The study collected standardized disease definitions, available data, and prevalence information on children younger than 5 years in the provinces and cities of China. The study included data from the Institute for Health Metrics and Evaluation (IHME) gathered by LBD CGF collaborators, which were available in public online repositories, data that were publicly available upon request from the data provider, and data that were not publicly available due to restrictions by the data provider and which were used under license for this study. More information about each data source is available on the Global Health Data Exchange (GHDx) [5,16,17]. The analysis covered relevant metrics in national, 33 provincial-level, and 362 municipal-level administrative divisions in mainland China, Hong Kong, and Macao from 2000 to 2019. Data were analyzed from October 17, 2023, to November 22, 2023.

LBD compiled an extensive geopositioned dataset for the study, which included data from 460 household surveys and reports representing 4.6 million children. The data covered estimates of annual prevalence in 105 LMICs for each year between 2000 to 2019. The collaborators first comprehensively mapped high spatial-resolution estimates of CGF indicators from 2000 to 2017 across 105 LMICs, providing an accurate public health tool to support efficient targeting of local-level interventions to vulnerable populations. Their maps identified high-prevalence areas even within nations otherwise succeeding in reducing overall CGF prevalence. They highlighted where the highest-need populations reside to support decision makers in planning interventions adapted locally and in efficiently directing resources toward reducing CGF and its health implications [5].

CGF, expressed as stunting, wasting, and underweight in children younger than 5 years (age 0-59 months), is a specific subset of malnutrition characterized by insufficient height or weight against age-specific growth reference standards [5]. Owing to the fact that WHO Child Growth Standards can reflect the true growth status and potential of children younger than 5 years and that the reference population is well representative, we mainly adopted them based on the comparability requirements between studies in China.

Height/length-for-age z-score (HAZ) can be used to assess the chronic malnutrition status of children. Stunting is defined as a HAZ that is more than two SDs below the median of the WHO standards for child growth [18].

Weight-for-height/length z-score (WHZ) can be used to investigate acute malnutrition in children. Wasting is defined as a WHZ that is more than two SDs below the median of the WHO standards for child growth. Severe wasting is defined as a WHZ that is more than three SDs below the median of the WHO standards for child growth [18].

Weight-for-age z-score (WAZ) can be used to determine the short-term and long-term nutritional status of children. Underweight is defined as a WAZ that is more than two SDs below the median of the WHO standards for child growth [18].

The prevalence of stunting, wasting, or underweight in children younger than 5 years in China was the main indicator used to reflect the health situation of children. Each prevalence is reported with 95% uncertainty interval (UI) according to the LBD algorithm. There are 4 age groups among children younger than 5 years, including early, late, and postneonatal infants, and children aged 1 to 4 years. All relevant metrics were estimated separately for the 2 sexes and various age groups (ie, 0-6 days, 7-27 days, 28 days through 11 months, 1-4 years, and 0-5 years) from 2000 to 2019. We used Excel 2017 (Microsoft Corp) to organize the data, established time series data files in CSV format between 2000 to 2019, and used Joinpoint Regression Software (Version 5.0.2; National Cancer Institute) for statistical analysis. In addition, Joinpoint regression (JPR) program was used to estimate the annual prevalence of CGF in China. Joinpoint is a statistical software for the analysis of trends using Joinpoint models, usually used to examine changes in trends. The reason we chose JPR was that JPR has certain advantages in processing long-term disease data from multiple trend segments. Moreover, JPR divides trend changes into several statistically significant trend segments through model fitting, and thus, this data processing method of segmentation is more reasonable than manual segmentation. After segmentation, researchers can clearly see the ascending segment, rapid ascending segment, gentle segment, descending segment, and rapid descending segment. Furthermore, the background and reasons for the generation of segmentation points can be analyzed, providing a basis for future medical policy formulation and resource allocation [19].

