how they were enforced and audited varied widely among the 9 DHAs (Table 2). Combining the survey results with the fluoroquinolone utilization data provided additional insights into drug utilization policies in Nova Scotia. Student ttests comparing drug utilization rates between hospitals with no utilization policies and those with more than one policy revealed no statistically significant differences in the use of ciprofloxacin (p = 0.20) or levofloxacin (p = 0.38).
The relation between the number of antimicrobial policies and fluoroquinolone use was explored for ciprofloxacin and respiratory fluoroquinolones (expressed in terms of both DDDs per 1000 bed-days per year and DDDs per 100 cases of community-acquired pneumonia with hospital admisison per year). Multilevel analysis revealed that fluoroquinolone use did not differ significantly between districts with more than one utilization policy and those with no policies. viagra 50 mg
Multilevel modelling analyses were then performed to estimate the variance of fluoroquinolone use across DHAs and the hospital-DHA interaction. This methodology was used to account for the hierarchical structure of the data set and to explore the variance (dispersion among measures) of 2 drug utilization variables, each represented by a different denominator (fluoroquinolone DDDs by hospital bed-days and by cases of community-acquired pneumonia with hospital admission). This analysis demonstrated that, except in the case of ciprofloxacin, the variability in fluoroquinolone utilization rates could not be explained simply by the DHA to which the hospital belonged. For ciprofloxacin use (expressed as DDDs per 1000 bed-days per year), the DHA did represent a small component of the variability. In other words, much of the variation in fluoroquinolone use occurred at the hospital level rather than the district level, and variation in drug utilization rates was not attributable to a clustering effect of hospitals. This also suggests that adherence to DHA pharmacy policies might have varied among hospitals within a given DHA.