Introduction

Preterm birth and small-for-gestational-age (SGA) in infants is a major adverse birth outcome. SGA rate differs largely depending on regions and tends to be high in developing countries [1, 2]. Although preterm birth rate globally stands at approximately 10% [2], the rate in Japan is much lower [3]. However, preterm birth is a risk factor for perinatal and infant mortality [4, 5], and adversely effects neurodevelopment or the incidence of cardiovascular diseases [6]. Infants with SGA, which is the other adverse birth outcome and a focus of epidemiological studies [7, 8], have an elevated risk of delayed development or childhood mortality [9,10,11]. Therefore, prevention of these adverse birth outcomes is a basic medical and societal need.

It is known that some characteristics of mothers, such as maternal age, smoking habits, and maternal socioeconomic status, are risk factors of adverse birth outcomes [12,13,14]. In Japan, epidemiological studies have shown that the mother’s pre-pregnancy body mass index (BMI) and the maternal level of education were risk factors for preterm and SGA births [15,16,17,18]. In addition, occupation is another major socioeconomic characteristic of parents. The association between maternal occupation and outcomes, such as infant mortality or low birth weight, has been investigated in Japan [19, 20]. However, there has been no systematic study for the relationship between the mother’s occupations and preterm and SGA births, using nationwide government statistics data in Japan. One study on maternal occupations in Sweden showed that “mechanics and iron and metalware workers” had a higher risk of giving birth to infants with SGA [21], and difference was also investigated and observed also in Korea and the United States [22, 23]. By identifying occupations that confer high risk for these adverse birth outcomes to parents also in Japan, we might be able to provide targeted support to people with those occupations at their workplace or in their communities.

To address this goal, we used the Vital Statistics data in Japan to investigate the association between maternal occupations and preterm and SGA births.

Main text

Methods

Data source and data preprocessing

We obtained nationwide individual-level data of the Report of Vital Statistics: Occupational and Industrial Aspects in the 2015 fiscal year from the Japanese Ministry of Health, Labor, and Wealth. Use of the data for research was permitted by the Ministry based on Article 33 of the Statistics Act. The survey for the report is conducted every 5 years, and the data include infants which were born between April 1, 2015 and March 31, 2016. The Report corresponds to the Vital Statistics with occupational and industry information. Vital Statistics is a survey by the Japanese government, and information from the birth certificates of all over the country are gathered by the Ministry of Health, Labor, and Wealth. It is mandatory for a parent to submit birth certificates of their children to a municipality within 14 days after childbirth. The birth certificate is a self-administered questionnaire, while health-related information, such as birthweight and gestational age, are written by physicians or midwives at the time of childbirth. The same data have been analyzed in previous studies in Japan [19, 24], and the maternal occupations at the time of birth were surveyed. We used the birth data, which included maternal age, maternal nationality, wedlock status, sex, birth weight, number of fetuses, gestational age, parity, household occupation, and mother’s occupation for each infant. Categories for maternal occupation in the data were 12 types, which was based on the Standard Occupational Classification for Japan. Household occupation means type of main occupation for each household, which is classified as farmer, self-employed worker, full-time worker 1, full-time worker 2, others, and unemployed. These six types of household occupations are the choices listed in the birth certificate for decades, and the parents must choose one from them.

Preterm birth was defined as gestational age below 37 weeks. SGA is often defined as a birth weight below the 10th percentile of the sex- and gestational age-specific standard reference value [8, 25]. In Japan, previous studies of SGA births generally used a sex- parity- and gestational age-specific standard reference for birthweight based on a Japanese neonatal anthropometric chart [7, 15, 26, 27]. Following this precedent, we defined SGA infants as those with birth weight below the 10th percentile of the same standard reference. For this study, we excluded multiples, and restricted the analysis to singleton infants. In addition, infants who were born between 22 and 41 weeks were used for the analysis of SGA births because the standard reference for birthweight in Japan covers only gestational ages of those periods [26].

Statistical analysis

We tallied the number of births, preterm births, and SGA births and preterm and SGA birth rates for each infant characteristic. We defined six disjoint groups for maternal age (≤ 19 years, 20–24 years, 25–29 years, 30–34 years, 35–39 years, and ≥ 40 years), three groups for gestational age (< 32 weeks, 32–36 weeks, and ≥ 37 weeks), and three groups for birth weight (< 1500 g, 1500–2499 g, ≥2500 g). Parity for infants was categorized into primiparous and multiparous. Maternal nationality was classified into Japanese and non-Japanese. Chi-square test was conducted between each of the birth characteristics and each of the outcomes. In addition, the relative risk of preterm and SGA birth for each type of occupation was calculated using a log binomial regression model with other characteristics as covariates. The log binomial regression model is a method used for calculating the relative risks for binary data [28, 29]. Unemployed persons were used as a reference in the regression analysis for maternal occupations because the number of births was the largest. We reported the relative risk, its 95% confidence interval, and the p-value.

Complete-case analysis was conducted in the chi-square test and the regression analysis for dealing with missing data. In addition, multiple imputation was used as a sensitivity analysis for dealing with missing data in the regression analysis [30]. All statistical analyses were performed by using the statistical software R (version 4.6.1) [31]. In addition, statistics shown in this study were analyzed by the author using the data provided from the Ministry of Health, Labor and Wealth, and those are different from statistics published by the Ministry.

