Weight gain during pregnancy and its associated factors: A Path analysis

Abstract Aim Weight gain during pregnancy is an important indicator in the prediction of morbidity and mortality in infants and mothers. This study aimed to determine the association factors for weight gain during pregnancy. Design A longitudinal study. Methods A total of 734 women were selected using multistage cluster sampling. Data were collected using demographic and midwifery questionnaires, economic and social status, psychological factors, domestic violence, perceived social support and food insecurity. Results Of participants 28.7%, 49.6% and 21.7%, respectively, received insufficient, adequate and excessive weight gain in pregnancy respectively. Among health determinants entered in the model, mother's age, prepartum body mass index and direct and indirect prenatal care, size of households, food insecurity, stress, anxiety, stress and pregnancy‐specific stress as well as violence had a positive and increasing effect on weight gain during pregnancy. Conclusion Considering the effect of inappropriate weight gain during pregnancy on undesirable pregnancy outcomes, related factors such as nutritional status, stress and depression in prenatal care should be assessed. Healthcare providers should consult, educate pregnant women.

recommended for obese women with a BMI over 30 (Cunningham, Leveno, Bloom, Spong, & Dashe, 2014). Gestational weight gains above and below recommendations have been shown to be associated with prenatal adverse outcomes (Brawarsky et al., 2005), So that excessive weight gain during pregnancy is associated with poor neonatal outcomes like early asphyxia, birth injury and hypoglycaemia (Asvanarunat, 2014;Catalano & Shankar, 2017;Reynolds et al., 2013), large-for-gestational-age infants, while inadequate weight gain is a major risk factor for adverse foetal/neonatal outcomes such as intrauterine growth restriction, preterm birth and low birth weight (Brawarsky et al., 2005;Rooney & Schauberger, 2002). Previous studies have identified various biological and non-biological factors and modifiable behaviours as risk factors for excessive and inadequate gain (Brawarsky et al., 2005;Olson & Strawderman, 2003).
Today, health has a broader perspective and social determinants of health(SDH) are considered as the most complex and controversial issues in the field of health policy (Marmot, Allen, Bell, Bloomer, & Goldblatt, 2012). Determinants or factors affecting health include the following: (a) Social, economic and political factors, culture and social system performance, (b) Structural factors, such as education, income, gender, ethnicity and employment status, which lead to social and economic inequalities, (c) Intermediate or intermediary factors (psychosocial factors: social support, self-esteem, couples' relationship; Behavioural Factors: Unhealthy Behaviours (Violence and Addiction) and Health Systems (Sharifi, Dolatian, Kazemi Fath Nezhad, Pakzad, & Yadegari, 2018).The structural determining factors of inequalities at the level of health and psychosocial conditions are two important elements affecting health. Economic and social factors have an influence on the psychosocial, behavioural and biological factors and ultimately affect the quality of health status (Marmot et al., 2012). With a more comprehensive look at health determinants, it turns out that the problem of weight gain during pregnancy depends on several factors (Garza, 2006). Also, changes in lifestyle, nutritional behaviours and reduced physical activity have increased the prevalence of obesity (Chu, Kim, & Bish, 2009;Fathnezhad-Kazemi & Hajian, 2019). In addition, according to the study of Davis and et al, chronic stress leads to change biological behaviours, which is followed by an imbalance diet, excessive weight gain during pregnancy and postpartum obesity (Davis, Stange, & Horwitz, 2012). Also, the results of other studies suggest that psychological factors such as stress, depression, anxiety, a history of previous psychological problems and social support are associated with being overweight during pregnancy (Mehta, Siega-Riz, & Herring, 2011;de Rooij, Schene, Phillips, & Roseboom, 2010) and several studies have confirmed a link between household food insecurity and overweight, especially in women (Holben & Pheley, 2006;Martin & Ferris, 2007). Understanding the relative contribution of each group of factors to the risk of excessive and inadequate gestational weight gain is important in identifying at-risk pregnant women and designing effective interventions. Considering the importance of weight gain during pregnancy, the unknown role of its associated factors, because of the lack adequate information in Ilam province, the present study was conducted to determine the gestational weight gain status and association factors with considering Social Determinants of Health, so the key factors were identified based on reviewing literature and the conceptual framework of the WHO Commission on Social Determinants of Health (SDH), study's model was designed to investigate the factors affecting weight gain during pregnancy.

| Study design and sampling criteria
This longitudinal study was conducted with 734 women in Ilam province On August 2016 -June 2017. The inclusion criteria were, Iranian nationality, ability to read and write, singleton pregnancy, gestational age of 24-28 weeks, absence of medical problems and satisfaction for participation in the study. Individuals were excluded from the research in case of lack of collaboration to participate in the study or completion of research questionnaires. And because this second trimester is safer than other trimesters and all pregnant women go to health centres at this time to get care and deliver test results, this time was chosen to complete the questionnaires. It should be noted all health records (such as test results, weight gain at any time of pregnancy,) were recorded from the beginning until the end of pregnancy.

