Kangaroo care: from stable babies to immediate care

Keeping a small baby skin-to-skin on a parent's chest, starting in hospital, reduces the odds of death by about 23%. That is the result of a 2025 meta-analysis of 29 randomised trials with 17,513 low-birthweight babies, most of them in lower-middle-income countries ¹. The same analysis found the odds of sepsis fell by 45% ¹.

Kangaroo mother care (KMC) needs no machine and no drug. The 15-year review follows the evidence from care after stabilisation to community initiation and immediate hospital care. These approaches need different levels of support.

The burden

Preterm birth is the leading killer of young children. WHO estimates that 13.4 million babies were born preterm in 2020 ⁷. Complications of preterm birth caused about 900,000 deaths in 2019, and they are the leading cause of death among children under 5 ⁷. When WHO launched its 2022 guideline, it put the numbers at about 15 million preterm babies and more than 20 million babies with low birthweight each year ⁶.

Newborn deaths are almost half of child deaths. In 2024, an estimated 4.9 million children died before age 5. Of these, 2.3 million were newborns in their first 28 days, which is 47% of all under-5 deaths ⁸. The UN inter-agency report released in March 2026 found that preterm birth complications caused 36% of newborn deaths ⁹.

Where the deaths are. Most preterm births happen in southern Asia and sub-Saharan Africa ⁷. Sub-Saharan Africa had 58% of all under-5 deaths in 2024 and Southern Asia a further 25% ⁸. The survival gap is large. More than 90% of extremely preterm babies (born before 28 weeks) die in low-income countries, against fewer than 10% in high-income settings ⁷.

The trend is slowing. Under-5 deaths have fallen by more than half since 2000 ⁹. But since 2015 the pace of reduction has slowed by more than 60% ⁹. The total has stayed at about 4.8–4.9 million a year since 2022 ⁸.

Most of these deaths can be prevented. WHO says that three-quarters of preterm deaths could be prevented with current, cost-effective interventions ⁷. It names kangaroo mother care immediately after birth among the simple steps that can substantially reduce deaths in preterm and low-birthweight babies ⁷.

How it works

What KMC is. The baby wears only a hat and a nappy. The baby lies on its front, skin-to-skin on the mother's (or another caregiver's) bare chest, held in place with a cloth wrap ². The aim is for this to continue for as many hours as possible, day and night, together with breastfeeding or breast milk feeding ³. In the Indian study below, "effective KMC" was defined as at least 8 hours of skin-to-skin care and exclusive breastfeeding in the 24 hours before discharge ³.

Why it helps. The parent's body keeps the baby warm. Close contact supports breastfeeding. The newest pooled evidence suggests that a large part of the benefit comes through fewer infections ¹. That is why the Basel team argues that kangaroo care should be part of standard infection prevention and control ¹.

What changed in 2022: "immediate" KMC. Until recently, small babies were first kept in an incubator or under a radiant warmer. KMC started only after the baby was stable, which could take around 3 to 7 days ⁶. In November 2022, WHO changed its advice. Its new guideline says that skin-to-skin contact with a caregiver should start immediately after birth, without any initial period in an incubator ⁶. This was WHO's first update to its preterm and low-birthweight guidelines since 2015 ⁶.

Three terms you will see.

• Conventional KMC starts after the baby is stable.
• Immediate KMC (iKMC) starts within hours of birth. In the WHO trial, babies of 1.0–1.799 kg began before full stabilisation ⁴.
• KMC before stabilisation is the Ugandan version. It included babies still getting oxygen, CPAP, intravenous fluids, antibiotics or anti-seizure drugs ².

What a health worker does. Teach the mother how to hold the baby, and why. Give her a wrap and an adjustable bed so she can keep the baby skin-to-skin for long periods ². Record the hours each day. Keep giving the baby the other care it needs: warmth when not in KMC, feeding support, and treatment of infection or breathing problems ².

What the evidence shows

2016–2019: established care and a community trial

The 2016 Cochrane review included 21 trials, mostly in babies stabilised before KMC began. Mortality at discharge or 40–41 weeks' postmenstrual age was lower with KMC (risk ratio 0.60, 95% CI 0.39–0.92; eight trials, 1,736 infants) ¹⁴. This is an earlier synthesis of overlapping evidence, not an additional population to add to the newer meta-analysis.

