Background
The health status of women and children directly reflect the level of national health and social development (Lebrun-Harris et al., Reference Lebrun-Harris, Ghandour, Kogan and Warren2022). In the eight Millennium Development Goals (MDGs) advocated by the United Nations, two MDGs are directly related to maternal and child health (MCH): MDG 4 (a two-third reduction in child mortality between 1990 and 2015) and MDG 5 (a three-quarter reduction in maternal mortality ratio between 1990 and 2015) (Zhang et al., Reference Zhang, Shi, Li and Bian2020). Being a United Nations member, the Chinese government formulated a series of policies and goals, such as the “Healthy China 2030” plan (2016) (Ning et al., Reference Ning, Pei, Huang, Li and Shao2024), the China National Program for Women’s Development (2021–2030) and the China National Program for Child Development (2021–2030) (Chen et al., Reference Chen, Li and Harmer2019), which place higher demands for MCH than set by the United Nations. Many studies found the temporal trends of MCH indicators in China have a remarkable improvement over the past few decades (Qiao et al., Reference Qiao, Wang, Li, Jiang, Zhang, Ma, Song, Ma, Fu and Pang2021; Yip et al., Reference Yip, Fu, Jian, Liu, Pan, Xu, Yang and Zhai2023). For example, MDG 4, which aims to reduce child mortality by two-thirds, has been reached in advance of 9 years; MDG 5, which seeks to decrease maternal mortality by three-quarters, has been achieved 1 year ahead of schedule (Campbell, Reference Campbell2017; Li et al., Reference Li, Zhang, Fang, Liu, Liu, Li, Liang and Fu2017).
However, unequal and inefficient MCH remains a persistent issue in China, particularly in rural areas (Gebremeskel et al., Reference Gebremeskel, Udenigwe, Etowa and Yaya2023; Yu et al., Reference Yu, Wang, Kang, Miao, Song, Ran, Zhu, Liang, Li and Dai2022). In rural China, the number of “left-behind women” has reached 47 million, with 59.04% of the 75.53 million children under the age of 5 residing in rural areas (Sun et al., Reference Sun, Zhang, Xu, Guo, Zhang, Xu, Jin, Zao and Huang2020; Xue et al., Reference Xue, Zhang, Zhang and Liu2021). There are large differences in the use of MCH between richer and poorer regions, between urban and rural areas. In 2018, the under-five mortality rate in urban areas was 4.4‰, while in rural areas it reached 10.2‰. Additionally, the maternal mortality rate in urban areas was 15.5 per 100,000, whereas in rural areas it was 19.9 per 100,000 (Dai and Menhas, Reference Dai and Menhas2020; Yu et al., Reference Yu, Wang, He, Kang, Miao, Wu, Yang, Zhu, Liang and Li2023). The efficiency of MCH services in rural areas is also much lower than in urban areas (Rizqi and Kurniawan, Reference Rizqi and Kurniawan2023). The Chinese government has implemented a series of measures aimed at enhancing the quality of MCH services for vulnerable populations in underdeveloped rural areas (Zhao et al., Reference Zhao, Diao, You, Wu, Yang and Liu2019). Including “Reduce Maternal Mortality Ratio and Eliminate Neonatal Tetanus,” (or Jiang Xiao Project), “The New Rural Cooperative Medical System,” “Subsidize Hospital Childbirths for Rural Women,” and so on. These programs and policies have directly and indirectly improved MCH in the central and western rural areas (Zhang et al., Reference Zhang, Chen, Lu, Hao, Zhang, Sun, Li and Chang2015). However, MCH outcomes in rural minority areas in western China remain poor (Gao et al., Reference Gao, Zhou, Singh, Powell-Jackson, Nash, Yang, Guo, Fang, Alvarez and Liu2017).
Guangxi (20 54 ′-26 20′ N, 104 26 ′-112 04′ E), located in western China, is one of the five major ethnic minority autonomous regions in China. According to the seventh census of China, the population of Guangxi is 50.1268 million, making it the most populous among ethnic minority province in the country, which is often considered an economically underdeveloped region based on China’s provincial Gross Domestic Product (GDP) per capita ranking (Xu et al., Reference Xu, Xue and Wu2022). Research indicates that in 2017, the quality evaluation of MCH in Guangxi ranked 19th among 31 provinces for maternal health care, while children’s health care was ranked 31st (Li et al., Reference Li, Wu, Feng and Xu2023). The Data Envelopment Analysis (DEA) efficiency evaluation value for MCH institutions at the municipal level in Guangxi was 0.953 (Yao et al., Reference Yao, Luo, Deng, Wei, He, Lei and Wei2023), significantly lower than that of developed regions such as Shanghai at 1.000 (Cai et al., Reference Cai, Zhang and Wang2022), Beijing at 0.981 (Zhao and Zheng, Reference Zhao and Zheng2023) and Guangdong at 1.031 (Li et al., Reference Li, Miao, Li, Liang, Wu and Zhao2019). This shows that there is considerable room for improvement in MCH for ethnic minorities in Guangxi. As the gatekeepers of MCH in rural areas, township health centers are the primary choice for mothers and children seeking health care services. In 2021, China began implementing the “Comprehensive Three-Child Policy” (Zhu et al., Reference Zhu, Zhang, Shen, Ye, Zhan, Cai, Huang, Wang and Chen2024). With the increased number of elderly “second-child” pregnant women, the number of high-risk and premature babies has also increased, placing immense pressure on MCH services in township health centers (Tian et al., Reference Tian, Ma, Du, Jin, Zhang, Xiao and Tang2023; Zhang et al., Reference Zhang, Huang, Lei, Liu and Liu2024). Therefore, the study of MCH in township health centers plays a key role in improving MCH care in rural areas.
