Each year 15 million babies are born preterm and their survival chances vary dramatically around the world [1]. For the 1.2 million babies born in high income countries, increasing complexity of neonatal intensive care over the last quarter of the 20th century has changed the chances of survival at lower gestational ages. Middle-income and emerging economies have around 3.8 million preterm babies each year, and whilst some countries such as Turkey and Sri Lanka have halved deaths for preterm babies within a decade, other countries have made minimal progress [2]. South Asia and sub-Saharan Africa account for almost two-thirds of the world's preterm babies and over three-quarters of the world's newborn deaths due to preterm birth complications [1]. Worldwide, almost half of preterm babies are born at home, and even for those born in facilities, essential newborn care is often lacking.

This paper is the fifth in a supplement entitled "Born Too Soon". Previous papers in this series have outlined the policy context [3], epidemiology [4], and interventions preconceptually [5] and during pregnancy [6]. In this paper we focus on care of preterm newborns.

We apply the simple WHO definition of all babies born at less than 37 weeks gestation, noting that this includes both provider initiated and spontaneous preterm birth, and many varying causations [4]. Most premature babies (>80%) are born between 32 and 37 weeks of gestation (moderate/late preterm), and many die needlessly for lack of simple, essential care such as warmth and feeding support (Figure 1). About 10% of preterm babies are born 28 to <32 weeks gestation, and in low-income countries more than half of those will die but many could be saved with feasible care, not including intensive care such as ventilation (Figure 1). For babies born before 28 weeks gestation, intensive care would be needed to save most of these, but it is important to realise that these are the minority - about 5% of premature babies. Yet in many countries, families and health care providers still perceive the deaths of any premature baby as inevitable.

In contrast, in high-income settings neonatal survival is extending to lower and lower extremes of gestational age. In 1990, few babies under 25 weeks gestation were surviving; yet by 2010, 95% of preterm babies under 28 weeks survived, and more than half of the babies born before 25 weeks gestation survived, although the latter have a higher risk of impairment [7].

Over the last few decades the survival gap for babies born in high-income countries and babies born in the poorest countries has widened dramatically, even though the pace of survival gains in high-income countries has slowed reaching the extremes of preterm gestation. For example, North America is still achieving an average annual reduction of more than 5% per year for preterm-specific mortality, yet Africa on average is improving mortality rates for preterm babies by only 1% a year (Figure 2). Those countries with the highest risk of death and the most feasible deaths to avert are still experiencing the least progress. The history of neonatal care in high-income countries shows that the major reduction in deaths occurred before neonatal intensive care was established. Yet the risk of a neonatal death due to complications of preterm birth is about twelve times higher for an African baby than for a European baby [8] (Figure 2).

An important but under-recognised issue for all countries is that of disability for survivors of preterm birth [1, 9]. In the early days of neonatal intensive care, disabilities were common amongst survivors, ranging from some school learning disability through to severe cerebral palsy. Impairment outcomes have a heavy toll on families and on the health system. Indeed a recent report estimated that the average baby born 28 to 31 weeks gestation in the United States costs $95,000 in medical care in the first year alone [10]. Overtime the pattern of impairment from preterm birth in high-income countries has shifted. The focus of intensive care has shifted to extremely premature babies (less than 28 weeks), or "micro preemies", as this smaller subset of babies has increased risk and severity of impairment [7, 11]. With the scale up of neonatal intensive care, a focus on follow up and family support is critical.

Recent data show that even late and moderate preterm (LAMP, or 32 to <37 weeks gestation) is also associated with significant adverse effects, including those on school learning, prompting increasing debate regarding avoidable causes of moderate preterm birth such as high caesarean birth rates [12, 13]. These long-term effects on society and on the health system as well as more evidence of the link with non-communicable diseases in later adult life [3] underline that the importance of addressing preterm birth is beyond survival alone.

Over the last four decades with an increasing focus on evidence-based care for premature babies in high-income countries, the risk of long-term impairments is reducing. Neonatal intensive care has also become less interventionist and hence some aspects are also potentially more feasible to adapt to lower-income settings. There have been notable advances in quality of intensive care for premature babies.

Widespread use of antenatal corticosteroids in high-and some middle-income countries for mothers at gestation of 32 weeks or less, following multiple RCTs and the National Institute of Health consensus statement [14], ensuring that babies are less likely to develop respiratory distress syndrome (RDS), or have less severe RDS [15–17]. All trials have been conducted in settings where intensive or special care for preterm infants is available. While the effect has biological plausibility, the magnitude of effect in low-income countries without intensive care is uncertain, although a meta-anlsysis for middle-income country trials showed a greater effect than in high-income settings. An NIH sponsored trial in low- and middle-income settings is ongoing.

