The study was conducted in Kazakhstan. It is a landlocked middle-income country in the center of Eurasia, with a small amount of its area belonging to Europe and the majority of it being Asia. The population of Kazakhstan is estimated at 19,644,067 by the State Statistics Committee (August 1, 2022). The area of the territory is 2,724,902 km². By territory, it holds the ninth-place position in the globe. Kazakhstan borders China, Russia, Kyrgyzstan, Uzbekistan, and Turkmenistan (23). Administratively, the Republic of Kazakhstan is split into 14 regions and three significant republican cities (Astana, Almaty, and Shymkent).

The work of healthcare institutions and the entire population of Kazakhstan in the period from 2000 to 2020 was studied using data from the annual statistical report (24). These reports include information on population health and health care services at the national and regional levels. To ensure comparability of statistics, health data collection practices in Kazakhstan comply with international standards. Socioeconomic data of the population was obtained from the official statistical portals of the Republic of Kazakhstan.

In this study, we analyzed several key factors sourced from the annual report. These factors included healthcare service and socioeconomic predictors.

We conducted a comprehensive examination of the maternal mortality ratio in Kazakhstan across all 14 regions and 3 cities of republican significance spanning a 20-year timeframe. Our study investigated the relationships between MMR and various socioeconomic and healthcare service variables both before and during the COVID-19 pandemic, employing a both-direction Stepwise Linear Regression analysis. Stepwise Linear Regression was utilized to identify the optimal subset of explanatory variables for multiple regression models for the years 2019 and 2020. This method involves comparing improvements in the Akaike Information Criterion by systematically removing or adding candidate variables within the specified bounds of regressors. We identified a total of 78 variables based on a thorough review of literature and data availability pertaining to the determinants of MMR in Kazakhstan during the aforementioned years. These variables were categorized into two distinct groups: Socioeconomic and Healthcare Service factors, leading to the derivation of four multiple regression models:

The coefficients of statistically significant variables derived from these models are outlined in Tables 1, 2.

We also described the MMR trends in Kazakhstan regions from 2000 to 2020. For statistical analysis, we applied IBM SPSS Statistics version 26.

In 2020, the national MMR almost tripled compared to 2019 (13.7 in 2019 and 36.5 per 100,000 live births in 2020). Also, there was a dramatic increase of MMR in 2020 in all regions of Kazakhstan. Thus, in the North Kazakhstan and Pavlodar regions, where no cases of maternal mortality were registered in 2019, during the pandemic there was a sharp rise in this indicator to 76.7 and 25.7 per 100,000 live births, respectively. In terms of MMR, the Karaganda region led in 2019 with an indicator of 39.5 per 100,000 live births, while the Kostanay region led in 2020 with an indicator of 94.4 per 100,000 live births (Figure 1).

MMR trends in various parts of Kazakhstan indicate a consistent fall beginning in 2000 and continuing until 2010–2015, then increasing by 2020. Maternal mortality per 100,000 live births has dramatically increased after 2010–2015 in the Aktobe region (from 10 in 2015 to 70.3 in 2020), in the Almaty region (from 5.1 in 2013 to 30 0.4 in 2020), West Kazakhstan region (from 7.9 per in 2013 to 21.9 per in 2020), Karaganda region (from 4.1 in 2013 to 39.5 in 2020), Kostanay region (from 15.3 in 2013 to 94.4 in 2020), Kyzylorda region (from 10 in 2014 to 56 in 2020), North Kazakhstan region (from 12.1 in 2011 to 76.7 in 2020) (Figure 2).

The dependent variable of the linear regression was MMR, with a mean of 12.91 and a standard deviation of 10.88, the MMR values ranged from 0 to 39.50 for 2019 (Table 3).

The descriptive statistics for variables of linear regression showed, that 88.20% of pregnant women received antenatal care, with a standard deviation of 2.47%. The volume of healthcare services provided ranged from 40.21 to 235.69 billion tenge and showed a mean of 83.23 with a standard deviation of 57.19 billion tenge. Abortions per 100 live births, including stillbirths, ranged from 6.70 to 50.10, with a mean rate of 22.64. The average availability of primary healthcare units indicated 115 units with a standard deviation of 76.57, suggesting differing levels of healthcare accessibility across regions. The density of nursing and midwifery personnel per 10,000 population averaged 53.15.

According to socioeconomic indicators, the mean divorce ratio was 3.26 thousand with a deviation of 0.71. The average gross enrollment rate in secondary education was 99.9%. The indicator “violence against women and girls” averaged 5.66, with a standard deviation of 5.68.

