Clinical and economic outcomes of pharmaceutical services related to antibiotic use: a literature review

Abstract

Aim of the review: To identify and review the clinical and economic impact of pharmacists' interventions on antibiotic use.

Method: A literature search was conducted on Medline (1966-2003) to identify original articles measuring the impact of pharmacists' interventions on antimicrobial therapy at patient's or prescriber's level.

Results: Forty-three articles were included: 20 uncontrolled before-after studies, four controlled before-after studies, five controlled trials, 12 randomized controlled trials and two interrupted time series (ITS). The described interventions were grouped into four categories: patient-specific recommendations (pharmacists' interventions concerning patient-specific drug therapy), implementation of policies, education, and therapeutic drug monitoring. These interventions were often combined to provide a multifaceted intervention, making it difficult to isolate the impact of one specific intervention. Measured outcomes were: appropriateness of prescribing (evaluated in 17 studies, 16 showing significant improvement), costs (analysed in 22 studies, nine showing a statistically significant reduction in costs after or with the intervention), and length of hospital stay (mixed results). Other measured outcomes were: drug use, prescriptions, length of treatment, dose intervals, switch to oral route, mortality rate, and treatment failure.

Conclusion: Over the years, the number of studies and quality of methodology has increased. The most frequently observed outcomes with a positive impact were appropriateness of prescribing and cost savings. The vast majority of studies used multiple interventions, in conjunction with pharmacists' recommendations to physicians. Coupled with the use of practice guidelines or educational strategies, these interventions demonstrated a positive impact on economic or clinical outcomes. However, the data are still sparse and sometimes contradictory; therefore, further studies with randomized controlled designs are needed.