AHRQ Analysis of Quality Improvement Strategies for Antimicrobial Prescribing

In 2006, the Agency for Healthcare Research and Quality (AHRQ) published a report on antimicrobial prescribing as a quality improvement measure (www. ahrq.gov/clinic/tp/medigaptp.htm). This report, prepared by investigators at the Stanford-UCSF Evidence-Based Practice Center, systematically examined the effects of quality improvement strategies for reducing inappropriate outpatient antimicrobial use, and for reducing inappropriate selection of broad-spectrum agents when narrow-spectrum agents are indicated. The authors reviewed RCTs, controlled before-after studies, and interrupted time series studies on the decision to use antimicrobials. The primary outcomes were either the proportion of patients receiving any antimicrobial, or the proportion receiving a drug that was in accordance with recommended first-line therapy. Two reviewers independently abstracted data on interventions, study populations, targets and outcomes. The different quality improvement strategies were compared based on the median effect for the primary outcomes.

Fifty-four studies met the inclusion criteria, including 34 studies addressing the decision to prescribe antimicrobials and 26 studies focusing on selection of the appropriate agent (6 studies addressed both issues). The authors concluded that the methodologic quality of the included studies was generally fair. Most trials were classified as RCTs, but often failed to describe the theoretical basis for interventions. The investigators reported the following conclusions from this analysis:

1 Quality improvement strategies are moderately effective at reducing the inappropriate prescribing of antimicrobials and improving the appropriate selection of antimicrobials. The median absolute reduction in antimicrobial use was only 8.9% (interquartile range 6.7-12.4%) in the reviewed studies. For studies targeting selection of the appropriate agent, the median absolute reduction was 10.6% (interquartile range 3.4-18.2%).

2 Although no single quality improvement strategy is clearly superior, active clinician education may be more effective in certain settings. There was no single intervention or group of interventions that was highly effective. Active educational interventions appeared to be more effective in studies focusing on the decision to prescribe antimicrobials, but the difference was not significant. Surprisingly, clinician education alone appeared to be less effective than clinician education plus audit and feedback of prescribing behavior for those studies that focused on selection of the appropriate drug.

3 Interventions targeting prescribing for all acute respiratory tract infections may exert a greater effect on overall prescribing than interventions targeting specific types of acute respiratory infections. The authors extrapolated antimicrobial prescribing data to the population level when possible for each study, and they found that interventions focused on all ARIs, rather than specific diagnoses, had the greatest potential impact on antimicrobial use. Interventions focused on particular diagnoses (such as sinusitis or pharyngitis) tended to have greater effect sizes at the individual level, but the population-level effects were more modest.

4 Study design and quality should be improved. Studies that formally evaluate the cost effectiveness of interventions to improve antimicrobial treatment and selection are needed, and studies should evaluate the potential harm of such interventions. A substantial number of studies suffered from methodologic limitations, such as lack of randomization and failure to document whether the educational interventions were received by the participants. Multifaceted intervention studies generally require analysis at the level of the clinic, population, or geographic area (such as a city or county). Randomized studies at the clinic level are feasible and have proven useful for evaluation of antimicrobial prescribing interventions. However, randomization is often impossible at the level of communities, counties, or other large geographic areas due to cost considerations and limited number of units available for allocation. In addition, delivery of interventions is more complex and difficult to measure in larger populations. As a result, non-randomized studies have predominated despite the methodologic superiority of group randomized trials.

0 0

Post a comment