Table 1 illustrates the factors exerting a statistically significant influence on the MMR in Kazakhstan during the pre-COVID period. Specifically, healthcare service indicators such as the antenatal care coverage (β = −2.51, p < 0.01) and the availability of nursing and midwifery personnel per 10,000 population (β = −0.89, p = 0.001) demonstrate associations with a reduction in the MMR.

An analysis of socioeconomic factors as predictors of MMR reveals notable associations in Kazakhstan for the year 2019. Specifically, divorce ratio (β = 14.96, p < 0.05) is linked with a significant rise in MMR, while a higher percentage of gross enrollment in secondary education correlates with a reduction in MMR. Additionally, it is noteworthy that instances of domestic violence against women (β = 1.53, p = 0.01) contribute to an elevated MMR.

The output of linear regression indicated as MMR during pandemic time, and the average MMR was approximately 41.97 deaths per 100,000 live births, with a standard deviation of 23.27 (Table 4). On average, about 85.1% of pregnant women receive antenatal care, including early registration, up to 12 weeks, with a standard deviation of 4.11%. Moreover, this indicator ranged from 76.40 to 91.80%. The average number of primary health care units available in the population was approximately 186.41 units, with a wide standard deviation of 90.41. Almost all population within the age group of 11–17 years are enrolled in secondary education. The number of reported cases of domestic violence against women varied from 1.14 to 13.02, also the mean was approximately 3.72, with standard deviation of 2.92.

A comprehensive regression analysis focusing on healthcare indicators pertaining to MMR unveils two statistically significant factors linked with a reduction in MMR: the number of primary health care units (β = −0.21, p < 0.01) and the antenatal care coverage, including early register of pregnancy up to 12 weeks (β = −9.46, p < 0.01). Furthermore, it is noteworthy that the coefficient for the predictor “Antenatal care coverage, including early register of pregnancy up to 12 weeks, %” experienced a notable shift during the pandemic period, dropping to −9.46 from −2.51 observed in the pre-pandemic period.

Similar to the findings in 2019, predictors of MMR in 2020 include the percentage of gross enrollment in secondary education (11–17 years) (β = −2.39, p < 0.05) and number of reported cases of domestic violence against women. Also, in 2020, the surge in domestic violence cases is significantly correlated with a notable rise in the average MMR in Kazakhstan (β = 6.04, p < 0.05).

The MMR and socioeconomic, medical, and morbidity indicator factors were shown to be significantly correlated in our study. Our analysis revealed several key determinants influencing MMR: socioeconomic factors (total divorce ratio, enrollment in secondary education, number of reported cases of domestic violence against women), healthcare related factors (volume of healthcare services provided, abortion rates, amount of nursing and midwifery personnel, number of primary health care units, antenatal care coverage). These findings align with the literature describing the individual-level factors (age, parity), household characteristics (location, access to clean water), community elements (socioeconomic status), and systemic aspects (healthcare facilities) as prominent influences on maternal mortality (2, 3, 38).

Since 2000, Kazakhstan’s MMR was the lowest among Central Asian nations, and it had been declining for the previous 10 years. This achievement can be attributed to the realization of the Government policy of the Republic of Kazakhstan, formulated through the consistent government programs: “Salamatty Kazakhstan” (2011–2015), “Densaulyk” (2016–2019), The Concept of healthcare development of the Republic of Kazakhstan until 2026 (39). Considering the challenges the nation’s healthcare system faces, the primary strategies have been established in a number of areas, with one of the top goals being the improvement of women’s and children’s health.

According to the Multiple Indicator Cluster Survey in the Republic of Kazakhstan, 2010–2011 99.2% of pregnant women received antenatal care at least once throughout their pregnancy, indicating a very high coverage rate for this type of care (40). In the research conducted by Dauletyarova et al. (41) 90.0% of the women expressed satisfaction with the antenatal care they received. Our research findings identified that 88.2% of pregnant women were covered with antenatal care services in 2019, whereas in 2020 this indicator dropped to 85.17%. Regression analysis revealed that the coverage of antenatal care had a higher impact on the decline in MMR in 2020 compared to 2019 (Tables 1,2).