However, previous studies have primarily focused on MCH at the provincial and municipal levels (Huang et al., Reference Huang, Luo, Wang, Cao, Wang, Bi, Huang and Yi2020; Cai et al., Reference Cai, Zhang and Wang2022) and at the district and county levels (Zhang et al., Reference Zhang, Shi, Li and Bian2020), using cross-sectional research designs. In terms of efficiency measurement, the main methods employed include the weighted RSR method and fuzzy comprehensive evaluation (Zhao et al., Reference Zhao, Chen, Li, Li and Li2021), as well as Banker–Charnes–Cooper (BCC)-DEA and Charnes–Cooper–Rhodes (CCR)-DEA efficiency evaluation models (Değirmenci, Reference Değirmenci2021; Ekinci, Reference Ekinci2024), and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method (Selamzade and Ersoy, Reference Selamzade and Ersoy2023). Additionally, there has been limited observation of changes in MCH over time (Trakakis et al., Reference Trakakis, Nektarios, Tziaferi and Prezerakos2021). Furthermore, research on MCH in township health centers in impoverished ethnic minority areas is very limited (Yan et al., Reference Yan, Tadadej, Chamroonsawasdi, Chansatitporn and Sung2020). Therefore, it is essential to conduct a comprehensive assessment of MCH in rural township health centers across different regions of China over a period of time.
Based on the above reasons, this study collected 6 years of longitudinal data, and used entropy weight coefficient method, BCC-DEA Model and Malmquist index Model to comprehensively evaluate and analyze the current situation, trend and efficiency of MCH in township health centers in minority areas. Our research aims to achieve the following objectives: First, to evaluate the current status and trends of MCH services. Second, to assess the efficiency of MCH services in township health centers using the BCC-DEA model, and to conduct a dynamic analysis of total factor productivity based on the Malmquist index model. Third, to provide targeted recommendations for improving the quality of MCH services in rural township health centers based on the research findings. The research findings offer valuable perspectives for policymakers and healthcare stakeholders in Guangxi, China, underscoring disparities in efficiency levels and illuminating potential avenues for enhancement. Through remedying these inefficiencies, policymakers can augment the efficiency of MCH in township health centers, ultimately improving the health of mothers and children in ethnic minority rural areas.
Materials and methods
Data sources and indicators selection
A multi-stage stratified random sampling method was employed to select a total of 49 township health centers for on-site investigation, including Lingui District, Lingchuan County, Yangshuo County, Gongcheng Yao Autonomous County, Longsheng Various Nationalities Autonomous County, and Yanshan District. Primary data sources for this study included the MCH Annual Reports and the “Gui Fu Er” information system from each township health center. Based on the “China Women and Children Development Plan 2011-2020” and in conjunction with a review of relevant literature (Duff et al., Reference Duff, Liu, Saavedra, Batycki, Morancy, Stocking, Gostin, Galea, Bertozzi and Zuniga2021; Esquivel et al., Reference Esquivel, Álvarez, Izquierdo, Martínez and Tamayo2014), considering the importance and availability of indicators, relevant indicators that could reflect the capacity of MCH services were identified for final use.
In the assessment of the current situation, we reference previous study results (Qin and Zhu, Reference Qin and Zhu2023; Wu et al., Reference Wu, Liu and Wang2023; Zhao et al., Reference Zhao, Chen, Li, Li and Li2021). The indicators included rates of early pregnancy health check-ups, prenatal examinations, postnatal visits, systematic management of pregnant women, high-risk maternal management, hospitalization for childbirth, maternal mortality, moderate to severe anemia among pregnant women, prenatal screening, exclusive breastfeeding for infants aged 0-6 months, neonatal mortality, mortality of children under 5, newborn visits, systematic management of children under 3, systematic management of children under 7, premarital medical check-ups, birth defects, and low birth weight (corresponding to X1-X18). In terms of input and output, based on previous research (Cheng et al., Reference Cheng, Jia, Zhang, Zhu and Wang2024; Huang et al., Reference Huang, Luo, Wang, Cao, Wang, Bi, Huang and Yi2020; Li et al., Reference Li, Miao, Li, Liang, Wu and Zhao2019), the number of full-time MCH care workers and equipment worth over 10,000 yuan are used as indicators for human resources and capital. The rates of systematic management for pregnant women and for children under 3 years old represent the outcomes of MCH services.
