Public (un)willingness to trade-off benefits and harms in decisions about cancer screening: A discrete choice experiment.
Le résumé fourni par la source
ObjectivesTo quantify the harms participants are willing to accept for varying benefits from cancer screening following provision of information.MethodsWe conducted an online discrete choice experiment with two samples of the UK public. Attributes were number of cancer diagnoses and deaths prevented (benefits) and overdiagnoses, false positives and false negatives (harms) for a population who are/are not screened for an unspecified cancer over their lifetimes. After selecting the best-fitting conditional logistic regression model for each sample, we used latent class analysis to investigate preference heterogeneity and calculated the probability that each class/group would take up screening with different benefit and harm levels.ResultsIn total, 1018 participants completed the discrete choice experiment. Cancer deaths prevented was the most important attribute, followed by the number of false positively/negatively detected cancers; cancers prevented and overdiagnosed were deemed relatively unimportant. Three latent classes were identified. The largest class (approximately two-thirds) chiefly opted for screening, focusing on deaths prevented; this group would strongly prefer screening participation even with harms only (probability 76%-92%). Approximately a quarter of each sample traded off benefits and harms; this group would likely take up screening with outcomes like those of colorectal screening (68%-87%).ConclusionsA minority of the public express preferences for screening that are influenced by trading-off potential benefits/harms, but most have strong preferences for cancer screening, focusing on the associated reduction of cancer deaths. The findings emphasise public health responsibility in appraising or generating evidence on benefits and harms to support decisions about the introduction or modification of screening programmes.
Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.