A shift to less intensive ventilator pressures and increasing use of continuous positive airway pressure (CPAP), now often the respiratory support method of choice [18].

Detailed quality of care protocols and "job aids" for almost every aspect of care have improved quality and also shifted more care to the responsibility of skilled neonatal nurses, particularly with respect to addressing infection prevention, feeding support, use of intravenous fluids, and safe oxygen use with careful tracking of oxygen saturation levels and follow-up services [19].

Deliberate attention to baby friendly care, reducing pain and over stimulation and more family friendly care, including family rooms linked to neonatal units and increased access for parents to their babies while in neonatal care units [20].

In low- and middle-income countries, there are limited comparable data on long-term outcomes after preterm birth [21, 22]. However, small studies suggest a high risk of moderate or severe neurodevelopmental impairment and an urgent need to improve awareness, data and care. Retinopathy of prematurity caused an epidemic of blindness for preterm babies in Europe and North America 50 years ago, especially after high or unmonitored use of oxygen. Data from Latin America show increasing rates of retinopathy of prematurity [23, 24] and it is likely that areas without data such as Southeast Asia are also experiencing an increase, recreating an avoidable problem. As neonatal care is improved and complexity increases, monitoring quality of care and tracking impairment outcomes are critical and should not be considered an optional extra in low-resource settings. Urgent attention is needed to develop standard, simpler measures of such impairments, to integrate these metrics into other measurement systems, and to provide support for such babies and their families [21].

All newborns are vulnerable given that birth and the following few days hold the highest concentrated risk of death of any time in the human lifespan. Every baby needs essential newborn care, ideally with their mothers providing warmth, breastfeeding and a clean environment. Premature babies are especially vulnerable to temperature instability, feeding difficulties, low blood sugar, infections, and breathing difficulties (Table 1). There are also complications that specifically affect premature babies (Figure 3).

Risk for all babies, especially those who are preterm	Essential care for all babies	Extra care for preterm babies
Hypothermia = low body temperature (increased risk of infections, mortality and for preterm babies increased risk of RDS)	Thermal care Drying, warming, skin-to-skin and delayed bathing	Extra thermal care Kangaroo Mother Care, baby hats, blankets, overhead heaters, incubators
Cord and skin infections, neonatal sepsis	Hygienic cord and skin care at birth and home care practices Hand washing and other hygiene Delayed cord clamping Consider chlorohexifine	Extra attention to infection prevention and skin care Consider chlorohexidine and emolients
Hypoglycemia = low blood sugar (Increased risk of impairment or death)	Early and exclusive breastfeeding	Extra support for breastfeeding e.g. expressing and cup or tube feeding, supplemented breast milk if indication Lack of breast milk is a risk factor for necrotizing entereocolitis in preterm babies
Hypoxia = low oxygen levels (Increased risk of impairment or death for preterm babies, higher risk of RDS and intracranial bleeding)	Neonatal resuscitation if not breathing at birth Bag-and-mask resuscitation with room air is sufficient for >99% of babies not breathing at birth	Safe oxygen use Monitored oxygen use e.g. in head box or with nasal cannula, routine use of pulse oximeters

Source: : Born Too Soon, Chapter 5 [113].

Recognition of small babies and distinguishing which ones are preterm are essential first steps in prioritising care for the highest risk babies. First trimester ultrasound assessment is the most accurate measure, but this is not available for most of the world's pregnant women [4]. Other options include Last Menstrual Period, using birthweight as a surrogate or assessment of the baby to estimate gestational age (e.g., Dubowitz or other simpler scoring methods). The highest-risk babies are those that are both preterm and growth restricted.

Care at birth from a skilled provider is crucial for both women and babies and all providers should have the competencies to care for both mother and baby, ensuring that mother and baby are not separated unnecessarily, promoting warmth, early and exclusive breastfeeding, cleanliness and resuscitation if required [26]. These practices are essential for full-term babies, but for premature babies, missing or delaying any of this care can rapidly lead to deterioration and death. For all babies at birth, minutes count.