Results

Figure 1 shows the flowchart of selecting birth data for analysis.

Fig. 1
figure 1

Flowchart of selecting birth data

Full size image

Table 1 shows number of live births, preterm births, and SGA births and rates of preterm and SGA births for each of the characteristics. The greatest number of births was registered for unemployed persons among maternal occupations. The preterm and SGA rates were particularly high for mothers of age 40 and older. Among maternal occupations, preterm birth rate was highest among carrying, cleaning, packaging, and related workers (5.65%) and lowest among security workers (4.24%). SGA birth rate was highest among manufacturing process workers (5.91%) and lowest among security workers (4.00%). Most of the characteristics such as maternal occupations were statistically significantly associated with the birth outcomes.

Table 1 Number of live births, preterm births, and SGA births and rates of preterm and SGA births for each of the characteristics
Full size table

Table 2 shows analysis results of regression analysis showing relative risk of maternal occupations for preterm and SGA birth. We found significantly elevated risks for preterm birth among manufacturing process workers compared with unemployed mothers, and significantly elevated risks for SGA birth compared with unemployed mothers were observed among sales workers, service workers, and manufacturing process workers. In contrast, security workers had a significantly decreased risk for SGA birth compared with unemployed mothers. In addition, the relative risk was the highest level among manufacturing process workers both for preterm and SGA births.

Table 2 Analysis results of regression analysis showing relative risk of maternal occupations for preterm and SGA birth
Full size table

Supplementary Table 1 shows analysis results of regression analysis showing relative risk of maternal occupations for preterm and SGA birth using multiple imputation. Similar results were obtained as the main analysis, while a significant risk ratio for SGA birth was not observed for sales and service workers.

Discussion

An association between the adverse birth outcomes and maternal occupations was revealed from an analysis of nationwide Vital Statistics data in Japan. Specifically, we have found that the risk of preterm and SGA births varied depending on maternal occupations. We now discuss the possible reasons for those associations.

Regarding the mother’s occupation, it was suggested that the risk for manufacturing process workers was the highest. The elevated risk for adverse birth outcomes among manufacturing process workers might indirectly reflect the prevalence of lower socioeconomic status among Japanese female manufacturing process workers. It is known that, in Japan, the proportion of university graduates is relatively low and the proportion of junior high school or elementary school graduates is relatively high for manufacturing process workers in the occupational classifications among women aged 20–49 years old [32]. The association of maternal education level or socioeconomic status with preterm birth and SGA is known [33, 34]. Specifically, lower socioeconomic status of the mother is positively associated with higher rate of preterm and SGA births in Japan [16, 18]. Furthermore, lower socioeconomic status is associated with the presence of medical risk factors, such as high smoking rate and low utilization of prenatal care, for adverse birth outcomes in Japan [35, 36]. However, the risk was not elevated among other types of workers whose educational level were relatively low, such as transport and machine operating workers, indicating that education level and socioeconomic status are not sufficient to fully explain the observations. Instead, other factors, such as employment status or occupation-specific factors could, in part, explain the elevated risk for preterm and SGA births among manufacturing process workers or other workers (service workers and sales workers). For example, long standing time at work or frequent night shifts are also known to be associated with the adverse birth outcomes [37,38,39], and it might have affected the results. In addition, security workers had the lowest risk for SGA births. As one possibility, physical characteristics of those workers might be related to the results. It is known that maternal height or BMI affects SGA births, and lower height or BMI is associated with higher risk of SGA birth [40,41,42]. Generally, adequate body weight and height are required for becoming security workers, such as police officers and self-defense officers, and it might have favorable effects on birth outcomes. An association of maternal occupations with socioeconomic and physical characteristics is needed to be surveyed in order to identify the reasons for the difference in risks.

One strength of this study is that we used a large set of nationwide birth data for the analysis, and the study results represent nationwide trends in Japan. In particular, the association between maternal occupations and SGA and preterm birth is demonstrated for the first time using nationwide Vital Statistics. We specifically highlighted a higher risk of preterm birth and SGA for mothers employed as manufacturing process workers. This finding has important implications. First, future studies are needed to investigate the causes for the observed association. Second, there is a need for implementing effective prevention measures to lower the risk of adverse birth outcomes. Adjusting the work load or improving working conditions might be needed to lower those risks for pregnant manufacturing process workers. A call for attention in working places or prenatal care for pregnant women with those occupations might help. Finally, providing health guidance targeted to husbands of pregnant women whose occupations have higher risk of adverse birth outcomes might be also effective.

Limitations

One limitation of this study is the lack of data on women’s behaviors or physical characteristics, such as smoking status, BMI, utilization of prenatal care, chronic diseases, and working hours, that are known risk factors for adverse birth outcomes. Another limitation is the indirect inference of maternal educational level or physical characteristics because of the absence of such data in the dataset we analyzed. Socioeconomic status indicators of a mother, such as educational level or household income were not available in the data, and we could not fully differentiate the effect of maternal occupations from that of maternal socioeconomic status. Thus, to properly account for those factors that the survey does not collect, an epidemiological study designed to target those factors needs to be conducted in the future.