Data analysis
Entropy weight coefficient method
The entropy weight coefficient method is an objective weighting approach that determines the weight of the total score based on the data variance of individual unit scores, effectively avoiding subjective errors from evaluators. In this study, an improvement was made to the entropy weight coefficient method, incorporating time indicators for comparison between different years, based on relevant literature (Kohl et al., Reference Kohl, Schoenfelder, Fügener and Brunner2019; Zou et al., Reference Zou, Guo and Wu2023). The entropy weight method calculation includes the following six steps, as detailed in the Appendix.
BCC-DEA model
This study adopts the DEA method to measure MCH efficiency. DEA is an analytical method proposed by scholars such as Charnes and Cooper (1978) for assessing the relative efficiency among Decision Making Units with multiple inputs and outputs (Ratner et al., Reference Ratner, Shaposhnikov and Lychev2023). Since the input-output indicators of relevant healthcare resources exhibit variable returns to scale, the study uses the input-oriented BCC-DEA model, and the specific calculation formulas are provided in the Appendix.
Malmquist index model
Traditional DEA model primarily facilitate static assessments of resource allocation efficiency using cross-sectional data (Nazarko, Reference Nazarko2024). However, this study spans a longer time frame, necessitating consideration of temporal variations. The Malmquist index provides a means to evaluate changes in efficiency over a specified period, reflecting the performance of decision-making units (Andrejić et al., Reference Andrejić, Kilibarda and Pajić2021). Consequently, we implemented the Malmquist Productivity Index method to analyze panel data and illustrate dynamic shifts in efficiency. The MPI includes input-oriented efficiency change (EFFCH) and technical change (TECHCH). Efficiency change can also be divided into scale efficiency change (SECH) and pure efficiency change (PECH). The MPI, also known as Total Factor Productivity Changes (TFPCH), is derived from the distance function and can be represented by the following mathematical equations (The specific calculation formulas are provided in the Appendix).
Results
Maternal and child health service implementation
Table 1 shows that from 2016 to 2021, MCH indicators, such as the high-risk maternal management rate, hospitalization rate for childbirth, prenatal screening rate, and premarital check-up rate, exhibited an upward trend. Meanwhile, the neonatal mortality rate, under-five mortality rate, and the systematic management rate for children under 7 years showed a downward trend. The breastfeeding rate for infants aged 0–6 months demonstrated a fluctuating downward trend, while the incidence rate of birth defects and the incidence rate of low birth weight displayed an upward trend.
Maternal and child health services, 2016–2021
Note: Maternal Health Care: # represents positive indicators, * represents negative indicators; Child Health Care (under 7 years old): + represents positive indicators, - represents negative indicators.
Maternal health entropy weight coefficient score
Figure 1 shows the maternal health care scores of township hospitals across various regions from 2016 to 2021. Overall, the scores exhibited a dynamic upward trend, although differences among regions were observed. Specifically, in 2017, the scores reached their lowest point across all regions, with Gongcheng Yao Autonomous County scoring the lowest, below 0.4, indicating a gap in maternal care compared to other regions. In 2020, the maternal care scores peaked, with Lingui District scoring the highest, above 0.7, reflecting the best performance in maternal health care. In comparison, Longsheng Autonomous County and Gongcheng Yao Autonomous County scored lower than other regions, suggesting room for improvement in maternal care in these areas.
Changes in maternal health scores by region from 2016 to 2021.
Health entropy weight coefficient scores for children under 7 years of age
Figure 2 illustrates the health care scores for children under seven years old in township hospitals across different regions from 2016 to 2021. Overall, the scores showed a dynamic downward trend, with significant differences among regions. Specifically, in 2020, the scores for children’s health care were at their lowest, with Lingchuan County experiencing a sharp decline compared to 2019. This highlights the urgent need to address children’s health care in that area. However, by 2021, the scores in Lingui District and Gongcheng Yao Autonomous County had improved to some extent, while other regions continued to experience a decline, with Lingchuan County scoring the lowest.
Changes in health scores for children under 7 years of age by region from 2016 to 2021.
Efficiency of maternal and child health services from 2016 to 2021
Basic information of indicators
As shown in Table 2, the average number of full-time MCH care workers increased from 20.50 in 2016 to 21.33 in 2021, representing a 4.05% growth. The average number of equipment valued at over 10,000 yuan rose from 127.50 units in 2016 to 286.67 units in 2021, a growth of 124.84%. This indicates that China has increased its investment in healthcare resources for primary healthcare institutions. The rate of systematic management of pregnant women showed a fluctuating downward trend, while the systematic management rate for children under three years old exhibited a dynamic upward trend. Overall, the increase in input and output has not been entirely balanced, indicating a need for further optimization and improvement. The large gap between the maximum and minimum values of indicators also reflects widening regional disparities.