A 2019 randomised trial in Haryana, India, enrolled 8,402 stable, feeding babies weighing 1,500–2,250 g at home within 72 hours of birth. Repeated home visits supported skin-to-skin contact and exclusive breastfeeding. Mortality from enrolment to 28 days fell (hazard ratio 0.70, 95% CI 0.51–0.96), as did mortality to 180 days (0.75, 0.60–0.93) ¹⁵. This supports community initiation for the studied babies with home-visit support. It does not justify managing an unstable newborn at home or show an effect on deaths before enrolment. The immediate-KMC hospital trials below address a different population.

The pooled evidence (Minotti et al., 2025)

Who. A team at the University Children's Hospital Basel led a systematic review and meta-analysis ¹. It was funded by the European Society for Paediatric Infectious Diseases, the EU and the Machaon Foundation ¹.

Design. The team searched for randomised trials published from January 2013 to February 2025 that compared hospital-started kangaroo care with conventional care in low-birthweight babies ¹. They found 29 trials with 17,513 babies, mostly from lower-middle-income countries. Most trials were of moderate to high quality ¹. The OMWaNA trial from Uganda (below) is included ¹.

Primary outcomes. Compared with conventional care, kangaroo care was linked to:

• All-cause mortality: odds ratio 0.77 (95% CI 0.67–0.89), from 25 trials, rated high-quality evidence, with no variation between trials (I² = 0%) ¹.
• Sepsis: OR 0.55 (0.37–0.82), moderate-quality evidence ¹.
• Invasive infection: OR 0.49 (0.33–0.74), moderate-quality evidence ¹.

Secondary outcomes. Infection-related death fell (OR 0.63, 0.47–0.84, high-quality evidence), as did hypothermia (OR 0.28, 0.16–0.46) and apnoea (OR 0.46, 0.25–0.85) ¹.

In plain terms, babies held skin-to-skin stay warmer and get serious infections about half as often ¹.

Caveats. The sepsis and hypothermia results varied between trials (I² = 53% and 72%) ¹. The authors found that this variation came from how much kangaroo care the control groups already got as part of usual care ¹.

The trial behind WHO's advice (WHO iKMC Study Group, 2021)

Who and where. WHO coordinated this trial in five hospitals in Ghana, India, Malawi, Nigeria and Tanzania. It was funded by the Bill & Melinda Gates Foundation ⁴. It is a central study within this review and helped inform the 2022 guideline and newer trials.

Design. 3211 babies weighing 1.0–1.799 kg were randomised to immediate KMC or to conventional care in an incubator or radiant warmer until stable, with KMC after that ⁴.

Results. Babies in the immediate KMC group had a median of 16.9 hours of skin-to-skin contact a day in the neonatal intensive care unit, against 1.5 hours in the control group ⁴. Deaths in the first 28 days were 12.0% against 15.7% (relative risk 0.75, 95% CI 0.64–0.89) ⁴. Deaths in the first 72 hours were 4.6% against 5.8% (RR 0.77, 0.58–1.04), which was not a significant difference ⁴. The trial was stopped early because of the lower mortality with immediate KMC ⁴.

Caveats. The hospitals were tertiary centres with neonatal intensive care ². The OMWaNA authors note that the effect seemed to be driven mainly by the trial site in India ².

Uganda: KMC before stabilisation (OMWaNA, Tumukunde et al., 2024)

Who and where. The MRC/UVRI and LSHTM Uganda Research Unit ran the OMWaNA trial (NCT02811432) in the neonatal units of five government hospitals across Uganda ². These were WHO "level 2" units, not intensive care ². Before the trial, the hospitals got infrastructure improvements, equipment and staff training ². The trial was funded by the UK Joint Global Health Trials scheme and the US National Institute of Child Health and Human Development ².

Design. It was an individually randomised trial of 2221 babies aged under 48 hours and weighing 700–2000 g, who were not yet stable ². Babies in the intervention group started KMC as soon as possible. Control babies stayed in an incubator or under a radiant heater until stable ². Recruitment ran from October 2019 to July 2022 ².

