Developing and evaluating interventions to reduce inappropriate prescribing by general practitioners of antibiotics for upper respiratory tract infections: a randomised controlled trial to compare paper-based and web-based modelling experiments.

Shaun Treweek,Martin Eccles,Debbie Bonetti,Claire Jones,Mark Weal,Ian W Ricketts,Jillian Francis,Nigel B Pitts,Frank Sullivan,Karen Barnett,Graeme Maclennan

doi:10.1186/1748-5908-6-16

Abstract

BackgroundMuch implementation research is focused on full-scale trials with little evidence of preceding modelling work. The Medical Research Council Framework for developing and evaluating complex interventions has argued for more and better theoretical and exploratory work prior to a trial as a means of improving intervention development. Intervention modelling experiments (IMEs) are a way of exploring and refining an intervention before moving to a full-scale trial. They do this by delivering key elements of the intervention in a simulation that approximates clinical practice by, for example, presenting general practitioners (GPs) with a clinical scenario about making a treatment decision.MethodsThe current proposal will run a full, web-based IME involving 250 GPs that will advance the methodology of IMEs by directly comparing results with an earlier paper-based IME. Moreover, the web-based IME will evaluate an intervention that can be put into a full-scale trial that aims to reduce antibiotic prescribing for upper respiratory tract infections in primary care. The study will also include a trial of email versus postal invitations to participate.DiscussionMore effective behaviour change interventions are needed and this study will develop one such intervention and a system to model and test future interventions. This system will be applicable to any situation in the National Health Service where behaviour needs to be modified, including interventions aimed directly at the public.Trial registrationClinicalTrials (NCT): NCT01206738

Highlights

Much implementation research is focused on full-scale trials with little evidence of preceding modelling work
The National Health Service and other healthcare providers need effective quality improvement interventions to be put into clinical practice, which requires effective behaviour change interventions
Intervention modelling experiments (IMEs) do this by delivering key elements of the intervention in a simulation that approximates clinical practice by, for example, presenting general practitioners (GPs) with a clinical scenario about making a treatment decision

Summary

Introduction

Much implementation research is focused on full-scale trials with little evidence of preceding modelling work. Intervention modelling experiments (IMEs) are a way of exploring and refining an intervention before moving to a full-scale trial. They do this by delivering key elements of the intervention in a simulation that approximates clinical practice by, for example, presenting general practitioners (GPs) with a clinical scenario about making a treatment decision. In an IME key elements of the intervention are delivered (generally in a randomised controlled trial, use of the word ‘experiment’) in a manner that approximates the real world but the measured outcome is generally an interim outcome, a proxy for the clinical behaviour of interest (clinical decision in response to a simulated clinical encounter) prior to entering the intervention into a fullscale trial. We will use the new system to run a web-based IME on antibiotic prescribing for upper respiratory tract infections (URTIs) in general practice

Objectives

Methods

Findings

Conclusion