Background

Several studies have investigated risk factors for poor psychological and physiologic health status of women during pregnancy and after delivery [1, 2]. Studies that focus on maternal health during pregnancy could show that there are significant negative associations between educational and occupation level and the prevalence of anemia in pregnant women [3], as well as between poverty and race and the occurrence of urinary tract infection, placenta disorders and preterm rupture of the membranes [4]. But not only the maternal health status during pregnancy but also the health status after delivery has been investigated earlier. Studies showed that higher maternal educational level and higher maternal occupational class as well as regular physical activity were associated with better maternal and better child health [5,6,7,8]. A study in Minnesota demonstrated that both physical and mental health problems 11 weeks postpartum (measured by the Short Form Questionnaire “SF-12” [9]) occurred more often in women with low coworker support and high job stress [10]. As a result, in particular the mental health status of mothers is considered at least as important as their physical health status. It could be shown that partnership and parity status, educational status, monthly income and residential property status were factors influencing depressive symptoms in mothers after delivery [11]. In addition, a study demonstrated that having children with chronic health conditions or developmental disabilities or activity limitations leads to a higher risk for adverse maternal mental health outcomes after delivery [12].

In general, previous studies focused mainly on one specific predictor for health outcomes whereas relatively little is known about the joint effects of numerous social determinants in relation to maternal health. In this study we use a comprehensive concept of health, including general, physical and mental health. A simultaneous consideration of socio-economic, environmental, lifestyle, psychosocial and birth related factors might help to identify women at risk for poor general, physical and mental health status four weeks after delivery. The current work extends prior work by examining the health status of mothers as well as a range of socioeconomic, environmental, lifestyle, psychosocial and birth and child related factors influencing maternal health outcomes using data from a prospective German birth cohort study.

Methods

Study design and study population

The study is based on data from the KUNOFootnote 1-kids health study, a population-based, prospective birth cohort study conducted at the hospital St. Hedwig in Regensburg, Germany. St. Hedwig is located in Regensburg (153,000 inhabitants). The catchment area comprises the city of Regensburg as well as its mostly rural adjacent regions. Unemployment rates are low [13]. Being a tertiary perinatal center, the proportion of high-risk births is large. Rationale and design of the study have been reported in detail elsewhere [14]. For our study all women who came to St. Hedwig for the delivery of their baby (or babies) from June 2015 until June 2018 were asked to participate. Overall response was about 33% [14]. Participation was voluntary and written informed consent was obtained for each case. Participating women were asked to complete a standardized questionnaire in addition to a physical examination of the child. Researchers were trained to conduct interviews during a pilot study, using data not included in the final analysis. Exclusion criteria were inadequate German language skills (no basic German language skills for the comprehension of study procedures existent) and underage mother (< 18 years). The study was approved by the ethics board of the University of Regensburg (file number: 14–101-0347).

Instruments

Assessment of social determinants

Maternal social determinants were assessed through a standardized questionnaire completed by mothers and, in addition, assessed by the KUNO-kids study team shortly after delivery. In addition, follow-up questionnaires were sent to the mothers four weeks after delivery. Socioeconomic determinants including age, family status, type of medical insurance, maternal education and employment before maternity leave, occupational group and maternal nationality were assessed within two days after delivery, as well as the level of social support and lifestyle factors, such as the level of physical activity, diet, height and weight before pregnancy, the number of doctors’ consultations during pregnancy and the history of psychiatric and somatic diseases. Further, birth related determinants, including birth mode, preterm delivery and season of birth were assessed right after delivery. Psychosocial factors such as the subjective position within society, the level of parental stress and social and emotional strains were assessed four weeks after delivery, as well as environmental and lifestyle factors, such as the housing situation, the drinking and smoking habits and secondhand smoke exposure. Child related determinants, including a visual analogue scale of the estimated child health “VAS child health”Footnote 2 and the factor “breastfeeding” were also assessed four weeks after delivery.