Most published trials come from high-income countries where care for premature babies assumes the presence of neonatal intensive care, and large multi-site trials often examine the incremental effect of a specific change in care. Few rigorous trials are undertaken in lower-income settings where severe morbidity and fatal outcomes are common, contextual challenges may be critical and the counterfactual or control group should really be women or neonates receiving no care at all as this is the real question for policymakers. For example, in the KMC RCTs the control group was those receiving routine incubator care, which may dilute the impact measured compared to a counterfactual of no care. Ironically, more of the large recently funded rigorous trials are community-based, such as those assessing chlorhexidine and emollients [51], and there is an urgent need for more facility-based research addressing quality of care and including cost analyses.

There were a number of interventions considered in a systematic review of essential interventions that are used in high-income settings for premature babies but were not included in the global recommendations for scale up due to lack of context-specific evidence on cost effectiveness-for example, caffeine citrate to reduce the risk of apnea of prematurity [83]. Thus, more evidence from low-income settings is required particularly with respect to context-specific adaptation and associated implementation realities.

National coverage data for many of the evidence-based interventions for premature babies are lacking even in high-income settings, hence it is difficult to assess the global situation for care of premature babies or indeed for several important newborn care interventions (Figure 4).

For the 50 million home births without skilled care, the poorest women in the poorest countries, a major care gap is obvious. In sub-Saharan Africa, more than half of home births are alone, with no attendant [84]. In South Asia, around one-third of home births are without traditional birth attendants. In these instances, the primary caregivers of babies are their mothers and their families. Ensuring that women and communities are informed about healthy home care and enabled to care for their newborns and especially their preterm babies in the best possible way is critical. Women's groups which offer peer counselling and community mobilisation have been shown to have a significant effect on maternal and neonatal mortality [44, 85, 86].

The increasing pace of policy and program change for home-visit packages during pregnancy and after birth provides an opportunity to empower women to have a better outcome themselves and for their babies [87]. An early postnatal visit (within two days of birth) is one of only seven coverage indicators along the continuum of care selected by the United Nations Commission on Information and Accountability and tracked by Countdown to 2015 [88, 89]. In the 75 Countdown to 2015 priority countries, only 1 in 3 women and babies have an early postnatal visit -- the lowest of the seven indicators. This early visit is critical for survival and health and an important opportunity to identify preterm babies. Novel methods for identification of premature babies include community health workers using foot size to identify those babies and then providing extra visits, breastfeeding support and referral to a facility if needed [90].

As well as gaps in coverage of crucial interventions for women and babies, there are equity gaps between rich and poor, public and private health sectors, provinces and districts and among rural, urban and peri-urban populations. Complex, facility-based interventions tend to have a higher level of inequity than simpler interventions that can be delivered closer to home [91]. For example, there is low inequity for immunisation and antenatal care, while higher disparities exist for skilled attendance coverage [92]. Among the 54 of 75 Countdown to 2015 priority countries with equity data, birth in a health facility is more than twice as likely for a richer family compared to a poorer family [93].

Many African and most South Asian countries are experiencing increases in health facility births, some very rapidly [89]. However, the quality of care has not kept pace with coverage, leaving a quality gap but also giving cost-effective opportunities for lifesaving care for women and babies who are reachable in health facilities. For example, midwives are skilled and equipped to provide essential newborn care and resuscitation if needed. However, often key commodities or attention to infection prevention are lacking. Perinatal audit data and process can be a powerful tool for improving quality of care and can also be collated and used for national or subnational improvement of care [94].

Figure 4 shows the coverage and quality gaps for premature baby care in the Countdown to 2015 priority countries, highlighting the data gaps. With just over 50% of all births taking place in health facilities, essential newborn care could be provided for all those babies. Yet data show even the apparently simple practices of hand cleansing and warmth in the labour room are poorly done around the world [95]. Early initiation of breastfeeding is tracked by national household surveys but the practices for premature babies and duration of breastfeeding for preterm babies is not known at national level.

Neonatal resuscitation scale up is benefitting from recent innovation in technology and from public-private partnerships [96] and also more attention since being listed as one of the 13 priorities for the United Nations Commission on Life Saving Commodities for Women and Children [97]. However, data from Service Provision Assessment surveys suggested that under half of all skilled birth attendants had resuscitation skills and/or the correct equipment in terms of bag-and-mask (Figure 4) [54].

