OBJECTIVES: This study aimed to determine EnCPAP rates in 36 neonatal intensive care units of the Malaysian National Neonatal Registry (MNNR) in 2013, to compare the outcomes of VLBW neonates with and without EnCPAP, and to determine whether the availability of CPAP facilities and unit policies played a significant role in EnCPAP rates.

METHODS: First, a retrospective cohort study was conducted of VLBW neonates born in the hospitals participating in the study without major congenital abnormalities in the MNNR. This was followed by a questionnaire survey of these hospitals focussed on CPAP facilities and unit policies.

RESULTS: Of the 2,823 neonates, 963 (34.1%) received EnCPAP. Amongst EnCPAP neonates significantly fewer deaths were recorded (10.9 vs. 21.7%; p < 0.001), less bronchopulmonary dysplasia was observed (BPD; 8.0 vs. 11.7%; p = 0.002) and fewer mechanical ventilation days were necessary (p < 0.001) than in non-EnCPAP neonates. Logistic regression analysis showed that EnCPAP was significantly associated with a lower mortality (adjusted OR 0.623; 95% CI 0.472, 0.824; p = 0.001) and BPD among survivors (adjusted OR 0.585; 95% CI 0.427, 0.802; p = 0.001). The median EnCPAP rate of the 36 hospitals was 28.4% (IQR 14.3-38.7). Hospitals with CPAP facilities in the delivery suites (p = 0.001) and during transport (p = 0.001) and a policy for EnCPAP (p = 0.036) had significantly higher EnCPAP rates.

METHODS: We generated updated estimates of child mortality in early neonatal (age 0-6 days), late neonatal (7-28 days), postneonatal (29-364 days), childhood (1-4 years), and under-5 (0-4 years) age groups for 188 countries from 1970 to 2013, with more than 29,000 survey, census, vital registration, and sample registration datapoints. We used Gaussian process regression with adjustments for bias and non-sampling error to synthesise the data for under-5 mortality for each country, and a separate model to estimate mortality for more detailed age groups. We used explanatory mixed effects regression models to assess the association between under-5 mortality and income per person, maternal education, HIV child death rates, secular shifts, and other factors. To quantify the contribution of these different factors and birth numbers to the change in numbers of deaths in under-5 age groups from 1990 to 2013, we used Shapley decomposition. We used estimated rates of change between 2000 and 2013 to construct under-5 mortality rate scenarios out to 2030.

FINDINGS: We estimated that 6·3 million (95% UI 6·0-6·6) children under-5 died in 2013, a 64% reduction from 17·6 million (17·1-18·1) in 1970. In 2013, child mortality rates ranged from 152·5 per 1000 livebirths (130·6-177·4) in Guinea-Bissau to 2·3 (1·8-2·9) per 1000 in Singapore. The annualised rates of change from 1990 to 2013 ranged from -6·8% to 0·1%. 99 of 188 countries, including 43 of 48 countries in sub-Saharan Africa, had faster decreases in child mortality during 2000-13 than during 1990-2000. In 2013, neonatal deaths accounted for 41·6% of under-5 deaths compared with 37·4% in 1990. Compared with 1990, in 2013, rising numbers of births, especially in sub-Saharan Africa, led to 1·4 million more child deaths, and rising income per person and maternal education led to 0·9 million and 2·2 million fewer deaths, respectively. Changes in secular trends led to 4·2 million fewer deaths. Unexplained factors accounted for only -1% of the change in child deaths. In 30 developing countries, decreases since 2000 have been faster than predicted attributable to income, education, and secular shift alone.

INTERPRETATION: Only 27 developing countries are expected to achieve MDG 4. Decreases since 2000 in under-5 mortality rates are accelerating in many developing countries, especially in sub-Saharan Africa. The Millennium Declaration and increased development assistance for health might have been a factor in faster decreases in some developing countries. Without further accelerated progress, many countries in west and central Africa will still have high levels of under-5 mortality in 2030.