We used the German version of the Mac-Arthur scale [15] to examine the mother’s subjective position within society. The level of social support (as experienced by the mother) was measured using the standardized short form social support questionnaire “F-SozU K-14”Footnote 3 [16] [17]. It comprises the following domains: emotional and practical support, social integration, perceived social support and social strain. Every item can be answered on a scale of one (total rejection of the statement) to five (total agreement with the statement) out of which a mean score is calculated. To determine parental stress we used three scales of the EBIFootnote 4 [18] a German version of the parenting stress index (PSI): parental competence, personal limitations and parent-child bond were each assessed with four items with a range of one to five points. Adding up the points for each domain results in a possible range of four to twenty points, with a higher score indicating a higher stress level. A total score was not calculated, as we only used three domains out of the original 12 domains.

For detailed information on all items, that have been used to assess social determinants see Table 1 and additional file 1.

Table 1 Assessment of socioeconomic, lifestyle, environmental, psychosocial and child and birth-related determinants
Full size table

Assessment of health outcome

Self-rated maternal health

Self-rated health was assessed four weeks after delivery on a visual analogue scale (VAS) ranging from 0 to 100 with 100 points indicating full health. The VAS is part of the EQ-5DFootnote 5 [19].

Questionnaire assessed maternal health

Physical and mental health status of the mothers was determined by means of the “SF-12-questionnaire” four weeks after delivery [9]. This questionnaire is a validated tool that reflects the overall health-related quality of life [20]. The SF-12 asks how they would describe their general health status (excellent, very good, good, fair, poor). They were also asked if they experienced any new limitations with moderate daily activities, if they had difficulties accomplishing daily activities because of emotional or physical health problems, if they experienced any pain that would interfere with their normal work and how often they felt calm, energetic or disheartened. It was also asked how often physical health or emotional problems interfered with social activities during the last four weeks. SF-12 scores for both the physical and the mental component core range between 0 and 100 points with scores above and below 50 being above and below the average German population. A score above 50 represents a better physical or mental health than the population norm, respectively [21].

Statistical analyses

The baseline characteristics of the study population were summarized descriptively. Linear regression was conducted to determine the relationship between health outcomes and potentially influencing factors. B-values and standardized ß-values were generated. Each variable with a p-value < 0.20 in univariable analysis was entered into a multivariable model to avoid missing variables that may show significant effects in multivariable analysis despite evident lack of associations in univariable analysis. For multivariable analysis the significance level was set to a p-value of < 0.05. All analyses were performed using SPSS 23 (IBM Corp., Armonk, NY, USA 2015). Mothers with missing data were excluded from the respective analyses.

Results

Social and health characteristics of the sample of mothers are shown in Table 2. A total of 1428 of the 2788 women (51%) had completed the interview and the questionnaires right after delivery in addition to the questionnaire four weeks later and were therefore included in the analysis. Four weeks after delivery, the overall average self-rated health of mothers was 86.17 points (SD: 11.90), physical and mental component scores of SF-12 were 47.63 points (SD: 8.22) and 47.28 (SD: 10.64), respectively. The scores ranged from 19.02 to 68.53 (physical component scale) and from 7.87 to 65.84 (mental component scale).

Table 2 Characteristics of the study population
Full size table

Univariable analysis

Results of univariable analyses are presented in Tables 3 and 4.

Table 3 Associations between self-rated health and potential determinants in mothers: results of univariable linear regression analysis
Full size table
Table 4 Associations between SF-12-PCS and -MCS and potential determinants in mothers: results of univariable linear regression analysis
Full size table

Self-rated health

Breastfeeding, the mother’s subjective position within society, the level of social support (F-Sozu) and VAS child health showed significant positive associations with mothers’ self-rated health four weeks after delivery. Social and emotional strains, obesity, having experienced a C-section, a history of both psychiatric and somatic diseases, parental stress (PSI) and the number of doctors’ consultations during pregnancy were inversely associated.