We used the normality test on the dependent variable, which followed a Poisson distribution. Therefore, we used a logarithmic linear model for analysis. The annual percentage change (APC), average annual percentage change (AAPC), and 95% CI are the main indicators for Joinpoint analysis. The number of joinpoints we set was three. APC is used to evaluate the trend of changes in various time periods, while AAPC is used to evaluate the trend of changes throughout the period, showing that the overall prevalence trend in the area increases or decreases monotonously between 2000 to 2019. If APC is <0, it indicates that the prevalence trend is decreasing year by year, while if it is >0, it indicates that the prevalence trend is increasing. All calculations were performed using Stata 17 (StataCorp). All P values were 2-sided, and P<.001 was considered statistically significant. In addition, we used the Holt-Winters model with trends but without seasonal components to predict the trends of prevalence in 2030, and the assessment was performed with Stata 17.

This study used data from the IHME, which were collected by LBD CGF collaborators, and this has been approved by the institutional review board of the University of Washington School of Medicine. Our research is based on the secondary analysis of data derived from the IHME. As this is a secondary analysis of existing data, no additional human participant research ethics review or informed consent was required. Study data were anonymized and deidentified to protect the privacy and confidentiality of the study participants. Our analysis complied with the Guidelines on Accurate and Transparent Health Estimate Reporting [20].

Overall, the estimated stunting prevalence in children younger than 5 years across China decreased by 36.9% from 19% (95% UI 17%-21%) in 2000 to 12% (95% UI 2%-27%) in 2019 (Multimedia Appendix 1). In 2019, the prevalence was similar between boys and girls (Multimedia Appendix 2). The 3 provinces with the highest prevalence were Guizhou (26%), Xizang (22%), and Hunan (20%) (Figure 1A; Multimedia Appendix 1).

From 2000 to 2019, the stunting prevalence in China showed a significant downward trend (AAPC –2.48%, 95% CI –2.62% to –2.34%; P<.001; Multimedia Appendix 1). JPR analysis showed that the steepest slope occurred in 2000-2004 (APC –4.34%, 95% CI –4.50% to –4.19%; P<.001) in both males and females. From 2004 to 2019, the rate of decline slowed down (Tables 1-4; Figure 2A; Multimedia Appendix 3). At the provincial level, the trends of stunting prevalence varied significantly across the 33 provinces (Figure 1B; Multimedia Appendix 1), with the highest AAPC in Ningxia (AAPC –5.08%, 95% UI –5.38% to –4.79%; P<.001) and the lowest AAPC in Shanghai (AAPC –0.72%, 95% UI –1.08% to –0.35%; P<.001). There were various directions in the trends of stunting in children younger than 5 years by province (Tables 1-4). Among the provinces, 25 showed significant downward trends of stunting in children younger than 5 years at all joinpoints, while the other 8 did not, and 6 provinces (Anhui, Jilin, Ningxia, Shanxi, Yunnan, and Macao) showed a gradual increase since 2017 (Tables 1-4). The disparities between their highest- and lowest-prevalence units from 2000 to 2019 reduced in most provinces, and densely populated areas may have a relatively low stunting prevalence (Multimedia Appendix 1). At the municipal level, the stunting prevalence in 98.9% (n=358) of units showed downward trends (Multimedia Appendix 4), and the 3 with the highest AAPC were Alxa (AAPC –7.23%; P<.001), Baynnur (AAPC –7.11%; P<.001), and Wuhai (AAPC –6.98%; P<.001). Four units with upward trends were Xiamen (AAPC 1.54%; P=.66), Shenzhen (AAPC 1.07%; P=.66), Zhuhai (AAPC 0.16%; P=.95), and Central and Western Hong Kong (AAPC 0.22%; P=.89).

From 2000 to 2019, the stunting prevalence declined in all age groups of children (Multimedia Appendix 5), and the prevalence was similar between boys and girls (Multimedia Appendix 6). The largest decline (AAPC –1.95%; P<.001) occurred in children aged 1 to 4 years. The decreasing trend in prevalence was the lowest among postneonatal infants (AAPC –1.95%; P<.001; Multimedia Appendix 5). The trends of stunting prevalence were similar in the early and late neonatal periods (Multimedia Appendix 5).