| Sampling and sample size
Sampling was performed using cluster random sampling. Ilam province in southwest of Iran has 10 cities. All cities in Ilam province were divided in five geographic regions (Central, North, South, East and West) which each zone was considered as a cluster that consisted of 3 or 4 health centres. We selected one urban health per cluster by using Randomizer software. Then, family health records were reviewed, and pregnant women were identified who had 24-28 gestational weeks and were selected randomly by using Randomizer software. Initially, they were called and invited to participate in the study. In case of not participating in the first study session, or not met other inclusion criteria, other eligible participants were requested to join the study according to the list to complete the final sample size. After explaining the research aims and confidentiality of information, informed consent was obtained from the samples and the questionnaires were completed by the participants. Participants completed the questionnaires in a meeting room at the health centre for 30 min. Due to a large number of questionnaires, completing of them lasted from 20 to 35 min.

| Measurement tools
Data were collected by questionnaires of demographic and midwifery profiles, structural health determinants (socioeconomic status), questionnaires to measure intermediate determinants of health including food insecurity, social support, stress, anxiety and depression, pregnancy-specific stress and domestic violence. In the present study, the Socio-demographic and obstetrics characteristics questionnaire was designed by the researcher. Validity and reliability of other questionnaire have been investigated in Iran and are available in Persian.

| Socio-Demographic and obstetrics questionnaire
The questionnaire was designed by the researcher and included questions about the age of the pregnant woman, the age of the spouse, the ethnicity, the gestational age, the number of pregnancies, the interval between pregnancies, the status of incidence of chronic diseases (diabetes, hypertension), consumption of supplements and so on. The section of economic and social status consists of the level of education of the pregnant mother, husband's education, woman's occupation, husband's job, number of people living at home, monthly household income, etc., the method of face and content validity was used to determine the validity of the questionnaire.

| Pregnancy-specific stress questionnaire
The questionnaire consists of 25 phrases in 6 subsets, including maternal, infant, childbirth and motherhood experience, child and maternal, personal-family and personal-occupational interests. The questionnaire has been designed based on Likert scale of 5 options with a minimum and maximum scores of 0 and of 100, respectively.
The score of the questionnaire indicates the level of concern of pregnant women and its worrying factors. Validity and reliability in Iran have been confirmed by Navidpour et al. (Navidpour, Dolatian, Yaghmaei, Majd, & Hashemi, 2015).

| Domestic violence questionnaire
The World Health Organization (WHO) has designed a questionnaire to measure the violence of a spouse in pregnancy and the rate of physical, psychological or sexual violence is calculated based on the Likert scale of 5 states. The questionnaire assesses physical, sexual and emotional violence with 10, 5 and 11 questions, respectively.
Individuals who have at least one positive answer to any of the questionnaire questions are considered to be violent. The Cronbach's alpha coefficient of the questionnaire in Iran has been calculated to be 0.92, 0.89, 0.88 for the three questionnaire areas of the physical, psychological and sexual violence (Hajian, Vakilian, Najm-abadi, Hajian, & Jalalian, 2014).

| Household food insecurity access scale (HFIAS)
HFIAS consists of 9 questions and 4 options in the frequency of occurrence (most often, sometimes, rarely and never) which provided information on food insecurity in terms of access to food at the household level. The lowest score for each question was zero while the highest score for each one was 3. The validity of this questionnaire was verified by Mohammadi et al. using face, content and structure validity. Cronbach's alpha of 0.86 in the questionnaire shows its high internal stability (Mohammadi et al., 2012).
Reliability of the questionnaire was determined by using test-retest method after conducting a pilot study on 30 pregnant women.
Internal consistency (Cronbach's alpha coefficient) was determined for DASS-21 standard questionnaire, it was 0.89, 0.92 and 0.88 for domain of depression, anxiety and stress respectively and Cronbach's alpha coefficient was 0.88, 0.86, 0.88, 0.89 for pregnancy-specific stress questionnaire, domestic violence questionnaire, multidimensional perceived social support scale and household food insecurity access scale, respectively.
In the current study, a conceptual model was designed to determine associated factors with weight gain during pregnancy; factors such as contextual factors, socioeconomic status, food insecurity, perceived social support, domestic Violence, stress and anxiety and depression and stressful life events, pregnancy care ( Figure 1). Path analysis method was used to examine model good fit and also percentage coverage of variance.

| Statistical analysis
After collecting data, SPSS19 software (IBM© SPSS© Statistics version 19 IBM© Corp.) and Lisrel 8.8 were used to analyse the data. Descriptive statistics were used to determine the frequency, percentage, mean and standard deviation. Chi-square and Fisher's analytical statistics were used to assess the relationship between determinants and weight gain. Fisher statistical test was used if an expected count in the table was less than 5. Also, the path analysis was used to determine the most important variable affecting weight gain and to specify the direct and indirect effects of variables.

