Abstract
Background: As the infant's best interests are determined through the perinatal decisions of family members and physicians, it is important to understand the factors that affect such decisions. This paper investigated the separate and combined effects of various factors related to perinatal decision making and sought to determine the best way to convey information in an unbiased manner to family members. Methods: In total, 613 participants were consecutively recruited. Each participant completed a series of surveys. All responses to four items were examined via a latent class analysis (LCA) to identify subgroups of participants with similar preferences for intensive care (IC) and comfort care (CC) regarding their potentially premature infant. Multiple logistic regression analyses were applied to identify the sociodemographic predictors for the classification of participants into different subgroups. Results: χ(2)-tests indicated that perinatal decision making for Item 2 was influenced by framing information, whereas decision making wasn't affected by presentation modes. Furthermore, the endorsement rates of IC significantly decreased with the information increased from brief to detailed, regardless of framing or presentation mode. The LCA indicated that a 3-subgroup model, which included the IC, CC, and variation subgroups, was optimal. Logistic regression analyses demonstrated that, compared with the IC subgroup, negative framing, higher education, parenthood, and poor numeracy predicted participants' preferences for CC. Meanwhile, worrying about physical or mental disabilities predicted preferences for CC and sensitivity to the amount of information provided regarding treatment options (variation subgroup). Conclusions: Perinatal decision making is affected by many factors, suggesting that more detailed information, improved understandability of numerical data, and a neutral tone of voice regarding therapeutic outcomes would be helpful for the families of premature infants to make unbiased decisions. Our findings should be replicated in future research.