Objectives: The purpose of this study was to explore the effect of information framing on parental decisions about resuscitation of extremely premature infants. Secondary outcomes focused on elucidating the impact of other variables on treatment choices and determining whether those effects would take precedence over any framing effects.
Methods: This confidential survey study was administered to adult volunteers via the Internet. The surveys depicted a hypothetical vignette of a threatened delivery at gestational age of 23 weeks, with prognostic outcome information framed as either survival with lack of disability (positive frame) or chance of dying and likelihood of disability among survivors (negative frame). Participants were randomly assigned to receive either the positively or negatively framed vignette. They were then asked to choose whether they would prefer resuscitation or comfort care. After completing the survey vignette, participants were directed to a questionnaire designed to test the secondary hypothesis and to explore possible factors associated with treatment decisions.
Results: A total of 146 subjects received prognostic information framed as survival data and 146 subjects received prognostic information framed as mortality data. Overall, 24% of the sample population chose comfort care and 76% chose resuscitation. A strong trend was detected toward a framing effect on treatment preference; respondents for whom prognosis was framed as survival data were more likely to elect resuscitation. This framing effect was significant in a multivariate analysis controlling for religiousness, parental status, and beliefs regarding the sanctity of life. Of these covariates, only religiousness modified susceptibility to framing; participants who were not highly religious were significantly more likely to be influenced to opt for resuscitation by the positive frame than were participants who were highly religious.
Conclusions: Framing bias may compromise efforts to approach prenatal counseling in a nondirective manner. This is especially true for subsets of participants who are not highly religious.