Input-output data from 2016 to 2021
Static analysis of service efficiency from 2016 to 2021 based on the BCC-DEA model
As shown in Table 3, healthcare service outputs were effective under DEA in 2016 and 2018, while they were deemed ineffective in 2017 and from 2019 to 2021. All the ineffective years are due to scale efficiency being less than 1. Considering returns to scale, scale returns remained constant in 2016 and 2018. In 2017, scale returns exhibited an increasing trend, indicating that the rate of increase in healthcare resource input was lower than the rate of output growth. This suggests the potential for further scientifically and reasonably increasing the proportion of resource input and appropriately expanding scale to achieve greater output. Scale returns for 2019 to 2021 showed a decreasing trend, signifying that the growth rate of healthcare resource input exceeded the growth rate of output during these three years, indicating the necessity to appropriately reduce the scale of resource allocation.
Static results of service efficiency from 2016 to 2021
PTE: Pure technical efficiency, SE: Scale efficiency, OE: Overall efficiency.
Static analysis of service efficiency in various regions in 2021
As shown in Table 4, from the efficiency analysis of different regions, it is evident that only Yanshan District exhibited effective service output under DEA. Gongcheng County, Lingui District, Yangshuo County, and Longsheng County are ineffective due to scale efficiency being less than 1. In contrast, Lingchuan is ineffective because both pure technical efficiency and scale efficiency are less than 1. Considering returns to scale, Yanshan District maintained constant scale returns. Gongcheng County, Lingui District, Yangshuo County, and Longsheng County demonstrated decreasing scale returns, indicating the need to appropriately reduce the scale of resource allocation. Lingchuan County exhibited increasing scale returns, suggesting the potential to expand scale appropriately to achieve greater output.
Static results of service efficiency in various regions in 2021
PTE: Pure technical efficiency, SE: Scale efficiency, OE: Overall efficiency.
An overall analysis of malmquist index for maternal and child health services
As shown in Figure 3, from 2016 to 2021, total factor productivity exhibited a dynamic upward trend, indicating an overall improvement in resource allocation efficiency during this period. The average annual growth rate of total factor productivity from 2016 to 2021 was 7.3%. An analysis of the components of total factor productivity reveals that the efficiency values were greater than 1 from 2016 to 2018 and from 2020 to 2021, primarily due to an increase in the technological change index. In 2018-2019, the efficiency value was greater than 1 primarily due to an increase in the efficiency change. In 2019-2020, the efficiency value was greater than 1, mainly due to both the technological change index and the efficiency change index being greater than 1.
Changes in total factor productivity of maternal and child health services from 2016 to 2021.
Efficiency changes can be decomposed into SECH and PECH. As shown in Figure 4, from 2016 to 2021, the efficiency change index exhibited a fluctuating downward trend. However, the efficiency change index was greater than 1 in 2018-2019 and 2019-2020, indicating an increasing trend in efficiency change during this period. On the other hand, the efficiency change indices from 2016 to 2018 and from 2020 to 2021 were both less than 1, indicating a downward trend in efficiency changes during these periods. The decomposition of efficiency reveals a consistent trend between efficiency change and SECH, suggesting that the decline in pure technical efficiency from 2016 to 2021 has somewhat slowed down the rate of efficiency change. The main driver of efficiency change is attributed to the SECH during the period.
Changes in efficiency of maternal and child health care services from 2016 to 2021.
Malmquist index analysis of maternal and child health services in different regions
As shown in Figure 5, the total factor productivity index in all regions is greater than 1. This indicates that the allocation efficiency of MCH resources in each region is generally on an upward trend. An examination of the components of efficiency reveals that Gongcheng County, Lingui District, and Yangshuo County had both efficiency changes and technological changes greater than 1, suggesting that the increase in efficiency in these areas resulted from the combination of both factors. The efficiency change in Yanshan District and Longsheng County is greater than 1, indicating that the increase in the total factor productivity index is primarily due to improvements in the efficiency change index. Lingchuan County had a technological change index greater than 1, indicating that the increase in efficiency was primarily due to advancements in the technological change index.
Changes in total factor productivity of maternal and child health services by regions from 2016 to 2021.
As shown in Figure 6, an analysis of the efficiency changes from 2016 to 2021 reveals that the efficiency change indices for Gongcheng County, Yanshan District, and Longsheng County are all greater than 1, indicating an overall upward trend in efficiency changes. Lingchuan County had an efficiency change index less than 1, indicating an overall downward trend in efficiency changes. An examination of the components of efficiency reveals that Gongcheng County and Yangshuo County had both pure technical efficiency and scale efficiency greater than 1, indicating that the improvement in efficiency changes was driven by simultaneous increases in both pure technical efficiency and scale efficiency. Lingchuan County had a pure technical efficiency change equal to 1 and a SECH index less than 1, suggesting that its decrease in efficiency was primarily due to a decline in scale efficiency.