Childhood wasting was less widespread than stunting, affecting 4% (3%-5%) of children younger than 5 years in China in 2000 and 3% (1%-8%) by 2019 (Multimedia Appendix 1). In 2019, the prevalence was similar between boys and girls (Multimedia Appendix 2). Among the 33 provinces, the wasting prevalence was the highest in Guangxi (5%) (Figure 1C; Multimedia Appendix 1).

From 2000 to 2019, there was a slow downward trend (AAPC –0.57%, 95% CI –0.59% to –0.55%; P<.001) in the wasting prevalence, which was smoother compared to the stunting prevalence in China (Multimedia Appendix 1). The analysis showed that the greatest increase occurred in 2004-2010 (APC –0.92%, 95% CI –0.94% to –0.90%; P<.001), and from 2010 to 2019, there was a steady trend in China as a whole and in most provinces (Tables 1-4; Figure 2B; Multimedia Appendix 3). At the provincial level, the trends of wasting prevalence varied significantly across the 33 provinces (Figure 1D; Multimedia Appendix 1), with the highest AAPC in Jilin (AAPC –0.72%, 95% UI –0.74% to –0.70%; P<.001) and the lowest AAPC in Macao (AAPC –0.05%, 95% UI –0.29% to 0.19%; P=.67). There were various directions in the trends of wasting in children younger than 5 years by province (Tables 1-4). Among the provinces, 25 showed significant downward trends of wasting in children younger than 5 years at all joinpoints, while the other 8 did not, and 3 provinces (Jiangxi, Shanxi, and Macao) showed a gradual increase since 2017, while 2 (Anhui and Gansu) remained stable since 2017 (Tables 1-4). There was a noticeable increase in the wasting prevalence from 2000 to 2004 in the Hong Kong province (APC 1.91%, 95% UI 1.75%-2.07%; P<.001), with an upward trend in the 11 years spanning from 2000 to 2011 (Table 4). At the municipal level, the wasting prevalence in 90.0% (n=326) of units showed downward trends (Multimedia Appendix 4), and the 3 with the highest AAPC were Chengde (AAPC –2.18%; P=.01), Zhangjiakou (AAPC –2.11%; P=.06), and Baicheng (AAPC –1.89%; P<.001). Three units with the most significant upward trends were Jixi (AAPC 0.95%; P=.07), Yanbian Korean Autonomous Prefecture (AAPC 0.90%; P=.02), and Shuangyashan (AAPC 0.86%; P=.03), while those with stable trends were Jilin (AAPC 0.00%; P=.99) and Wong Tai Sin of Hong Kong province (AAPC 0.00%; P=.99).

The prevalence was the highest among early neonatal infants in both 2000 and 2019 (16% and 11%, respectively) (Multimedia Appendix 5). From 2000 to 2019, the wasting prevalence declined in all age groups of children (Multimedia Appendix 5), and the prevalence was similar between boys and girls (Multimedia Appendix 6). The largest decline (AAPC –1.95%; P<.001) occurred in the early neonatal period. The trends of wasting prevalence were similar in the late and postneonatal periods, as well as in children aged 1 to 4 years and those aged younger than 5 years (Multimedia Appendix 5).

During the same period, the estimated underweight prevalence in children younger than 5 years decreased by 42.9% from 7% (6%-9%) in 2000 to 4% (2%-9%) in 2019 (Multimedia Appendix 1). The 3 provinces with the highest prevalence were Xizang (9%), Guizhou (9%), and Guangxi (8%) (Figure 1E; Multimedia Appendix 1). The prevalence was similar between boys and girls in 2019 (Multimedia Appendix 2).