Due to difficulties in obtaining healthcare, variations in pandemic containment strategies, and a high frequency of COVID-19 risk factors there may be an increased risk of maternal deaths as a result of COVID-19 in low- and middle-income countries affected by the pandemic (7, 42). El-Shal et al. (43) suggests that health emergencies, such as the COVID-19 pandemic, can instantly increase maternal, under-five, and neonatal mortality by 0.3, 0.3, and 0.2% and, after 1 year by 35, 80, and 26%, respectively (44). Our findings show that while Kazakhstan experienced a general decline in MMR from 2000 through around 2015, there was an alarming increase by 2020. In contrast to 2019, the MMR rapidly climbed in 2020, rising by 2.6 times to 36.5 per 100,000 live births (45). A nearly threefold rise in MMR nationally from 2019 to 2020 suggests significant systemic stressors at play. The growth of excess mortality is being observed around the world, and the pandemic contributed to it. The study from Brazil reported that, even after accounting for the anticipated excess mortality from COVID-19, the excess maternal mortality in Brazil in 2020 was 1.40 (95% CI 1.35–1.46) (46). The pandemic also resulted in a 33.3% relative rise in maternal fatalities in the US (47). Furthermore, in the USA the number of maternal mortality cases grew with maternal age (43). A study from Kazakhstan reported that higher age associated with greater mortality risk in this period, and from 2019 to 2021 in Kazakhstan, there was a rise in mortality rates among women aged 35–39 from 0.48 to 4.37 per 100,000 individuals (48, 49). Overall, even before the pandemic, obesity and high school non-completion were identified as important contributors to maternal mortality, while access to healthcare services and resources, family factors, such as revenue and education, and ecological determinants as poverty, gender inequality, human development also played a role (3, 6, 50).

Our linear regression analysis showed that factors such as antenatal care coverage, percentage of population enrolled in secondary education, number of nursing and midwifery personnel, were statistically significantly associated with a decreased level of MMR. These findings contribute to the existing body of literature by providing specific insights into the Kazakhstani context. Thus, increasing coverage of midwife-delivered interventions potentially significantly influence the reduction of stillbirths, neonatal fatalities, and maternal mortality. 41% of maternal deaths, 39% of neonatal deaths, and 26% of stillbirths might be avoided with a significant increase in coverage. Universal coverage has the potential to make the largest impact in low- to medium-HDI countries, potentially saving up to 4.3 million lives yearly by 2035 (51). The decrease in the availability of competent healthcare professionals during childbirth due to reductions in government health spending has been found to significantly contribute to an increase in maternal mortality rates (52). A well-equipped, empowered, and effectively deployed healthcare workforce is critical for the attainment of health-related SDGs. Most high-income countries have an adequate number of health professionals to provide necessary treatment, and they generally have strong educational and regulatory systems in place. Similar trends can also be observed in upper-middle-income countries, with their policies being supportive of the healthcare sector (53–55). The Universal Health Coverage (UHC) Index for Kazakhstan has seen a steady increase from 2000 to 2021, which can be considered as a positive indicator in the context of maternal mortality. Starting at a score of 56 in 2000, the index rose gradually over two decades to reach and maintain a score of 80 by 2019 through 2021. Improvements in the UHC suggest that more people in Kazakhstan are accessing essential health services, which includes maternal and child healthcare (56). Moreover, since universal health coverage is aligned with SDG 3.8.1, which specifically aims to achieve equity in healthcare services including reproductive, maternal, newborn, and child health, this upward trajectory correlates with efforts directed toward lowering the MMR (57). Since the mid-2000s, Kazakhstan has established youth health centers that play a significant role in enhancing maternal health among adolescents. These centers provide young women with education on prenatal care, which is essential for reducing pregnancy-related complications and maternal mortality. These centers strive to prevent unsafe abortions by guiding adolescents toward effective family planning methods and contraceptive use, thereby reducing unwanted pregnancies and associated health risks (58). In the context of maternal health and mortality, Kazakhstan’s fall to 72nd place in gender equality list in 2019 highlights significant disparities affecting women’s health and survival. The country’s decline from 32nd since 2006 signals pressing issues like wage gaps, few women leaders, and unequal domestic workloads. Pregnant women and mothers face severe discrimination that can hinder job access and growth, leading to financial issues which negatively impact maternal healthcare availability and increase mortality risks (59). Central Asian Sample Survey on the prevalence of violence against women indicated a significant incidence of domestic abuse in Kazakhstan. Of the ever-partnered women aged 18–75 surveyed, 17% reported encountering physical or sexual violence from a partner, while 21% faced psychological abuse (60). Domestic violence had a larger influence on the rise in MMR in 2020 than it did in 2019, according to our data, even though the number of instances fell from 5.66 thousand in 2019 to 3.718 thousand in 2020. Recently, the Kazakhstan government has enacted a New Law on the Rights of Women and Safety of Children. This law aims to challenge male chauvinism embedded within society and legal frameworks by advocating for legislation that historically favors men’s interests at the expense of recognizing and protecting women’s vulnerabilities (61).

In low and low-middle-income countries, maternal education was revealed to be one of the most important individual factors of maternal mortality, along with the location of delivery, delays in seeking health care, prenatal care, and competent birth attendance (62). According to a 2014 OECD study, 28% of women pursued higher education compared to 23% of men (20). Furthermore, both women and men have very high literacy rates (99.9%), and the labor force participation rate in Kazakhstan stands at 63.2% for women and 74.5% for men (21).

