Rebekah Hall1, Antonieta Medina-Lara2, Willie Hamilton2, Anne E Spencer2. 1. College of Medicine and Health, University of Exeter, South Cloisters, St Luke's Campus, Heavitree, Exeter, EX1 2LU, UK. rh591@exeter.ac.uk. 2. College of Medicine and Health, University of Exeter, South Cloisters, St Luke's Campus, Heavitree, Exeter, EX1 2LU, UK.
Abstract
BACKGROUND: Evidence from discrete choice experiments can be used to enrich understanding of preferences, inform the (re)design of screening programmes and/or improve communication within public campaigns about the benefits and harms of screening. However, reviews of screening discrete choice experiments highlight significant discrepancies between stated choices and real choices, particularly regarding willingness to undergo cancer screening. The identification and selection of attributes and associated levels is a fundamental component of designing a discrete choice experiment. Misspecification or misinterpretation of attributes may lead to non-compensatory behaviours, attribute non-attendance and responses that lack external validity. OBJECTIVES: We aimed to synthesise evidence on attribute development, alongside an in-depth review of included attributes and methodological challenges, to provide a resource for researchers undertaking future studies in cancer screening. METHODS: A systematic review was conducted to identify discrete choice experiments estimating preferences towards cancer screening, dated between 1990 and December 2020. Data were synthesised narratively. In-depth analysis of attributes led to classification into four categories: test specific, service delivery, outcomes and monetary. Attribute significance and relative importance were also analysed. The International Society for Pharmacoeconomics and Outcomes Research conjoint analysis checklist was used to assess the quality of reporting. RESULTS: Forty-nine studies were included at full text. They covered a range of cancer sites: over half (26/49) examined colorectal screening. Most studies elicited general public preferences (34/49). In total, 280 attributes were included, 90% (252/280) of which were significant. Overall, test sensitivity and mortality reduction were most frequently found to be the most important to respondents. CONCLUSIONS: Improvements in reporting the identification, selection and construction of attributes used within cancer screening discrete choice experiments are needed. This review also highlights the importance of considering the complexity of choice tasks when considering risk information or compound attributes. Patient and public involvement and stakeholder engagement are recommended to optimise understanding of unavoidably complex choice tasks throughout the design process. To ensure quality and maximise comparability across studies, further research is needed to develop a risk-of-bias measure for discrete choice experiments.
BACKGROUND: Evidence from discrete choice experiments can be used to enrich understanding of preferences, inform the (re)design of screening programmes and/or improve communication within public campaigns about the benefits and harms of screening. However, reviews of screening discrete choice experiments highlight significant discrepancies between stated choices and real choices, particularly regarding willingness to undergo cancer screening. The identification and selection of attributes and associated levels is a fundamental component of designing a discrete choice experiment. Misspecification or misinterpretation of attributes may lead to non-compensatory behaviours, attribute non-attendance and responses that lack external validity. OBJECTIVES: We aimed to synthesise evidence on attribute development, alongside an in-depth review of included attributes and methodological challenges, to provide a resource for researchers undertaking future studies in cancer screening. METHODS: A systematic review was conducted to identify discrete choice experiments estimating preferences towards cancer screening, dated between 1990 and December 2020. Data were synthesised narratively. In-depth analysis of attributes led to classification into four categories: test specific, service delivery, outcomes and monetary. Attribute significance and relative importance were also analysed. The International Society for Pharmacoeconomics and Outcomes Research conjoint analysis checklist was used to assess the quality of reporting. RESULTS: Forty-nine studies were included at full text. They covered a range of cancer sites: over half (26/49) examined colorectal screening. Most studies elicited general public preferences (34/49). In total, 280 attributes were included, 90% (252/280) of which were significant. Overall, test sensitivity and mortality reduction were most frequently found to be the most important to respondents. CONCLUSIONS: Improvements in reporting the identification, selection and construction of attributes used within cancer screening discrete choice experiments are needed. This review also highlights the importance of considering the complexity of choice tasks when considering risk information or compound attributes. Patient and public involvement and stakeholder engagement are recommended to optimise understanding of unavoidably complex choice tasks throughout the design process. To ensure quality and maximise comparability across studies, further research is needed to develop a risk-of-bias measure for discrete choice experiments.
Authors: Jennifer S Lin; Margaret A Piper; Leslie A Perdue; Carolyn M Rutter; Elizabeth M Webber; Elizabeth O'Connor; Ning Smith; Evelyn P Whitlock Journal: JAMA Date: 2016-06-21 Impact factor: 56.272
Authors: Lara Moons; An Mariman; Peter Vermeir; Luc Colemont; Els Clays; Hans Van Vlierberghe; Dirk Vogelaers Journal: Acta Clin Belg Date: 2019-01-04 Impact factor: 1.264
Authors: Glyn Elwyn; Dominick Frosch; Richard Thomson; Natalie Joseph-Williams; Amy Lloyd; Paul Kinnersley; Emma Cording; Dave Tomson; Carole Dodd; Stephen Rollnick; Adrian Edwards; Michael Barry Journal: J Gen Intern Med Date: 2012-05-23 Impact factor: 5.128