From 2000 to 2019, the underweight prevalence in China demonstrated a significant downward trend (AAPC –2.95%, 95% CI –3.02% to –2.87%; P<.001) (Multimedia Appendix 1). JPR analysis showed that the most substantial changes occurred in 2000-2005 (APC –4.52%, 95% CI –4.58% to –4.46%; P<.001) in China. The trend decreased steadily from 2005 to 2017 and remained stable after 2017 (Tables 1-4; Figure 2C; Multimedia Appendix 3). At the provincial level, the trends of underweight prevalence varied significantly across the 33 provinces (Figure 1F; Multimedia Appendix 1), with the highest AAPC in Sichuan (AAPC –5.36%, 95% UI –5.47% to –5.25%; P<.001) and the lowest AAPC in Hong Kong (AAPC –0.93%, 95% UI –0.98% to 0.89%; P<.001). There were various directions in the trends of underweight in children younger than 5 years by province (Tables 1-4). Most provinces showed significant downward trends of underweight in children younger than 5 years before 2010, and then, the downward trend slowed down (Tables 1-4). At the municipal level, the underweight prevalence in 95.9% (n=347) of units showed downward trends (Multimedia Appendix 4), and the 3 units with the highest AAPC were Kunming (AAPC –7.87%; P<.001), Qujing (AAPC –7.97%; P<.001), and Zhaotong (AAPC –8.69%; P<.001), which are all in Yunnan province. Three units with the most significant upward trends were Yanbian Korean Autonomous Prefecture (AAPC 1.28%; P=.003), Turfan (AAPC 0.98%; P=.63), and Changji Hui Autonomous Prefecture (AAPC 0.93%; P=.57).

The prevalence was the highest among early neonatal infants in both 2000 and 2019 (15% and 9%, respectively) (Multimedia Appendix 5). From 2000 to 2019, the underweight prevalence declined in all age groups of children (Multimedia Appendix 5), which was similar between boys and girls (Multimedia Appendix 6). The largest decline (AAPC –3.03%; P<.001) occurred in children aged 1 to 4 years. The downward trend in prevalence was the lowest among early neonatal infants (AAPC –2.36%; P<.001; Multimedia Appendix 5). The trends of underweight prevalence were similar in the late and postneonatal periods, as well as in children aged 1 to 4 years and those aged younger than 5 years (Multimedia Appendix 5).

It was estimated that the stunting prevalence in children younger than 5 years would decrease to 11.1% (boys: 11.7%, girls: 10.5%) by 2030 (Table 5; Figure 3A). China will achieve the WHO GNT (50% reduction in the number of children with stunting from the baseline in 2012) of 13.5% for stunting (Figure 3A). In fact, China had achieved the target by 2010 (Table 5) at the national level. Although China will achieve the WHO GNT for stunting, it will not meet the target (<5%) of the China National Program for Child Development (2021-2030) based on current trajectories. At the provincial level, 82% of provinces (27/33) will achieve the WHO GNT, and only 9% (3/33; Ningxia, Zhejiang, and Hong Kong) will meet the target (<5%) of the China National Program for Child Development (2021-2030) (Multimedia Appendix 7). The variation of temporary trends according to sex was similar (Table 5; Figure 3). The wasting prevalence is estimated to decrease to 3.2% (boys: 3.2%, girls: 3.2%) in 2030. China might not nationally meet the WHO GNT for wasting (<3%) by 2030 (Table 5; Figure 3B). At the provincial level, 36% of provinces (12/33) will achieve the WHO GNT for wasting (Multimedia Appendix 7). The underweight prevalence will decrease to 4.1% (boys: 4.4%, girls: 3.6%) in 2030 (Figure 3C). Only 33% (11/33) of provinces will meet the WHO GNTs for both stunting and wasting by 2030. The stunting prevalence in all age groups had achieved the 2030 WHO target in 2019. In contrast, the wasting prevalence in all age groups has not achieved the WHO target.

In this study, the prevalences of stunting, wasting, and underweight decreased by 36.9%, 25.0%, and 42.9%, respectively. In addition, the prevalence of CGF showed downward trends from 2000 to 2019 in China as a whole, in all 33 provinces, and in the vast majority of cities, which is in concordance with a previous study based on 2 national surveys between 2002 and 2013 [21].

