Strengthening the science of addressing antimicrobial resistance: a framework for planning, conducting and disseminating antimicrobial resistance intervention research

. 2020 Jun 8;18:60. doi: 10.1186/s12961-020-00549-1

Box 1

Current state of the evidence base on AMR policy interventions
Around the world, 129 governments are currently in the process of developing or implementing a National Action Plan to address antimicrobial resistance [19] Global capacity for AMR surveillance is lacking; discrepancies between methods and monitoring systems, data quality concerns and lack of representativeness make it challenging to compare AMR data between countries [20] Many evaluations of AMR policy interventions are conducted retrospectively by academics who were not involved in the design or implementation of the intervention [6, 21] A systematic review of experimental and quasi-experimental studies evaluating government policy interventions to reduce the use of antimicrobials [6, 22] found that 30 of the 69 studies used low-quality study designs, such as uncontrolled before–after designs, which severely limits the validity of their findings. Among these 69 studies, only 4 used a randomised controlled design which is considered the gold standard for evaluating interventions. Another systematic review of 221 interventions for improving antibiotic prescribing among hospital inpatients found the quality of the reporting for the 163 non-randomised studies was so poor that it was difficult for professionals to use the research findings or to implement interventions that were shown to be useful; further, this systematic review found that no useful evidence could be gleaned from studies using controlled before–after and non-randomised trial designs [17] Reporting of AMR policy intervention studies is weak; studies often fail to describe the intervention in sufficient detail for replication and many do not report the reason the intervention is expected to work [21] In the broader field of public health, researchers have estimated that at least 50% of published research is not sufficiently clear, complete or accurate for others to interpret or use [23, 24] There are no standardised measures and metrics for AMR research; many AMR intervention studies report antimicrobial use in defined daily dose per 1000 population or a simple prescribing rate [21]

Current state of the evidence base on AMR policy interventions

Around the world, 129 governments are currently in the process of developing or implementing a National Action Plan to address antimicrobial resistance [19]

Global capacity for AMR surveillance is lacking; discrepancies between methods and monitoring systems, data quality concerns and lack of representativeness make it challenging to compare AMR data between countries [20]

Many evaluations of AMR policy interventions are conducted retrospectively by academics who were not involved in the design or implementation of the intervention [6, 21]

A systematic review of experimental and quasi-experimental studies evaluating government policy interventions to reduce the use of antimicrobials [6, 22] found that 30 of the 69 studies used low-quality study designs, such as uncontrolled before–after designs, which severely limits the validity of their findings. Among these 69 studies, only 4 used a randomised controlled design which is considered the gold standard for evaluating interventions.

Another systematic review of 221 interventions for improving antibiotic prescribing among hospital inpatients found the quality of the reporting for the 163 non-randomised studies was so poor that it was difficult for professionals to use the research findings or to implement interventions that were shown to be useful; further, this systematic review found that no useful evidence could be gleaned from studies using controlled before–after and non-randomised trial designs [17]

Reporting of AMR policy intervention studies is weak; studies often fail to describe the intervention in sufficient detail for replication and many do not report the reason the intervention is expected to work [21]

In the broader field of public health, researchers have estimated that at least 50% of published research is not sufficiently clear, complete or accurate for others to interpret or use [23, 24]

There are no standardised measures and metrics for AMR research; many AMR intervention studies report antimicrobial use in defined daily dose per 1000 population or a simple prescribing rate [21]