How do we discharge preterm infants safely without unnecessary delay?

Discussions about admission practice, length of stay or discharge planning often assume that we are referring to the same group of patients at the same level of care. We may also take for granted that we share a common understanding of what a good and safe practice looks like, and we would probably all denounce any kind of discrimination reflecting race or socio-economic background. But is this the truth? The Dartmouth health atlas project has documented unwarranted geographic variation in medical care independent of medical needs and patient preferences over the last 40 years.1 In 2016 the Norwegian neonatal health atlas documented a more than two-fold geographic variation in neonatal intensive care unit (NICU) admission rates for near-term and term infants as well as for antibiotic and ventilator treatment.2, 3 Recently Norman et al.4 found significant regional differences in care and management of very preterm infants in the Nordic countries. Also, hospital admission rates for children vary according to parental level of education.5 Eichenwald et al.6 observed a more than 1 week difference between NICUs in post-menstrual age (PMA) at discharge in a homogeneous population of premature infants. The same author later found variation in the proportion of moderately preterm infants diagnosed with apnoea, and this variation significantly affected the length of stay.7 Care sensitive to personal or unit preferences, such as admission of late preterm or term infants, usually demonstrates large variation between regions and NICUs, whereas the variation of admission of preterm infants born before 34 weeks is low, otherwise the consequences would be unacceptable.2 But even if admission is necessary, care when admitted may be highly preference-sensitive with a wide variation in outcomes such as length of stay, growth, breastfeeding rate, or parents' readiness to take care of their infant after discharge. In this issue of Acta Paediatrica, Arwehed and a group of authors representing all five Nordic countries publish results from a survey of discharge criteria used by 83 Nordic. NICUs represent 93% of all available NICUs in the region.8 First, the authors should be complimented for their high response rate which provides an excellent starting point for a description of discharge practices and preparations among Nordic NICUs. These NICUs are all parts of fully financed public healthcare systems providing treatment not restricted by socio-economic background or type of health insurance. Access and adherence to ante- and postnatal healthcare programmes for mothers and infants are free of charge and almost universal, population-based outcomes after preterm birth are good4 and infant mortality rates are very low. Parents also have the right to financial support post-partum facilitating parental presence and family-centred care in the NICU. Unexpected death shortly after discharge in otherwise healthy and stable preterm infants is rare, although more common than in term infants. Still, most experienced neonatologists have one or more times encountered the tragedy of losing a preterm infant to sudden infant death syndrome (SIDS) shortly after discharge. To try to increase the margin of safety, observation time after the last registered apnoea may be increased, apnoea monitoring system or definition of clinically significant apnoea may be changed, or one may implement a higher PMA or weight limit as criteria for discharge. It may be hard to argue against the saying ‘it's better to be safe than sorry’, but evidence for improved safety after such measures is non-existent and they are most likely without effect. What we do know is that they may reduce the overall quality of care due to unnecessary long hospital stay with a negative effect on breast feeding, growth and empowerment of parents. Except from the general recommendations to prevent SIDS, we do not know how to prevent sudden death in preterm infants after discharge. The survey identifies large variations in discharge preparations and criteria both between and within countries. However, variation is low when physicians are asked about their opinion on early discharge; more than 80% of respondents consider early discharge and the home environment as best for growth, breastfeeding, and empowerment of parents. Only 58% of units had a standard definition of apnoea, and the standards varied widely between units. Forty-seven per cent of units deemed short feeding-related apneas or bradycardias unacceptable for discharge, contrasting recommendations from the American Academy of Pediatrics who states that such events are not predictive of later clinically significant apneas or SIDS.9 The survey does not ask whether gestational age (GA) is included in discharge evaluation and risk estimation. There is a strong correlation between low GA and delayed maturation of cardio-respiratory reflexes, and clinically significant apneas resolve at a higher post-menstrual age in very- and extremely preterm infants. But since infants born <32 weeks GA only constitutes 20%–25% of all infants born preterm, most infants may bear a low risk for significant apnoea's and bradycardias after 34–35 weeks PMA and may therefore be good candidates for early discharge. Parental involvement in care and a non-paternalistic staff-parent collaboration valuing shared decision-making is the foundation of family-centred care, but only 42% of units include parents' clinical assessment of their infant's discharge readiness in their evaluation, and parent assessment of readiness is done by nurses in 95% of units. Eighty per cent of the units ask parents if they are ready, but we know little about how these answers are interpreted. Real shared decision-making is challenging and demands a change in our culture of care. We are not there yet, despite good intentions. One disturbing result from the survey is how parents' language skills, smoking habits, access to a car or other indicators of socio-economic status prohibit early discharge due to defined criteria. At first sight, it might seem reasonable to make sure that parents are able to communicate with the home service staff or drive by their own car if they need to get to the hospital, but such criteria strongly prohibit early discharge based on indicators reflecting race and socio-economic status. As such, these criteria may facilitate structural discrimination and should be omitted and replaced by measures to increase the use of qualified interpreter services and extended support for parents in situations deemed challenging. The authors state that discharge criteria that do not increase safety or enhance care quality should be omitted. Hopefully, this survey is the first step in a project aiming to determine safe discharge criteria without delaying discharge. Nordic countries have good health registers allowing studies of unwarranted outcomes like early SIDS in preterm infants after discharge. Large variation are unwarranted but do not tell us which practice is better. However, with most respondents agreeing that early discharge is better than prolonged stay if safety is taken care of, we are given an opportunity to hold the effect of different discharge criteria and practices up against a standard deemed better by four out of five responding neonatologists. Atle Moen: Conceptualization; writing – review and editing. The author has no conflict of interest to declare.