SF-12-physical component scale (PCS)

The mother’s age, education of more than ten years, social and emotional strains, parental stress (PSI), humidity stains inside the house or flat, number of doctors’ consultations during pregnancy, having experienced a C-section, having the first child and a positive history of psychiatric and somatic diseases were negatively associated with the PCS-12 four weeks after delivery.

SF-12-mental component scale (MCS)

We found significant positive associations between mother’s age, health insurance, education more than ten years, mother’s subjective position within society, social support (F-Sozu), VAS child health and the MCS-12 four weeks after delivery: The older the mothers, the higher their subjective position within society, the higher their social support and the healthier mothers estimated their child, the better was their mental health status. Also having health insurance and having had an education of more than ten years showed better mental health status. Employment before maternity leave, migration background, social and emotional strains, parental stress (PSI), a history of somatic and psychiatric diseases and having the first child were negatively associated.

Multivariable analysis

Self-rated health

Table 5 shows significant associations between determinants and the self-rated health of mothers four weeks after delivery. Breastfeeding and VAS child health remained as significant positive predictors in the multivariable analysis. Social and emotional strains, two of three domains of the PSI (personal limitations, parental competence), obesity, having experienced a C-section, and a positive history of somatic diseases were negatively associated with self-rated maternal health. In addition, education of more than ten years was also negatively associated with self-rated health of mothers.

Table 5 Associations between self-rated health and potential determinants in mothers: results of multivariable linear regression analysis
Full size table

SF-12-physical component scale (PCS)

Age, employment before maternity leave, social and emotional strains, one PSI domain (personal limitations), C-section, having the first child and a history of somatic diseases were negatively associated with the PCS-12. For detailed information see Table 6.

Table 6 Associations between SF-12-PCS and potential determinants in mothers: Results of multivariable linear regression analysis
Full size table

SF-12-mental component scale (MCS)

Education of more than 10 years and a solid social environment, indicated by a high score in the F-Sozu, showed significant positive associations with MCS-12. Social and emotional strains as well as two of three PSI domains (personal limitations and parental competence) showed significant negative associations with the MCS-12 (Table 7).

Table 7 Association between SF-12-MCS and potential determinants in mothers: Results of multivariable linear regression analysis
Full size table

Discussion

Summary of main findings

The aim of our study was to provide a contribution to the literature on how a multitude of socio-economic, lifestyle, environmental, psychosocial and birth related determinants effects maternal health. We identified risk factors for maternal health outcomes four weeks after delivery, especially social and emotional strains, parental stress and education of more than 10 years. We could also determine positive factors for maternal health, such as a high subjective position within society or a high score in the VAS child.

Comparison of findings with previous studies

Some of our findings are consistent with earlier studies, emphasizing the importance of a stable social environment for mothers’ health [1, 6]. Dennis et al. identified a lack of social support as a risk factor for mothers’ self-rated health. Also having had a C-section (or forceps or vacuum delivery), no or little physical activity and low income were negatively associated [1]. Some of our findings on the other hand seem to be surprising in the context of prior work. Education more than ten years appears to have a significant negative influence on one’s self-rated health. Former studies show that the self-rated health varies with perception of what constitutes good health and with expectations regarding one’s own health [22]. It is assumed that more highly educated individuals are better informed of treatment options and are less tolerant of a given health condition [23]. A study in Germany in 2010 showed that the higher the educational level of women, the higher the number of doctor’s consultations [24]. An explanation for this relation might be that the more knowledge mothers have, the more they tend to be worried about any complications and therefore report a lower self-rated health. Our finding warrants further investigation though.

We took the physical and mental component scale of the SF-12-questionnaire to picture the physical and mental health status of the mothers. Previous studies using the SF-12 found that job stress, amongst others, was a negative predictor for mental health status of mothers [10]. Interestingly, we found that employment before maternity leave was associated with a lower physical health status in mothers. This finding might inform a reconsideration of maternity protection in Germany.

