A Systematic Review of Discrete Choice Experiments and Conjoint Analysis on Genetic Testing.

BACKGROUND: Although genetic testing has the potential to offer promising medical benefits, concerns regarding its potential negative impacts may influence its acceptance. Users and providers need to weigh the benefits, costs and potential harms before deciding whether to take up or recommend genetic testing. Attribute-based stated-preference methods, such as discrete choice experiment (DCE) or conjoint analysis, can help to quantify how individuals value different features of genetic testing. OBJECTIVES: The aim of this paper was to conduct a systematic review of DCE and conjoint analysis studies on genetic testing, including genomic tests. METHODS: A systematic search was conducted in seven databases: Web of Science, CINAHL Plus with Full Text (EBSCO), PsycINFO, PubMed, Embase, The Cochrane Library and SCOPUS. The search was conducted in February 2021 and was limited to English peer-reviewed articles published until the search date. The search keywords included relevant keywords such as 'genetic testing', 'genomic testing', 'pharmacogenetic testing', 'discrete choice experiment' and 'conjoint analysis'. Narrative synthesis of the studies was conducted on survey population, testing type, recruitment and data collection, survey development, questionnaire content, survey validity, analysis, outcomes and other design features. RESULTS: Of the 292 articles retrieved, 38 full-text articles were included in this review. Nearly two-thirds of the studies were published since 2015 and all were conducted in high-income countries. Survey samples included patients, parents, general population and healthcare providers. The articles assessed preferences for pharmacogenetic testing (28.9%), predictive testing and diagnostic testing (18.4%), while only one (2.6%) study investigated preferences for carrier testing. The most common sampling method was convenience sampling (57.9%) and the majority recruited participants via web-enabled surveys (60.5%). Review of literature (84.6%), discussions with healthcare professionals (71.8%) and cognitive interviews (53.8%) were commonly used for attribute identification. A survey validity test was included in only one-quarter of the studies (28.2%). Cost attributes were the most studied attribute type (76.9%), followed by risk attributes (61.5%). Among those that reported relative attribute importance, attributes related to benefits were the most commonly reported attributes with the highest relative attribute importance. Preference heterogeneity was investigated in most studies by modelling, such as via mixed logit analysis (82.1%) and/or by using interaction effects with respondent characteristics (74.4%). Willingness to pay was the most commonly estimated outcome and was presented in about two-thirds (n = 25; 64.1%) of the studies. CONCLUSION: With the continuous advancement in genetic technology resulting in expanding options for genetic testing and improvements in delivery methods, the application of genetic testing in clinical care is expected to rise. DCEs and conjoint analysis remain robust and useful methods to elicit preferences of potential stakeholders. This review serves as a summary for future researchers when designing similar studies.

Duke Scholars

Author Semra Ozdemir Population Health Sciences