KMC, despite being established for more than 20 years, has had limited scale up (Figure 5). It is currently implemented on a large scale in only a few countries such as Colombia, Brazil and South Africa. There has also been rapid uptake in neonatal intensive care units in highincome countries, including for ventilated babies [98]. Systematic scale up of KMC is making progress in some countries in sub-Saharan Africa and South Asia including Malawi [99], Tanzania, Rwanda, Ghana [100], Indonesia and Vietnam [101]. In other countries, a KMC unit established in one teaching hospital over a decade ago has yet to benefit babies in the rest of the country. Lessons are being learned in overcoming barriers such as lack of knowledge by policy makers and service providers. Countries that are making more rapid progress have a national policy for KMC, a learning site, national champions, and a plan for national implementation. IN addition, they have integrated training along with essential newborn care and resuscitation into pre-service medical and nursing education (Figure 5). KMC can be safely delivered by trained patient attendants under the supervision of nurses, allowing nurses to look after the sickest neonates - a successful example of taskshifting [99]. A major impediment to program tracking and accountability is the lack of data for coverage of KMC, although this indicator could feasibly be tested for inclusion in household surveys.

Quality of service provision requires the availability of people with the right skills (Figure 6) as well as essential equipment and drugs. Indeed, for newborn survival, skilled people are at least as critical as equipment and commodities (Table 3) [102]. Shortages of qualified health workers and inadequate training and skills for the care of premature babies are a major reason for poor progress in reducing neonatal deaths [92, 103]. Nurses or midwives with skills in critical areas such as resuscitation, KMC, safe oxygen management, and breastfeeding support are the frontline worker for premature babies, yet in the whole of sub-Saharan Africa there are no known neonatal nurse training courses. Urgent systematic attention is required for pre-service and in-service training, non-rotation of nurses with skills in neonatal care, and where appropriate the development of a neonatal nurse cadre, as well as rewarding for those who work against the odds in hard-to-serve areas [2].

Addressing newborn care in district hospitals is a key priority for improving newborn survival and health. In most countries, district hospitals are understaffed and poorly resourced compared to teaching hospitals. Design and implementation of context-specific hospital newborn care packages is critical, especially as more births occur in facilities, also with referral transport and communications linkages between home and hospital. Newborn units at this level should aim at providing warmth (using KMC, or radiant warmers), assisted enteral feeding of expressed breast milk (by feeding tube, spoon, 'paladai'), intravenous fluids for sick babies, antibiotics, oxygen, and, if possible, CPAP. There are a number of large-scale examples of improved newborn care in district hospitals including a network in rural Western Kenya [104]. In Limpopo, South Africa, a network of more than 30 district hospitals instituted an accreditation scheme and targeted quality improvement with mentor teams [105], and in another province, KwaZulu-Natal, a program called Neonatal Experiential Leaning reaches 16 hospitals with standard guidelines, resuscitation workshops, a 2-week neonatal training course, and monthly mentor visits.

Across several Indian states, peripheral hospitals have developed a dedicated newborn care space ("Newborn corner") including basic equipment, while referral hospitals have upgraded special-care baby units. According to the government's most recent data, there are now 13,219 newborn care corners, 1,574 newborn stabilisation units and 448 special newborn care units. In an attempt to remove financial barrier to neonatal care, India has introduced a program (Janani Shishu Suraksha Karyakaram) that entitles all pregnant women and neonates to free care at public facilities including free drugs and free transport from and to home. An evaluation of the special newborn care units (SNCU) concluded that it is possible to set up and manage quality SNCUs and improve the survival of small and preterm newborns and those with sepsis, although several challenges relating to human resources, maintenance of equipment, and asepsis remain [106]. Skilled and motivated nurses are the key to successful neonatal units. Some also have experimented with the use of alternative cadres of ward aides specially trained in newborn care and restricted from rotations to other wards [107].

The neonatal mortality rate (NMR) in the United Kingdom and the United States was reduced to below 15 per 1,000 live births before neonatal intensive care was widely available, and the largest reduction in NMR from 40 to 15 was related to obstetric care and simpler improvements in individualised newborn care such as warmth, feeding, infection prevention, and case management (Figure 7).

Seven low- and middle-income countries have halved their preterm deaths within a decade [2]. These countries are Sri Lanka, Turkey, Belarus, Croatia, Ecuador, El Salvador, Oman and China. Some of these countries also had fertility rate reductions, which may have contributed [112], but the likely explanation is national focus on improved obstetric and neonatal care, and systematic establishment of referral systems with higher capacity of neonatal care units and staff and equipment helped in some cases by larger national budgets [2]. Over time, as neonatal care increases in scope, people skills, commodities and equipment become more critical and at a NMR below 15 per 1,000 live births, intensive care plays an increasing role. Hence low- and middle-income countries should be able to halve the risk of their newborns, their most vulnerable citizens, of dying with the right people and the right basic commodities. Yet human resource planning has not addressed this key need, and courses for nurse training in neonatal care are rare in sub-Saharan Africa and much of South Asia. Investing in frontline workers and skills is crucial to overcoming nervousness of many workers when looking after tiny babies, and building their lifesaving skills. A phased approach, for example using KMC as an entry point to show that babies under 1,000g at birth can and do survive and thrive can be a turning point for clinical staff as well as also hospital management.