Strengths and limitations

Our work adds valuable new findings to already existing studies and knowledge about maternal health. First, we identified several potential risk factors for maternal health that have not been the focus of previous research, including a high maternal education or employment before maternity leave. To the best of our knowledge it is one of the first studies investigating the simultaneous influence of a multitude of factors, not focusing on only specific factors. It is important to understand how socioeconomic, environmental, lifestyle, psychosocial and birth related factors together influence maternal health outcomes. Health is a complex multidimensional construct. In our study we used a comprehensive approach to describe maternal health four weeks after delivery including the self-rated health of mothers on a visual analogue scale and the physical and mental component scales of the SF-12, a well-established and validated tool to measure health-related quality of life [9].

Previous studies tend to dichotomize self-rated health into two groups of poor/fair self-rated health and good/excellent self-rated health or use other items to describe it, like the Personal Health Scale (PHS) or the General Health Questionnaire (GHQ) [1, 25]. The visual analogue scale we used with a response range from 0 to 100 offers very fine nuances and prevents respondents of being bound to predetermined categories [26].

Some limitations of this study must be considered.

All of our data are self-reported and therefore subject to recall bias. To overcome this limitation, mothers were interviewed shortly after delivery and again after four weeks in order to potentially improve the mothers’ recall of events. All members of the study team were trained to conduct the interview in a professional and sensitive way. Still information may have been misreported due to social desirability bias [27].

The response rate in our study is 33% (investigated in detail during a 2.5 months’ time period) [14]. This seems low, compared to other earlier German birth cohort studies, such as the ULM-Spatz birth cohort study with a response rate of 48% or the GINIplus birth cohort study with a response rate of 55% [28]. Still, response rates between studies vary. Other birth cohort studies in Europe, such as the Danish national birth cohort study have achieved a similar response rate of 35% [29] and also data from the currently ongoing German National Cohort NAKO suggests that participation rates in population based studies are low [30]. In the KUNO-kids study potential differences between respondents and non-respondents have been investigated. The most frequent reasons for non-participation were insufficient German language skills, mothers’ perception that the study procedures are associated with too much effort and lack of interest [14]. For the analysis of determinants of maternal health we investigated differences between the initial sample of mothers who provided written informed consent and participated in the baseline interview and the final study population (who also returned the four-weeks-questionnaire). The rate of single mothers and mothers with migration background in the final study sample is lower compared to the original sample. Educational status and rate of employment before maternity leave in the final sample on the other hand are higher compared to the original study population (see additional file 2). Compared to data from registration offices our final study population has a lower rate of single mothers and mothers with migration background [13]. However, educational level in our study is similar to the educational level of women aged between 25 and 35 years in Germany [31] and child and birth related characteristics in our study reflect the average perinatal statistics in Bavaria with similar ratios of C-section or preterm delivery [32]. In summary, we cannot deny the presence of selection bias in our study. In the context of prior birth cohort studies this observation does not seem too surprising [33] as people from a higher social class tend to participate in health studies more often than people from a lower social class.

Conclusion

We noticed that some risk factors for maternal health become more and more common, such as a high educational level for women, employment before maternity leave and high-aged primipara. As it could be shown in our study, these factors are associated with lower maternal self-rated health and a lower physical health status four weeks after delivery which underlines the importance of reinvestigating these relations. It appears interesting that we found an association between higher education and a lower self-rated health. This is against all previous findings and should definitively lead to more studies examining this relation.

By investigating several factors and their influence on maternal health the importance of mothers’ social support became clear. Social and emotional strains as well as parental stress seem to be negatively associated with both the self-rated health and the physical and mental health status of mothers. In this context the enforcement of social support networks, as well as stable family structures seem important. Concretely, it could reduce maternal stress and strains to establish family-friendly work concepts. Particularly in rural areas there is the need to extend social support offers, such as educational counseling or cry baby clinics. This could improve maternal health outcomes in the future.