Changes in efficiency of maternal and child health care services by regions from 2016 to 2021.
Discussion
In 2009, according to the policies of the new healthcare reform, which aimed to “ensure that urban and rural residents have equal access to the most basic and effective public health services and to narrow the urban-rural gap,” China increased its investment in rural health care projects. Local governments strengthened the capacity building of rural health care institutions through urban-rural coordination (Wang et al., Reference Wang, Liao and Feng2020; Zhou et al., Reference Zhou, Han, Zhang, Fan, Liu, Liu, Fan and Qu2024). This study shows that MCH services in township health centers have overall improved. However, the incidence of birth defects and low birth weight has been on the rise. These findings are consistent with studies conducted in Heilongjiang Province, China (Tian et al., Reference Tian, Wu, Li, Zhu, Liu, Zhang and Liu2022), and Ethiopia, Africa (Kassaw et al., Reference Kassaw, Abebe, Kassie, Abate and Masresha2021). It is reported that approximately five million infants with birth defects are born globally each year, with over 85% of these cases occurring in developing countries (Zhong and Jiang, Reference Zhong and Jiang2019), and nearly 15% of babies worldwide are low birth weight and more than half of them are born in Asia (Kouser et al., Reference Kouser, Bala, Sahni and Akhtar2020). Environmental and biological factors, along with social progress leading to increased detection rates of congenital abnormalities, continuous improvements in prenatal examinations and screening methods, enhanced diagnostic levels and quality, improved tertiary surveillance networks, and an increase in hospital deliveries (Findley et al., Reference Findley, Parchem, Ramdaney and Morton2023; Fomda et al., Reference Fomda, Velayudhan, Siromany, Bashir, Nazir, Ali, Katoch, Karoung, Gunjiyal and Wani2023; Huang et al., Reference Huang, Chen, Zeng, Lin, Herbert, Cottrell, Liu, Ash and Wang2021) all contribute to the causes of this rise. Additionally, recent studies have suggested a potential association between the occurrence of low birth weight and maternal exposure to air pollution during the preconception (Bekkar et al., Reference Bekkar, Pacheco, Basu and DeNicola2020) and early pregnancy periods, as well as the supply of nutrients during pregnancy (Harper et al., Reference Harper, Rothberg, Chirwa, Sambu and Mall2023). Therefore, through the implementation of a special management model for high-risk infants and young children, a planned management approach is adopted to identify and monitor high-risk cases, ensuring early detection, evaluation, diagnosis, and intervention for developmental abnormalities and deformities.
According to the entropy weight coefficient method, we found that there are still disparities in MCH services across different regions. Longsheng County and Gongcheng County perform poorly in MCH services, showing significant room for improvement compared to other areas. First, Longsheng County and Gongcheng County are ethnic minority areas that face a scarcity of MCH service resources. For example, Lingui District, which performs well in MCH services, ranks second in economic level within the city. It is a directly governed district with relatively abundant MCH resources, and it is close to several municipal public hospitals and MCH centers. Other studies have found that regions with strong economic development and abundant healthcare resources often exhibit better MCH services (Paul et al., Reference Paul, Paul, Gupta and James2022; Rios Quituizaca et al., Reference Rios Quituizaca, Gatica-Domínguez, Nambiar, Ferreira Santos, Brück, Vidaletti Ruas and Barros2021). Therefore, the differences in the capacity of MCH services among regions may be attributed to a combination of varying economic levels (Janevic et al., Reference Janevic, Zeitlin, Egorova, Hebert, Balbierz and Howell2020)and the strength of MCH capabilities in each area (Ramadan et al., Reference Ramadan, Gutierrez, Feil, Bolongaita, Bernal and Uribe2023). Second, township health centers in minority counties are often located in remote mountainous areas with inconvenient transportation, which may lead to difficulty in covering all populations in need (Kaiser and Barstow, Reference Kaiser and Barstow2022). Third, there is a loss of talent in MCH. Numerous studies have demonstrated that satisfaction with compensation is a significant factor that affects the professional loyalty of primary healthcare workers (Soesanto et al., Reference Soesanto, Yanto, Irani, Pranata, Rejeki and Sasmito2022; Zhang et al., Reference Zhang, Shi, Li and Bian2020). Compared to clinical doctors, MCH physicians earn significantly lower salaries (Brekke et al., Reference Brekke, Holmås, Monstad and Straume2020), leading them to pursue higher wages in clinical departments, which results in a decline in the quality of MCH services. Therefore, it is necessary to improve transportation and raise the salaries of MCH workers so that they can better serve women and children in rural areas.
