A good start in life begins in the womb
In 2015, 20.5 million newborns, an estimated 14.6 per cent of all babies born globally that year, suffered from low birthweight. These babies were more likely to die during their first month of life and those who survived face lifelong consequences including a higher risk of stunted growth,1 lower IQ,2 and adult-onset chronic conditions such as obesity and diabetes.3 To grow a healthy baby, mothers need good nutrition and rest, adequate antenatal care, and a clean environment. Together, these ingredients for a healthy pregnancy can help to prevent, identify and treat the conditions that cause low birthweight and thus foster achievement of the World Health Assembly (WHA) nutrition target to reduce low birthweight by 30 per cent between 2012 and 2025.
Progress on reducing low birthweight prevalence has been limited in all regions. In Latin America and the Caribbean and Western Europe, there was no apparent change in low birthweight prevalence between 2000 and 2015, while in North America, there was a statistically significant, albeit small, increase over the same period (7.3% [7.2-7.5] in 2000 and 7.9% [7.8-8.0] in 2015). No region has experienced a statistically significant decrease in low birthweight prevalence during this 15-year period. At the global level, the annual average rate of reduction (AARR) in low birthweight prevalence was only 1.2 per cent between 2000 and 2015, yet an ARRR of 2.74 per cent between 2012 and 2025 is required to meet the WHA low birthweight target by 2025.
Progress has been slow in reducing the number of low birthweight babies with 22.9 million affected in 2000 and 20.5 million in 2015. In 2015, approximately three quarters of all low birthweight newborns in the world were born in just 3 regions: South Asia (47 per cent of all low birthweight births worldwide), Eastern and Southern Africa (13 per cent of all low birthweight births worldwide) and West and Central Africa (12 per cent of all low birthweight births worldwide). While there seems to have been some increase in the number of low birthweight babies between 2000 and 2015 in regions including West and Central Africa and Eastern and Southern Africa, the only region to have experienced a statistically significant increase was Northern America (315,684 [310,373 – 320,996] in 2000 and 345,743 [341,630-349,856] in 2015).
Birthweight data were not available for 29.1 per cent of newborns in 2022. These estimates reflect newborns who were unweighed and those who were weighed but whose birthweights were not captured by key data sources. Estimates of newborns without birthweight data from administrative systems (e.g., Health Management Information Systems) include unweighed births and weighed births not recorded in the system. Estimates from household surveys include births where weight was not available from an official document (e.g., health card) or could not be recalled by the respondent at the time of interview. West and Central Africa was home to the highest percentage of newborns without a recorded birthweight (54.1 per cent of births). Among countries with recent data, 8 had a percentage of unrecorded birthweights greater than 70 per cent, rendering the data from these countries unsuitable for generating low birthweight estimates.
References
Christian P, et al. Black RE. Risk of childhood undernutrition related to small-for-gestational age and preterm birth in low- and middle-income countries. International Journal of Epidemiology 2013;42:1340–55.
Gu H, Wang L, Liu L, et al. A gradient relationship between low birth weight and IQ: A meta-analysis. Sci Rep. 2017;7(1):18035. Published 2017 Dec 21. doi:10.1038/s41598-017-18234-9
Jornayvaz FR, Vollenweider P, Bochud M, Mooser V, Waeber G, Marques-Vidal P. Low birth weight leads to obesity, diabetes and increased leptin levels in adults: the CoLaus study. Cardiovasc Diabetol. 2016; 15: 73.
Blencowe H, Krasevec J, de Onis M, Black R E, An X, Stevens G A, Borghi E, Hayashi C, Estevez D, Cegolon L, Shiekh S, Hardy V P, Lawn J E, Cousens S. National, regional, and worldwide estimates of low birthweight in 2015, with trends from 2000: a systematic analysis.
Low birthweight
Low birthweight data
Build and download your own customisable dataset
Resources




Notes on the data
Definitions
Two key indicators about birthweight are described in this section, (i) low birthweight prevalence and (ii) prevalence of newborns without birthweight data.
Birthweight is the first weight of the newborn obtained after birth. For live births, birthweight should preferably be measured within the first hour of life, before significant post-natal weight loss has occurred. Low birthweight is defined as less than 2,500 grams (up to and including 2,499 grams).
The prevalence of newborns without birthweight data reflects newborns who were unweighed and those who were weighed but whose birthweights were not captured by key data sources. Estimates of newborns without birthweight data from administrative systems (e.g. Health Management Information Systems) include unweighed births and weighed births not recorded in the system. Estimates from household surveys include births where weight was not available from an official document (e.g. health card) or could not be recalled by the respondent at the time of interview.
