Samantha Husbands1, Susan Jowett2, Pelham Barton2, Joanna Coast3. 1. Population Health Sciences, Bristol Medical School, University of Bristol, 1-5 Whiteladies Road, Bristol, BS8 1NU, England, UK. Samantha.husbands@bristol.ac.uk. 2. Health Economics Unit, Institute of Applied Health Research, University of Birmingham, Birmingham, England, UK. 3. Population Health Sciences, Bristol Medical School, University of Bristol, 1-5 Whiteladies Road, Bristol, BS8 1NU, England, UK.
Abstract
INTRODUCTION: Decision-analytic models play an essential role in informing healthcare resource allocation decisions; however, their value to decision makers will depend on model structures being clinically valid to determine cost-effectiveness recommendations. Clinician involvement can help modellers to develop clinically valid but straightforward structures; however, there is little guidance available on methods for clinician input to model structure. This study aims to provide an in-depth exploration of clinician involvement in structural development, highlighting key issues and generating recommendations to optimise practices. METHODS: A qualitative study was undertaken with a range of modellers and clinicians working in different modelling contexts. In-depth interviews and case studies using observations were carried out to understand how clinicians are involved in model structural development and to identify problems and optimal approaches from informants' perspectives. RESULTS: Twenty-four interviews and two case studies were undertaken with modellers and modelling teams. Key issues included the number and diversity of clinicians contributing to structural development, potentially impacting the generalisability of structures, and problems with clinician understanding of important information to contribute to model pathways. Modellers and clinicians suggested that clinician training in modelling could enhance structural processes. CONCLUSIONS: Recommendations to optimise current practices include recruiting clinicians from a variety of backgrounds and using discussions between experts to develop valid and generalisable structures. Future research should focus on developing training materials for clinicians and finding ways to help modellers recruit clinicians from different settings.
INTRODUCTION: Decision-analytic models play an essential role in informing healthcare resource allocation decisions; however, their value to decision makers will depend on model structures being clinically valid to determine cost-effectiveness recommendations. Clinician involvement can help modellers to develop clinically valid but straightforward structures; however, there is little guidance available on methods for clinician input to model structure. This study aims to provide an in-depth exploration of clinician involvement in structural development, highlighting key issues and generating recommendations to optimise practices. METHODS: A qualitative study was undertaken with a range of modellers and clinicians working in different modelling contexts. In-depth interviews and case studies using observations were carried out to understand how clinicians are involved in model structural development and to identify problems and optimal approaches from informants' perspectives. RESULTS: Twenty-four interviews and two case studies were undertaken with modellers and modelling teams. Key issues included the number and diversity of clinicians contributing to structural development, potentially impacting the generalisability of structures, and problems with clinician understanding of important information to contribute to model pathways. Modellers and clinicians suggested that clinician training in modelling could enhance structural processes. CONCLUSIONS: Recommendations to optimise current practices include recruiting clinicians from a variety of backgrounds and using discussions between experts to develop valid and generalisable structures. Future research should focus on developing training materials for clinicians and finding ways to help modellers recruit clinicians from different settings.