In the efficiency analysis of MCH services, we found that all years, except for 2016 and 2018, were in a DEA inefficient state. Specifically, over 60% of the areas demonstrated low efficiency in MCH, with a decline in scale efficiency, indicating that these inefficient regions are using more healthcare resources than currently needed for health services. According to the input slack data, almost all inefficient regions had input redundancies, particularly in the number of medical equipment (≥CNY 10,000) (Chinese Yuan). Several factors could explain this outcome. One possibility is that the demand for MCH services was overestimated, resulting in redundant MCH resources in these inefficient regions. While this chance may be minimal, it should not be overlooked in decision-making. From the 1970s to the early 2000s, China’s health institutions gained considerable autonomy with limited government subsidies during the market-oriented reform period (Huang et al., Reference Huang, Luo, Wang, Cao, Wang, Bi, Huang and Yi2020). Influenced by market competition, these institutions significantly increased investments such as facility expansion, bed installation, and medical equipment purchases (Jiang and Pan, Reference Jiang and Pan2020). A previous survey indicated that less than 50% of large-scale medical devices were in use in China (Huang et al., Reference Huang, Luo, Wang, Cao, Wang, Bi, Huang and Yi2020). This underutilization of invested resources can lead to significant waste, ultimately reducing overall MCH efficiency. This indicates that health administrators should be aware of potential resource redundancy when formulating regional health plans.
Another possibility is that many families prefer municipal MCH hospitals to local primary health service centers in search of higher-quality MCH services, as municipal tertiary hospitals typically have more skilled healthcare personnel and better technology (Zhou et al., Reference Zhou, Han, Zhang, Fan, Liu, Liu, Fan and Qu2024). This phenomenon suggests a certain level of inequality in the allocation of MCH resources and healthcare utilization, with urban areas having more experienced specialists and advanced equipment than rural areas. Sharing training programs at municipal MCH hospitals could be a potential way to improve the quality of healthcare services in underdeveloped rural health centers (Helmyati et al., Reference Helmyati, Dipo, Adiwibowo, Wigati, Safika, Hariawan, Destiwi, Prajanta, Penggalih and Sudargo2022). It may be useful to improve the quality of medical care in economically disadvantaged rural areas in a sustainable manner that grassroots health workers return to work after a period of receiving the technical training from municipal tertiary hospitals. For regions with inefficiencies, transferring their excessive input resources to more available health services for mothers and children is another reasonable method to improve the efficiency of MCH resources. The disparities in MCH performance across various regions may contribute to the observation that, at equivalent input levels, the volume of services generated by MCH in less efficient areas is noticeably lower than that in their more efficient counterparts. In this regard, attention should be directed toward enhancing the capacity for improved service delivery. The inefficient regions are primarily tasked with improving staff performance and the quality of their management practices in order to effectively utilize inputs and provide available health care services for the mothers and children in economically disadvantaged rural areas (Alotaibe et al., Reference Alotaibe, Alharthi, Al Rogi, Al Awadh, Alqarni, Alshammeri, Almoteri, Alzhrani, Al Rizq and Almoder2022).
A study and analysis of panel data from different regions between 2016 and 2021 using the Malmquist index revealed varying degrees of improvement in TFPCH. Overall, there was a dynamic growth trend in TFPCH changes, which is consistent with the findings of Junxu Zhou’s analyzing the efficiency of Chinese primary healthcare institution (Zhou et al., Reference Zhou, Peng, Chang, Liu, Gao, Zhao, Li, Feng and Qin2023). Zhang’s analysis based on the Malmquist index showed an average annual growth rate of 2.5% in the total factor productivity of county-level MCH institutions in China from 2010 to 2014 (zhang et al., Reference Zhang, Sun, Li, Zhu and Ren2017). These studies all support the findings of the present research. The reasons for this trend could be attributed to the impact of policies such as the “New Medical Reform,” the “Outline for the Development of Women and Children,” and the “Decision on Optimizing Birth Policies to Promote Long-Term Balanced Population Development,” which have led to an overall increase in healthcare resource inputs. The TFPCH can be broken down into the TECHCH and the EFFCH, with the EFFCH further divided into PECH and SECH (Zhang et al., Reference Zhang, Chen, Wang, Wang and Zhang2023). However, our study revealed that TFPCH growth varied across regions. This also validates the evaluation of MCH by entropy weight coefficient to a certain extent. The growth in TFPCH is primarily driven by an increase in the EFFCH, while the rise in EFFCH is mainly due to the improvement in PECH. Guangxi has set up a four-level MCH service system covering provinces, cities, counties, and townships, and implemented a training program for MCH personnel, which included training targeted at admitting medical students to work in rural areas after graduation and strengthening the training of registered nurses (Zhou et al., Reference Zhou, Han, Zhang, Fan, Liu, Liu, Fan and Qu2024). In contrast, changes in SECH may be the primary factor contributing to the decline in EFFCH, as seen in Longsheng County. Therefore, to improve the efficiency of MCH and support the sustainable growth of MCH services, Guangxi needs to further improve its internal structure and management levels to enhance scale efficiency. It is also essential to strengthen the service capabilities of primary health institutions (Ahmed et al., Reference Ahmed, Hasan, MacLennan, Dorin, Ahmed, Hasan, Hasan, Islam and Khan2019), promote collaboration among medical facilities at various levels (Lal et al., Reference Lal, Erondu, Heymann, Gitahi and Yates2021).