Indicator | Numerator | Denominator |
Prevalence of low birthweight | The number of live births that weigh less than 2,500 grams in a given time-period[1] | Total number of live births in the same time-period |
Prevalence of newborns without birthweight data
(from household surveys)
|
Total number of live births for which a birthweight was not available from an official document (e.g. health card) or could not be recalled by the respondent at the time of interview. | Total number of live births in the survey sample. |
Prevalence of newborns without birthweight data
(from administrative sources)
|
Total number of live births which were not weighed and weighed births not recorded in the administrative system in a calendar year. | Total number of live births captured in the administrative data source in a calendar year.
|
Data sources and methods
Nationally representative estimates of LBW prevalence can be derived from a range of sources, broadly defined as administrative data or nationally representative household surveys. National administrative data include those from national systems including Civil Registration and Vital Statistics (CRVS) systems, national Health Management Information Systems (HMISs), and birth registries. Nationally representative household surveys include Demographic and Health Surveys (DHSs), Multiple Indicator Cluster Surveys (MICSs), and other national surveys for which raw data are available to assess data quality and undertake adjustment for missing birthweights and heaping.
The UNICEF/WHO estimates were derived using the following types country input data (i) administrative data covering ≥90 per cent of the UN estimated livebirths and with an associated facility birth rate of ≥90 per cent (labelled high coverage administrative estimates); (ii) administrative data covering ≥80 but <90 per cent of UN estimated livebirths and with an associated facility birth rate of ≥80 per cent (labelled moderate coverage administrative estimates) and (iii) household surveys: (a) that were nationally representative; (b) that had publicly available raw data; (c) that had birthweight data; and (d) that had maternal perception of size at birth data (or in absence of size at birth data, had a birthweight for ≥99 per cent of births in the sample), (e) that met all inclusion criteria (Table 1); and (f) for which an estimate adjusted for missing birthweights and heaping could be derived.
Table 1: Survey quality inclusion criteria
Indicator | Inclusion criteria |
Percentage of births with a birthweight in the dataset | ≥30 per cent |
Total number of births with a birthweight in the dataset | ≤200 |
No indication of severe heaping or implausible distribution among births with a birthweight in the dataset | a) ≤55% of all birthweights fell on the three most frequent birthweights (i.e. if 3,000g, 3,500g and 2,500g were the three most frequent birthweights, when added together, they made up ≤55 per cent of all birthweights in the dataset);
b) ≤10 per cent of all birthweights weighed ≥4,500g; c) ≤5 percent of birthweights on tail ends of 500g and 5,000g |
Methods applied to generate annual country estimates varied by availability and type of country input data as follows:
- b-spline: data for the 57 countries with ≥8 data points from higher coverage administrative sources with ≥1 prior to 2005 and ≥1 more point more recent than 2010 were smoothed with b-spline regression to generate annual low birthweight estimates. A b-spline regression model was used to predict the standard error and calculate 95 per cent confidence intervals for the country-level low birthweight estimates. These low birthweight estimates follow very closely those included in the countries’ own administrative reports.
- hierarchical regression: data for the 91 countries not meeting requirements for b-spline but with ≥1 LBW data point from any source meeting inclusion criteria were fitted into a model using covariates to generate annual LBW estimates, as well as uncertainty ranges, using a bootstrap approach. The model included (natural log) of neonatal mortality rate; the proportion of children underweight (weight for-age z score below minus two standard deviations from median weight for age of reference population); data type (higher quality administrative, lower quality administrative, household survey); UN region (e.g., Southern Asia, Caribbean); and a country-specific random effect. These LBW estimates may vary substantially from estimates reported by countries in administrative and survey reports
- no estimate: the 54 countries for which LBW input data were not available and/or did not meet inclusion criteria are indicated in the database as “no estimate
Global estimates were derived by summing the estimated number of live births weighing less than 2,500g for the 195[2] countries in the United Nations regional grouping for each year and dividing by all live births in each year in those 195 countries. Regional estimates were similarly derived, based on countries in each regional grouping. To obtain the global and regional level estimates of uncertainty, 1,000 low birthweight point estimates were made for each country for each year using either b-spline (by randomly sampling from a normal distribution plotted using the calculated standard error) or hierarchical regression approach (using a bootstrap approach). The country low birthweight estimates for each of the 1000 samples were summed at worldwide or regional level and the 2.5th and 97.5th centiles of the resulting distributions were used as the confidence intervals. For further details on methods, please refer to Blencowe et al. 2019.
[1] Usually a calendar year from administrative sources and multiple years from household surveys
[2] The UNICEF-WHO low birthweight estimates are presented for various regional groupings of which the UNICEF regional grouping has the largest number of countries (n=202). Seven of the 202 countries did not have LBW input data or covariate data. It was therefore not possible to generate any estimates for these seven countries or include them in the regional and global estimates which are based on a total of 195 countries.