Consensus group on the design, analysis and reporting of antibiotic stewardship trials

Overuse of antibiotic is a major contributor to increasing resistance to antibiotics. One of the tools to reduce the inappropriate use of antibiotics is by antibiotic stewardship. Antibiotic stewardship interventions prevent antibiotic misuse by influencing the prescription behaviour of doctors. Research of antibiotic stewardship interventions are often of insufficient quality to demonstrate the safety of antibiotic stewardship. This is partly because it is unknown what are the best methods to design and analyse these interventions.

We will establish a working group of experts on antibiotic stewardship and research design to determine the best way to design, analyse, and report antibiotic stewardship interventions. The goal of the working group is to formulate recommendations in a research guideline on antibiotic stewardship studies. The guideline aims to improve the quality of antibiotic stewardship research, which will lead to more effective use of recourses of safe and effective antibiotic stewardship interventions.

The primary goal of antibiotic stewardship is to reduce use of inappropriate antibiotics while ensuring non-inferior clinical outcome. The highest level of evidence is provided by antibiotic stewardship trials (ASTs). Currently, up to 80% of the publications on ASTs are of insufficient methodological quality to provide interpretable conclusions (Molina 2015). This results - at least in part - from the absence of guidelines and standardized methods on how to design ASTs. To ensure that ASTs do not negatively affect the quality of patient care, non-inferiority designs are often used. However, predefined non-inferiority margins are subjective and non-inferiority designs require large sample sizes, increasing the risk of underpowered studies (Schulz, 2005). As a result, many ASTs are descriptive and not sufficiently powered for evidence-based policymaking. A new alternative method using a composite endpoint of antibiotic use and clinical outcome was recently proposed for evaluating ASTs; the RADAR method (Evans, 2015). This method uses a superiority design, requiring smaller sample sizes. However, with this method outcomes are difficult to interpret and susceptible to subjective choices in the analysis (Phillips, 2016; Schweitzer, in preparation). It is therefore fair to conclude that a consensus on how to properly design and analyse ASTs is lacking. We will establish an international working group consisting of experts on trial methodology and antibiotic stewardship with the aim to establish a best-practice guideline for designing, analysing and reporting ASTs. The working group will evaluate existing methods, develop new methods or improve existing methods, evaluate which settings are important to consider and define the optimal methods for the design and analysis of ASTs per relevant setting. The final product will be a best-practice guideline containing recommendations for designing, analysing and reporting ASTs.