There are two limitations in this study. Firstly, Due to the ongoing COVID-19 pandemic during the 2022 survey, there may have been complexities and challenges in grassroots MCH care services. The data related to grassroots MCH services in this survey were based on the Outline for the Development of Women and Children, the Service Evaluation Guide for Township Health Centers, and Basic Public Health Services. Discrepancies may exist compared to the actual work statistics and assessment indicator systems. Secondly, due to limitations in time and resources, the survey’s comprehensiveness might be limited. Future research endeavors aim to broaden the scope of data collection, conduct further investigations, refine data structures, elaborate on indicators, and further strengthen empirical research on grassroots medical and health systems.
Conclusion
The new healthcare reform policies have catalyzed the development of primary healthcare institutions in our country, offering decision-making support for their rational and effective progress. This is particularly evident in the advancement of township health centers. The findings of this study hold significant implications for the further enhancement of primary MCH in rural China, as well as for the strategic planning of healthcare systems, especially in the less developed regions. This study conducted an empirical investigation into the current status, efficiency, and trends of MCH services in township hospitals in rural ethnic minority areas. The results indicate that, first, there has been an overall improvement in the level of MCH services. Furthermore, there has been a modest overall increase in the use of MCH, despite concerns about the increasing inaccessibility of MCH services in rural China. However, attention should be paid to the rising trend in the incidence of birth defects and low birth weight. Second, there are regional disparities in efficiency evaluation, which are related to the capacity of MCH services and the economic levels of each area. Experiences from other countries support this view. Third, the improvement in productivity has relied solely on technological advancements rather than enhancements in internal management and institutional innovation. Although the input of various resources has certainly contributed to improvements in MCH, the enhancement of SE has not kept pace. In the future, internal management should be strengthened by setting service goals and updating management concepts. Existing resources should be fully and rationally utilized, and the relationship between scaling up and improving quality should be addressed through the flow of resources among institutions.
Data availability statement
The datasets generated and/or analyzed during the current study are not publicly available due protect confidentiality of data from primary healthcare institutions but are available from the corresponding author on reasonable request. The formulas for data calculations in the appendix were also obtained from the corresponding author upon reasonable request.
Author contributions
Zhuanzhi Tang drafted and revised the manuscript; Ranfeng Hang, Siyuan Wang, and Jianying Liu collected, analyzed, and interpreted the data; and Wuxiang Shi designed the study. All authors read and approved the manuscript.
Funding statement
This research is not funded by a specific grant or contract, but it received financial support from the “Bagui Scholars Program in Social Medicine and Health Management” in Guangxi, China.
Competing interests
The authors declare no conflict of interest.
Ethical standards
This study has been approved by the Ethics Committee of Guilin Medical College. All methods were carried out in accordance with relevant guidelines and regulations (Declaration of Helsinki). All participants signed informed consent.
Code availability
Not applicable
Appendix
Entropy weight coefficient method
Step 1: Construct the corresponding output data matrix X. This converts original data into relative values between 0 and 1, where values closer to 1 indicate better performance. Indicators are classified as positive (higher values mean better performance) or negative (higher values mean worse performance):
