Outcomes for all of these categories of newborns can be influenced by national health policies. For example, a high rate of termination for the “most severe” hypoplastic left heart syndrome will lead to better surgical outcomes for babies with this condition. A decision to do cesarean sections only for fetuses over 25 weeks of GA will likely select bigger non-growth restricted neonates at each GA and will affect outcomes. Another striking example is what happened in Holland in 2007. The Dutch government started offering structural ultrasound at 20 weeks’ gestation for all pregnant women at no extra cost. This has resulted in fewer births of babies with severe anomalies, such as spina bifida (because of termination of pregnancy), which in turn has led to fewer cases of neonatal euthanasia [24]. Another example is the policy statements for treatment of extremely premature infants. In Canada, palliative care is advised in babies born at less than 23 weeks of gestational age, and cesarean section is not recommended below 24 weeks of gestational age [14]. In the United States, gestational age is not mentioned. In the Netherlands, non-intervention is now advised below 24 weeks of gestational age, and between 24 and 25 weeks is defined as the “gray zone” where interventions are optional, depending on parental preferences [15]. The latter policy will result in fewer survivors less than 25 weeks of gestational age, fewer survivors with handicap, and also fewer survivors overall. The policy may also result in less desire to intervene below 25 weeks gestational age, thus contributing to a self-fulfilling prophecy.

For these reasons, we should consider making all fetuses and neonates who are alive at more than 450 g the denominator of all deaths, even if these fetuses are in utero. This classification is more complex than categorizing neonatal deaths and also involves another set of ethical values (and other disciplines).

The examples discussed in this paper underline that national and/or professional healthcare policies can change medical practice and shape patient outcomes. Conversely, outcomes could also influence national health policies and local/personal values. If the prevalence of individuals with disability from prematurity or congenital malformation in our communities decreases, society might adapt. It would be interesting to learn what that might do to health care infrastructure and the availability of support and acceptance of children with disabilities among the public. What effect might this situation have on policy makers and on physicians’ perinatal and counseling and decision-making?

Knowledge about how babies die is important because decisions about WH/WD in the delivery room and in the NICU are in part based on personal, local and/or national values. Those values shape the patient mix, outcome in NICUs, and the context of clinical and ethical decisions. Reflection on those values and on the differences in values between units, as well as proper comparison of outcome data, can take place only if we start including in our calculations data about how babies die and who never makes it to the NICU. Only if we can recognize our biases and make more transparent the true reasoning behind our decision-making processes can we be empowered to respond appropriately and consistently to the needs of sick children and their families.

Today, decisions on when to start WH or WD life supportive interventions in critically ill children are among the most difficult decisions in pediatric practice. These decisions directly affect the health of our societies. It is difficult to remain neutral when we make these decisions. We expect that full transparency of the circumstances around these decisions will allow us to compare outcomes better, reflect on similarities and differences in end-of-life care, and ultimately make the necessary improvements in the interest of the child.