Primary outcome. Seven-day mortality was 7.5% in both arms (81 of 1083 vs 83 of 1102; adjusted RR 0.97, 95% CI 0.74–1.28) ².

Secondary outcomes. At 28 days, mortality was 11.3% with KMC and 12.8% with standard care (RR 0.88, 0.71–1.09; p = 0.23) ². Hypothermia at 24 hours fell from 53.1% to 40.9% (adjusted RR 0.76, 0.70–0.83) ². Daily weight gain at 28 days also improved ². There was evidence of a larger effect on 28-day mortality among babies born at 28 weeks or earlier ².

Delivery. Skin-to-skin contact started within 24 hours in 82.7% of the intervention group. The median was 10.1 hours a day, against 0 hours in the control group ². That is less than the 16.9 hours in the WHO trial ⁴.

Cost. The intervention had a 97% probability (provider view) and 84% probability (societal view) of being more cost-effective than standard care, even when no value was given to deaths averted ².

A whole district in India (Thomas et al., 2024)

Who and where. The state government of Karnataka worked with two local institutions, St John's Research Institute in Bengaluru and the Karnataka Health Promotion Trust, to scale up KMC across Koppal district ³. Koppal has 1.53 million people and poorer health indicators than the rest of the state ³. Its newborn mortality rate was 24 per 1000 live births in 2017 ³.

Design. This was a prospective birth cohort within an implementation research project. It covered 23,667 live births from March to December 2018, of whom 1152 (4.9%) weighed under 2000 g ³. Babies were born in 90 public and private facilities or at home ³. Babies were stabilised first and then started on KMC, so this was conventional KMC, not immediate KMC ³.

Coverage. KMC was started in 816 babies (71%) of these small babies, and in 88% of those eligible ³. This shows that high coverage is possible across a district under real conditions, not only in a trial hospital.

Outcomes. Among 874 babies followed to 29 days, 6.4% of those who started KMC died, against 74.8% of those who did not ³. That gap is not a measure of the effect of KMC (see below). The coverage is the real result.

How. Mothers got a KMC kit with a wrap, a bag to hold the baby, a cap, socks, a breast milk cup and a paladai, a traditional infant feeding cup ³. Staff, equipment and supplies came from within the existing health system, to make it last ³.

Results at a glance

LimitsWhat didn't work (or the limits)

OMWaNA found no significant difference in deaths. Seven-day mortality was 7.5% in both arms, and the 12% relative fall in 28-day deaths was not significant ². This was the first large African trial of KMC started before stabilisation. The result suggests that in hospitals with weak basic newborn care, KMC alone is not enough to show a clear drop in deaths.

Why OMWaNA may have missed an effect. The trial assumed a baseline death rate of 25% ². Deaths in both arms were lower than expected, which reduced the trial's power ². Babies also got fewer hours of skin-to-skin contact than in the WHO trial ²⁴. The authors conclude that safe respiratory support and other newborn care need to be in place before KMC is started in unstable babies ².

Early mortality remains a distinct question. The positive community trial measured deaths only after enrolment ¹⁵. The WHO trial found no significant difference at 72 hours ⁴. OMWaNA and a trial in The Gambia also found no fall in early deaths ².

The Indian district numbers are observational. Sicker babies were less likely to start KMC. So the very large difference in death rates (adjusted RR 0.05) is inflated by confounding and must not be read as the effect of KMC ³. Overall newborn mortality among the small babies was 18% (213 deaths), and 56% of those deaths happened in the first 3 days ³. Some babies die before KMC can start.

Coverage is not the same as enough hours. In the earlier Karnataka coverage study, 85% of babies got "any KMC", but only 53% got "effective KMC" at discharge ³.

No new district-hospital results yet. The strongest mortality trial (the WHO trial) was done in tertiary centres ⁴. Evidence from the smaller district hospitals where many small babies are born is still limited. The PRISM trial in India is designed to fill this gap (see below) ¹⁰.