$x{'_{ij}} = \left\{ {{{{{{x_{ij}} - \min \left\{ {{x_{1j'}}} \right....,\left. {{x_{nj}}} \right\}} \over {\max \left\{ {{x_{1j'}}} \right....,\left. {{x_{nj}}} \right\} - \min \left\{ {{x_{1j'}}} \right....,\left. {{x_{nj}}} \right\}}}} \over {{{\max \left\{ {{x_{1j'}}} \right....,\left. {{x_{nj}}} \right\} - {x_{ij}}} \over {\max \left\{ {{x_{1j'}}} \right....,\left. {{x_{nj}}} \right\} - \min \left\{ {{x_{1j'}}} \right....,\left. {{x_{nj}}} \right\}}}}}} \right.$
Step 2: Calculate the weight P ij of the j-th indicator for the i-th evaluation criterion:
${p_{ij}} = {{x{'_{ij}}} \over {\sum _{i = 1x'ij}^n}},i = 1,2 \cdot \cdot \cdot ,\,\,n,j = 1,2, \cdot \cdot \cdot m$
Step 3: Calculate the entropy e ij of the j-th indicator:
${{\rm{e}}_{ij}} = - k\sum\limits_{i = 1}^n {{p_{ij}}} in({p_{ig}}),i = 1,2, \cdot \cdot \cdot n,j = 1,2, \cdot \cdot \cdot m$
Step 4: Calculate the redundancy degree d j of the entropy:
${{\rm{d}}_j} = 1 - {e_j},j = 1,2, \cdot \cdot \cdot m$
Step 5: Calculate the objective weight w j for each indicator:
${w_j} = {{{d_j}} \over {\sum _{j = 1}^m{d_j}}} \times 100\% ,j = 1,2, \cdot \cdot \cdot m$
Step 6: Calculate the comprehensive score s j for each evaluation unit:
${s_j} = \sum\limits_{i = 1}^m {{w_j}} x{'_{ij}},j = 1,2, \cdot \cdot \cdot m$
BCC-DEA Model
The study employs the input-oriented BCC-DEA model, which can be expressed as:
${\rm{(BCC}})\left\{ \matrix{ \min \theta \hfill \cr \sum\limits_{j = 1}^{\rm{n}} {\mathop \lambda \nolimits_j \mathop Y\nolimits_j + \mathop S\nolimits^ - = \mathop {\theta X}\nolimits_0 } \hfill \cr \sum\limits_{j = 1}^n {\mathop \lambda \nolimits_j \mathop Y\nolimits_j - \mathop S\nolimits^ - = \mathop Y\nolimits_0 } \hfill \cr \sum\limits_{j = 1}^n {\mathop \lambda \nolimits_j = 1,\mathop \lambda \nolimits_j \ge 0,j = 1,2,...,n} \hfill \cr \mathop S\nolimits^ + \ge 0,\mathop S\nolimits^ - \ge 0 \hfill \cr} \right.$
In this expression, θ is the efficiency of DMU j , λ j is the unit’s weight, X j and Y j are the input and output quantities of the j-th unit, n is the number of units, and S+ and S- are the input and output slack variables. The effectiveness of DMU j is determined by the model’s optimal solution values.
(1)If θ=1 and both S + and S - are 0, DMU j is DEA efficient; (2) If θ=1 but S + or S - is not 0, DMU j is weakly DEA efficient; (3) If θ<1 and S + or S - is not 0, DMU j is non-DEA efficient.
Malmquist index Model
The MPI, also known as Total Factor Productivity Changes (TFPCH), is derived from the distance function and can be represented by the following mathematical equations:
$MPI_I^{\rm{t}} = {{E_I^t({x^{t + 1}},{y^{t + 1)}}} \over {E_I^t({x^t},{y^t})}},MPI_I^{t + 1} = {{E_I^{t + 1}({x^{t + 1}},{y^{t + 1)}}} \over {E_I^{t + 1}({x^t},{y^t})}}$
To thoroughly understand the technical level during both periods, we took into account the geometric mean:
$MPI_I^{\rm{G}} = {(MPI_I^tMPI_I^{t + 1})^{{1 \over 2}}} = {[{{E_I^t({x^{t + 1}},{y^{t + 1)}}} \over {E_I^t({x^t},{y^t})}} \times {{E_I^{t + 1}({x^{t + 1}},{y^{t + 1)}}} \over {E_I^{t + 1}({x^t},{y^t})}}]^{{1 \over 2}}}$
The productivity function includes input-oriented efficiency change (EFFCH) and technical change (TECHCH). Efficiency change can also be divided into scale efficiency change (SECH) and pure efficiency change (PECH).
$\eqalign{ & MPI_I^{\rm{G}} = (EFFC{H_I}) \times (TECHCH_I^{\rm{G}}) = {{E_I^t({x^{t + 1}},{y^{t + 1)}}} \over {E_I^t({x^t},{y^t})}} \times {[{{E_I^t({x^{t + 1}},{y^{t + 1)}}} \over {E_I^t({x^t},{y^t})}} \times {{E_I^{t + 1}({x^{t + 1}},{y^{t + 1)}}} \over {E_I^{t + 1}({x^t},{y^t})}}]^{{1 \over 2}}} \cr & SECH = {[{{E_{vrs}^{t + 1}({x^{t + 1}},{y^{t + 1}})/E_{crs}^{t + 1}({x^{t + 1}},{y^{t + 1}})} \over {E_{vrs}^{t + 1}({x^t},{y^t})/E_{crs}^{t + 1}({x^t},{y^t})}} \times {{E_{vrs}^t({x^{t + 1}},{y^{t + 1}})/E_{crs}^t({x^{t + 1}},{y^{t + 1}})} \over {E_{vrs}^t({x^t},{y^t})/E_{crs}^t({x^t},{y^t})}}]^{{1 \over 2}}} \cr & PECH = {{E_{vrs}^{t + 1}({x^{t + 1}},{y^{t + 1}})} \over {E_{crs}^{t + 1}({x^t},{y^t})}} \cr} $
Open access