The explanation for the observed downward trends is likely complex. The crude exclusive breastfeeding rate under 6 months was 20.7% in 2013 [22]. A survey report on the influencing factors of breastfeeding in China showed that in 2019, the exclusive breastfeeding rate under 6 months was 29.2% [23]. A number of national programs promoted healthful food choices and increased support for families and communities. For example, a series of the China National Program for Child Development (2001-2010, 2011-2020, and 2021-2030) whose implementation would provide important guarantees for the realization of children’s rights to survival, development, protection, and participation [24]. The 2000 Dietary Guidelines for Chinese helped the people make beneficial and healthy dietary choices and behavioral changes [25]. Steady progress is seen on mother’s education, the nutrition of women before and during pregnancy, the development of the economy, levels of exclusive breastfeeding, and the implementation of the policy, which are likely to have contributed to these changes [26-29]. Therefore, providing in-depth analysis of the major drivers behind these concerning trends and evidence-based policy recommendations to address them is important.

The downward trend slowed down during the assessment period. Most provinces had stagnated processes of trends after 2017. The signs of increasing hunger globally that first began to appear since 2017 may have led to the trends of CGF in this study. In 2017, China had a relatively low prevalence of national CGF in Asia [5]. Therefore, it is understandable that the downward trend has slowed down.

In first administrative-level units, most provinces reduced disparities between their highest- and lowest-prevalence units between 2000 and 2019. A possible explanation for this might be that health plans (“pilot project to improve children’s nutrition in poor areas”) have been released by the government to reduce the gap between provinces [30]. To improve the health level of children in impoverished areas, the central government has specially arranged subsidy funds for the children’s nutrition improvement project, providing Yingyangbao (YYB; complementary food supplement) with protein, vitamins, and minerals for infants and young children aged 6 to 23 months in concentrated and contiguous poor areas every day [31]. Although most provinces successfully reduced the prevalence of CGF, there remain areas with slow progress. In second administrative-level units, the declines were still uneven. The further reduction of inequality will require targeted strategies to lessen CGF in each province and will need further and much more detailed analysis.

The age group with the highest stunting prevalence was children aged 1 to 4 years, and the highest prevalence of wasting and underweight was noted in early neonatal infants. The prevalence of stunting and underweight declined significantly in all age groups of children from 2000 to 2019. The largest decline in the wasting prevalence occurred in early neonatal infants, and the downward trend among late and postneonatal infants was relatively low. A recent study showed that early postnatal stunting predisposed children to subsequent and persistent stunting [32]. Children who were wasted before the age of 6 months had faster recovery and shorter episodes than did children who were wasted at older ages. The findings of these studies suggest that the period from the prenatal stage to 6 months should be the focus of preventive wasting interventions [33,34]. Thus, early intervention is important. A guideline for healthy nurturing and care of infants and young children younger than 3 years was released in 2022, which targeted children younger than 3 years. Targeted strategies are needed to reduce CGF in children younger than 1 year, and more attention should be paid to late and postneonatal infants.

According to the predictions in this study, China has achieved the goal regarding stunting prevalence. However, the goal regarding wasting prevalence has not been achieved. A national assessment of the patterns of CGF may help policymakers to develop appropriate prevention and management strategies.

This study has some limitations. The analysis heavily depended on data from LBD. There were differences between the data estimated through complex modeling in this study and in earlier studies [15]. In addition, the data were gathered before the popularity of COVID-19. Owing to the potential impacts of the pandemic, predictions based on prepandemic data may be affected. A study showed that the relatively high stunting rate in children younger than 3 years in 2021 may be associated with the influence of COVID-19 [35]. In the future, the data based on China after the pandemic should be modeled.

Over the past 20 years, the prevalence of CGF has gradually decreased. The downward trend has slowed down, and there remain areas with slow progress. The findings of this cross-sectional study suggest that although widespread declines in CGF have occurred in China, local underachievement and stagnated processes of trends after 2017 are still present. China has nationally met the WHO GNTs for stunting but not for wasting. Furthermore, early interventions for children are crucial.

These findings can be used by governments at the subnational level to identify major problems of CGF and facilitate priority setting. The findings can help to evaluate the effectiveness of national and provincial public health policies and intervention programs in the same period based on the investigation of trends at the provincial level in China in the past 2 decades. Although densely populated areas may have relatively low prevalence of CGF, the absolute number of affected children may still be high; thus, both relative and absolute estimates are important to determine where additional attention is needed.