Starting from existing program platforms at community level (e.g. home visit packages, women's groups) and at facility level to ensure effective care for all births at health facilities, is cost effective and more likely to show early results. However whilst families remain unreached, for example because of financial barriers to facility birth care, these gaps often mean those most at risk are unreached.

Action for preterm birth will start from increased visibility and recognition of the size of the problem -- deaths, disability, later chronic disease, parent suffering, and wider economic loss (Table 6). In many higher-income countries, visibility is driven by empowered parents, professionals or a synergy of the two (Figure 8). Parents of premature babies are both those who experience the greatest pain and those who hold the greatest power for change. Societal mobilisation has made it unacceptable for women to die while giving birth. The voice of women and families in low-income countries is yet to be mobilised for the issue of newborn deaths and stillbirths, and these deaths too often continue to be accepted as the norm despite the existence of highly cost-effective and feasible solutions.

Invest and plan Assess and advocate for newborn and preterm baby care, mobilise parent power • Review existing policies and programs to integrate high-impact care for premature babies • Train nurses for newborn care and include skilled personnel for premature baby care in human resource planning for all levels of the health system where babies are cared for • Ensure essential equipment and commodities are consistently available Implement Seize opportunities through other programs including For all facility births ensure: • immediate essential newborn care and neonatal resuscitation if needed • infection prevention and management At community level scale up: • Pregnancy and postnatal home visits, including behaviour change messages for families, as well as identification, extra care and referral for premature babies, • Breastfeeding promotion through home visits, well baby clinics, baby friendly hospital initiative Reach high coverage with improved care for premature babies especially • Kangaroo Mother Care and improved feeding for small babies • Antenatal corticosteroid use • Respiratory distress syndrome support, safe oxygen use • Audit and quality improvement processes • Provide family support Where additional capacity consider: • Additional neonatal care such as CPAP, • Referral level neonatal intensive care, with safeguards to ensure the poor can also access this care Careful attention to follow up of premature babies (including extremely premature babies) and early identification of impairment Inform and improve program, coverage and quality • Improve the data including morbidity follow up and use this in programmatic improvement e.g. gestational specific survival, rates of retinopathy of prematurity etc. • Address key gaps in the coverage data especially for Kangaroo Mother Care Innovate and undertake research • Establish prioritised research agenda with emphasis on implementation • Invest in research and in research capacity • Conduct multi-country studies of effect, cost and "how to" and disseminate findings linked to action

Source: Born Too Soon, Chapter 5 [113].

Globally, progress is being made in reducing maternal deaths and child death after the first month of life. Progress for neonatal deaths is slower. Severe neonatal infection deaths may possibly be reduced through "trickle down" from child health programs. Neonatal deaths due to intrapartum complications ("birth asphyxia") are also beginning to decline, although slowly, perhaps related to increased investments in care at birth and maternal health and care. However the over 1 million deaths among premature babies are less likely to be reduced though "trickle down" from other programs, and indeed it was the specific vulnerability and needs of the premature baby that catalyzed the specialty of neonatology. There are simple solutions that will reduce deaths among premature babies immediately for the poorest families at home in the lowest income settings -- for example promotion of early and exclusive breastfeeding, and handwashing, chlorhexidine cord applications and skin-to-skin care. Women's groups and other community mobilisation approaches are key; however, higher-impact care in facilities is also needed, such as KMC, feeding support and KMC and management of infections and respiratory complications and this is dependent on nurses and others with skills in caring for small babies, as well as more innovative technology, and can be phased over time to add increased complexity. Starting with intensive care will fail if simple hygiene, careful attention to feeding and other basic building blocks are not in place. Many countries cannot afford to rapidly scale up neonatal intensive care but no country can afford to delay doing the simple things well for every baby and investing extra attention in survival and health of newborns especially those who are preterm.

The time of JEL and MVK was funded by a grant from Bill & Melinda Gates Foundation to Save the Children's Saving Newborn Lives programme. The Born Too Soon report was funded by March of Dimes, the Partnership for Maternal, Newborn and Child Health and Save the Children.