Cost and scale

KMC does not cost more than incubator care. Across the five OMWaNA hospitals, the mean cost to society per baby was US$359.1 with KMC and US$365.9 with standard care (adjusted difference −$7.2, 95% CI −21.5 to 7.2) ². Costs to the health provider were US$273.3 and US$282.4 ². Staff time and electricity cost less with KMC. This more than balanced slightly higher costs for setup, water, drugs, supplies and transport ².

Earlier studies agree. The OMWaNA authors note that economic studies in South and Central America, sub-Saharan Africa, south Asia and Europe all found KMC saved money for health providers ².

What is needed to deliver it. The OMWaNA hospitals needed infrastructure, equipment and training before the trial, and adjustable beds so mothers could keep babies skin-to-skin ². In Karnataka, the district used its own staff and supplies and gave mothers a simple kit ³.

Coverage at scale. Koppal district reached 71% KMC initiation among all babies under 2000 g, and 88% among those eligible ³. That was across 90 facilities, both public and private ³.

Who funds and delivers. Delivery in these studies was by government hospitals and district health systems, in Uganda and in Karnataka ²³. Research funding came from the UK government and Wellcome Trust scheme, the US NIH, the Gates Foundation and European sources ¹²⁴.

What to watch

India: iKMC in district hospitals (PRISM, CTRI/2024/09/074269). This stepped-wedge cluster randomised trial will run in ten district hospitals in Chhattisgarh over 42 months ¹⁰. It will include babies of 1000–1799 g in whom KMC can start within 12 hours of birth ¹⁰. The primary outcome is death or culture-positive sepsis by 28 days ¹⁰. The protocol says the intervention was due to start sequentially after a 6-month baseline, in ten steps three months apart ¹⁰. Results are therefore several years away.

Pakistan: community KMC (NCT03545204). Researchers at Aga Khan University designed a cluster randomised trial of a community KMC package in one rural district ¹¹. It includes stable babies weighing 1200–2500 g, with KMC kits, counselling and village "KMC champions" ¹¹. The primary outcome is newborn mortality. Data collection was planned to end in December 2021, and a correction to the protocol was published in July 2026, but we have not found published results ¹¹.

Europe: kangaroo care as infection control (NeoIPC, NCT05993442). The NeoIPC project, which includes the Basel team behind the 2025 meta-analysis, is testing optimised kangaroo care in 24 neonatal intensive care units in five European countries ¹². It will measure severe infections and colonisation with resistant bacteria ¹². If it works, it will strengthen the case that KMC is an infection-control tool in every setting.

Zambia: longer stays on the KMC ward (NCT06545760). The University Teaching Hospital in Lusaka, the University of Zambia and the University of Alabama at Birmingham are testing whether a longer stay on the KMC ward prevents postpartum depression in mothers of low-birthweight babies. The trial plans 1908 participants ¹³. It will also look at child development at 6, 12 and 18 months, and at cost-effectiveness ¹³.

Open questions. OMWaNA's authors list the research priorities: how to give KMC to unstable babies in the settings with the highest death rates, at different levels of the health system, and how to track national coverage and quality through routine data ².

What it means

Hospital and community trials support KMC for defined groups of small babies ⁴¹⁵, while OMWaNA shows that measured mortality benefits are not uniform across delivery settings ². The newer meta-analysis also supports a reduction in serious infection ¹. Koppal district shows that 70–90% coverage is achievable at scale ³.

The Uganda trial shows the limit. KMC works best as part of a stronger unit for small and sick newborns, with warmth, feeding support, infection prevention, safe breathing support and staff who can manage early complications ². Because more than half of deaths in the Indian cohort happened in the first 3 days, the first hours matter most ³.

For ministries of health. Make immediate KMC a routine part of every newborn unit, in line with WHO's 2022 guideline ⁵⁶. Budget for beds, wraps and space for mothers, and for the basic equipment a level 2 unit needs ². Add KMC coverage and hours to routine data, so that scale-up can be measured ².

For health workers. Start skin-to-skin contact early and aim for long hours each day. The WHO trial reached a median of 16.9 hours ⁴. Keep giving the other care a small baby needs: warmth when not in KMC, breastfeeding support, and fast treatment for infection or breathing problems ².

For funders. KMC is one of the cheapest ways to reduce newborn deaths, and it cuts infections ¹². It works best when paired with investment in basic newborn care, so fund both together ².