Strategies to contain the emergence of antimicrobial resistance: a systematic review of effectiveness and cost-effectiveness.

OBJECTIVES: To conduct a systematic review of the literature to describe and critically appraise studies reporting on the cost and/or effectiveness of interventions proposed to control the emergence of antimicrobial resistance (AMR).
METHODS: The search for relevant studies encompassed consultation with world experts in AMR, and electronic bibliographic database search of: Medline (1960-2000); ISI (1981-2000); EMBASE (1988-2000); Grey Literature (1999-2000); Database of Reviews of Effectiveness (DARE) and the NHS Health Economic Evaluation Database (HEED) at York University's Centre for Reviews and Dissemination (CRD) (numerous years); OPAC (1975-2000); and the Cochrane Library Online (1990-2000). Only studies that concerned the effectiveness or cost-effectiveness of measures specifically designed to contain the emergence of AMR were reviewed. Standardised data extraction sheets, based on existing checklists for effectiveness and cost-effectiveness, were used to assess the validity of each study using the 'risk of bias criteria' suggested in the Cochrane Handbook. Only studies categorised as being at low or moderate risk of bias were reported fully. The reliability of the data review process was monitored by comparison of several, random, independent assessments by all authors. The mix of study methods (i.e. including studies based on non-randomised controlled trials) meant that formal meta-analysis was not possible, and thus a qualitative review was performed.
RESULTS: In total, 43 studies were reviewed, with 21 classed as being at moderate or low risk of bias and therefore reported in the paper. These studies covered policies on: restricting the use of antimicrobials (five studies, suggesting that restriction policies can alter prescriber behaviour, although with limited evidence of subsequent effect on AMR); prescriber education, feedback and use of guidelines (six studies, with no clear conclusion); combination therapies (seven studies, showing the potential to lower drug-specific resistance, although for an indeterminate time period); vaccination (three studies showing cost/effectiveness). Most of these studies were: from the developed world, principally the USA; hospital-based, with few community level interventions; and concerned with effectiveness, not cost-effectiveness.
CONCLUSIONS: Overall, there is an absence of good evidence concerning what is effective, and especially cost-effective, in reducing the emergence of AMR. However, in addition to more research concerning these forms of intervention, the paper highlights four specific areas for further investigation: validating intermediate or surrogate outcome measures to enable better use to be made of the literature on intermediate measures; development and evaluation of 'macro' strategies; research into specific aspects of AMR in developing countries; and empirical and methodological research concerning the economic evaluation of interventions.