118 5 CHAPTER 5 10 | GOODARZI et al. 4 | DISCUSSION 4.1 | The complexity of risk selection This nationwide longitudinal retrospective cohort study in a healthy Dutch pregnant population offers novel insights into the associations between maternal characteristics, gestational age, and PM. The absolute number of fetal and neonatal death and the probability of TMwas low. The differences between subgroups of women have declined substantially over the past two decades. In most subcategories, the increase in TM rate with advancing gestational age was minimal. Overall, we observed the highest probability of TM among non-Dutch women, women with ≥3 previous births, women aged ≥40 years, and low-SES women. However, interaction analyses showed that the probability of TM differed among subgroups of women when taking a second maternal characteristic into account. We observed a higher probability of TM among low-SES women but only for non-Dutch and not for Dutch women. We observed the highest probability of TM among non-Dutch low-SES women, nulliparous women aged ≥40 years, and women aged up to 29 years with ≥3 previous births. 4.2 | The limitations of a routine approach Currently, in some countries, shared decision-making is used to offer healthy nulliparous women and women at advanced age IOL at term. Women are informed about the probability of TM associated with their parity or age and the benefits and harmful side effects of IOL. These harmful side effects have not been found in randomized controlled trials,23 which may be due to the low incidence of these side effects, and the non-representative samples in these studies.24 Furthermore, some consequences of IOL have been examined insufficiently, such as the long- term care outcomes for the mother and child and maternal experiences.11 It is presumed that women can make a well-informed decision about IOL based on the information about the benefits and harmful side effects of IOL. However, studies show that women often feel insufficiently involved and informed in the decision-making process regarding IOL, and women's choice is led by the presented information which is focused on concerns for their child's short-term outcomes.25,26 The results of this study do not support routinely offeringwomen IOL based on the probabilities of TMassociated with a single maternal characteristic. Furthermore, this is a sliding scale approach: because the probability of TM is never zero; ultimately, all healthy women should be informed about their risk of TM and be offered the choice of IOL. Also, care providers are not obliged to informwomen about every very small risk inherent to healthy pregnancy and birth.27 More importantly, the results of this study indicate that using single determinants as indications for IOL has poor predictive value.28,29 Our study results point to the difficulty of distinguishing which women would benefit from IOL. Interventions for which the balance between benefits and harms varies substantially among subgroups are also referred to as “gray zone” interventions.30,31 Brownlee and colleagues (2017) emphasize that for gray zone interventions, “even when robust consensus has established criteria defining the appropriateness of tests and treatments […], appropriateness can remain uncertain in many individual cases.”30 (p157) Risk selection regarding gray zone interventions is associated with professional bias, and overuse or underuse of care.30,31 Lastly, to prevent one case of TM, hundreds of healthy women and children who will not experience TM if labor is not induced will be needlessly exposed to the discomfort and disadvantages of IOL.4 Thus, applying a routine approach to IOL based on single maternal characteristics might result in underestimation or overestimation of probabilities of TM, leading to overuse or underuse of care.11 This approach also disproportionately puts the focus on mortality risks, causes fear, and shifts the responsibility for outcomes to women.4,32 However, it does not mean that TM risks should be ignored. Childbearing women need a high value care system that not only provides timely intervention using and allocating resources optimally, but also helps them stay safe and healthy by preventing unnecessary medical interventions.11 4.3 | Toward value-based health care The results of this study call for a comprehensive approach, in which women are individually assessed within their own context to better identify those women who would benefit from IOL while preventing inappropriate care. This requires further understanding of the interaction between maternal characteristics, gestational age and TM and the differences in TM between subgroups, which has been identified as one of the top research priorities necessary to improve risk selection.28,29 Furthermore, there is a call to expand the contemporary research focus on individual risk and include the social context as well.33,34 We recommend a value-based health care (VBHC)35 strategy– foregrounding women centered, evidence-based, appropriate, cost-effective, accessible, and equitable care–by investing in better screening and diagnostic methods,11 addressing social inequities,33,36 and using a multi- determinant approach in research and practice.20,21,28,29