While government revenue is crucial, the quality of governance plays an even more key role in achieving positive results for mothers and children. This impact is more pronounced in low-income countries. As government revenue reaches a level of around 5,000 USD per capita, the impact of governance on outcomes becomes less critical. The incidence and maternal mortality rates for COVID-19 were higher in the localities with less access to healthcare and greater socioeconomic disparities (63). In Kazakhstan, the correlation between high out-of-pocket healthcare spending and low government health investment is evident. With public healthcare expenditure at merely 2% of the GDP, it falls significantly short of the average 6.5% reported by countries in the OECD (64). Kazakhstan’s recent implementation of a mandatory health insurance system, where the Social Health Insurance Fund became the purchaser of publicly funded health services in 2020, represents a significant shift in healthcare financing. These reforms aim to enhance the accessibility, equity, and efficiency of health services. The introduction of mandatory health insurance complements the state-guaranteed basic package, diversifying funding sources for healthcare. Despite these efforts, public spending on health decreased from approximately 75% in 2009 to around 60% in 2019. Meanwhile, out-of-pocket payments rose to constitute 33.9% of total health expenditure by 2019 (65).

A greater risk of maternal morbidity and mortality exists among pregnant women who have been diagnosed with COVID-19. There are findings showing that COVID-19 during pregnancy leads to a constant and significant increase in severe maternal health issues, mortality, and neonatal complications in comparison to pregnant women who do not have COVID-19. SARS-CoV-2 infection increases the risk of unfavorable pregnancy outcomes, as do pre-existing comorbidities such chronic hypertension, diabetes, advanced maternal age, and high body mass index (66). In Kazakhstan, living conditions and income levels have a significant association with health outcomes (67).

Our study revealed several socioeconomic and healthcare-related determinants associated with the maternal mortality in Kazakhstan in the pre- and during pandemic periods. To enhance the effectiveness of interventions aimed at reducing maternal mortality, it’s essential to systematically address these factors. Given the amount of research showing a correlation between social and economic conditions and maternal deaths, the social and economic components should receive significant attention. Addressing domestic violence prevention and improving access to education and healthcare services are critical steps in this regard. Pregnancy care coverage, the availability of nurseries, and midwifery services are examples of health care determinants that should be guaranteed. Moreover, to inform decision-making and intervention strategies, there’s a pressing need to strengthen scientific research in both public health and clinical medicine. Building a robust evidence base is fundamental for developing effective interventions and improving maternal health outcomes. By systematically addressing socioeconomic and healthcare-related determinants of maternal mortality through targeted interventions and multisectoral collaboration, significant strides towards reducing maternal mortality and improving maternal health outcomes can be done.

Our study presents several strengths, including its comprehensive coverage of regional trends in MMR in Kazakhstan and the use of statistical methods compliant with various studies and publications in the field. We pursued to unravel the role of structural and contextual predictors of MMR. The study also benefits from using officially recognized data sources, which enhances the reliability of our findings.

There are limitations that should be acknowledged. First, as we utilized secondary data obtained from annual statistical reports, there may be inherent biases in the data collection procedure. Given the concerns regarding the potential for biases in data collection and reporting, it is important to highlight that the process of diagnosing, registering, and reporting maternal mortality cases in Kazakhstan is conducted in accordance with rigorous national standards. The Rules for the Provision of Maternal Mortality Information, as outlined by the Ministry of Justice of the Republic of Kazakhstan, ensure that each case of maternal death is thoroughly documented and reported through a structured framework. This legal document is publicly accessible and provide guidelines that dictate how healthcare institutions should report instances of maternal mortality promptly and accurately (68). Additionally, the Standard of Pathological and Anatomical Diagnostics in the Republic of Kazakhstan issued by the Minister of Health further specifies the procedures and protocols for post-mortem examinations. This standard outlines the technical aspects of how maternal deaths are to be diagnosed pathologically (69). It includes criteria for autopsy practices as well as histological examinations when required. These documents collectively establish a comprehensive protocol ensuring that all cases of maternal deaths are not only appropriately diagnosed following medical best practices but also systematically recorded and reported in a manner consistent with legal requirements.

Second, potential confounding variables that might have influenced maternal mortality were not fully explored. Structural and contextual factors such as cultural practices, access to transportation, geographic barriers to care, and personal health behaviors could have significant implications on MMR but were not accounted for in our analysis.

Third, the generalizability of our findings beyond the study sample might be constrained due to unique socioeconomic, cultural, or health system factors present in Kazakhstan that may not be applicable elsewhere.