| RE SULTS
The mean age of the samples was 28.7 (SD 4.4) years, and the mean age of the spouses was 33.4 (SD 5.5) years. All samples were Iranian, 81.6% of the participants had Kurdish ethnicities. 46.3% of the participants had their first pregnancy, and 86.6% of them had wanted pregnancy and 13.4% of pregnancies were unwanted or unplanned.
About half of the samples (50.7%) and their spouses (48.2%) had university education. Most women were housewives (86.2%), and more than half of their spouses (64%) were self-employed. In addition, the number of employees in each family was one individual in most of the families (86. 8%). The income of half of households (57.4%) was 10-20 million Rials ( income, average household cost and household size in three groups (p > .05) ( Table 2). In examining the relationship between intermediate determinants and weight gain in pregnancy, it was observed that there was a significant difference among food insecurity, social support and prenatal care in three groups of insufficient weight gain, sufficient weight gain and excess weight gain (p < .05).
People with food insecurity had too much weight gain in pregnancy. Sufficient and high levels of overweight in pregnancy were also found in pregnant women with high social support. Also, there was no significant difference in the general violence among three groups of people with insufficient weight gain, sufficient weight gain and overweight (p = .99) ( Table 3).
Among the structural determinants, variables of mother's age, mother's education, spouse's education and household size had an impact on weight gain in pregnancy. Among these determinants, only the mother's age from direct and indirect paths had an effect on weight gain. From the direct path, mothers with older ages had reduced weight gain in pregnancy (β = 0.08) and from indirect path, due to the impact of mother's BMI (β = 0.12), BMI had a negative effect on the rate of weight gain in pregnancy (β = −0.08). Female and her husband's education had a negative and decreasing effect on high overweight, as the rate of overweight dropped with the increase in the level of education (Figure 2 and Table 4).
Among intermediate health determinants, two variables of mother's prepartum BMI and the condition of pregnancy cares had a direct effect on weight gain. The greatest impact was observed in the amount of pregnancy care and thus, individuals with more care had more weight gain (Table 4).
In an indirect path, food insecurity variable had the greatest impact on excessive overweight and had a positive and increasing effect on weight gain (β = 1.087) and excessive weight gain in pregnancy was augmented with increasing food insecurity. Food insecurity from three indirect routes had an effect on symptoms of depression, stress, tension and pregnancy-specific concerns as well as pregnancy weight gain, and BMI contributes to excessive overweight in pregnancy (   (Table 4). Criteria used to fit the model included goodness of fit index, normalized fit index, comparative fit index and root-mean-square error. In this model, criteria were appropriate, and the model fit was good (Table 5).

| D ISCUSS I ON
In the present study, 2.8% of the samples were lean, 57.7% had normal BMI and 39.6% had prepartum overweight. In the study of Chen et al., 13.2% of women had overweight, 3.6% of them were obese and 15.8% were lean before pregnancy (Chen et al., 2010). The mean weight gain in pregnancy was 11.27 ± 3.23.
28.7%, 49.6% and 21.7% of participants respectively had insufficient, sufficient and excessive overweight in pregnancy. In the study of Deputy et al, 20.9%, 32% and 47.2% of women had insufficient, sufficient and excessive overweight in pregnancy, respectively (Deputy, Sharma, Kim, & Hinkle, 2015).  & Story, 2010). In the study of Davis et al., biological, metabolic, socioeconomic and psychological factors were identified as determinants of weight gain in pregnancy (Davis et al., 2012). In the present study, rate of weight gain was higher with increasing age and prepartum BMI, which was consistent with the results of studies by Mohammadi et al., (2011);Panahandeh, (2009). In a research by Wells et al., inadequate weight gain during pregnancy was associated with underweight and obesity of mother, rural housing, low level of education, smoking and excessive weight gain while excessive overweight was related to obesity of mother and 12 years of education or less (Wells, Schwalberg, Noonan, & Gabor, 2006).
The results of path analysis showed that mother's age and education, husband's education and household size among structural determinants had an impact on overweight in pregnancy. Among these determinants, only the mother's age had an impact on weight gain both from direct and indirect paths. Individual's education and her husband's education also had a negative and decreasing effect on excessive overweight. Rate of excessive weight gain in pregnancy reduced with increased levels of education. In the study by Delaram et al., no relationship was observed between occupation of mother and overweight in pregnancy but education had a significant relation with weight gain in pregnancy and those with higher education had higher average weight gain (Delaram & Akbari, 2008 (Herring et al., 2008;Mehta et al., 2011).
In a review of Hartley et al, there was a significant relationship between pregnancy weight gain and depression as well as social support but no relationship was observed between stress, anxiety, self-confidence and excessive weight gain (Herring et al., 2008).
Also, the results of path analysis indicated that among intermediate determinants of health, the prepartum BMI of the mother and the condition of pregnancy both had a direct effect on weight gain and the greatest impact was on the amount of pregnancy care so that those with more care had more weight gain. Other determinants also influenced the weight gain from indirect path, where the food insecurity had the greatest impact on weight gain during pregnancy. The results of present study were consistent with the outcomes of the study by Laraia et al and women with food insecurity had a higher pre-pregnancy BMI and their weight gain was higher during pregnancy (Laraia, Siega-Riz, & Gundersen, 2010).
Studies by Martin et al. and also Holben et al. confirmed the relationship between excessive overweight and household insecurity, especially in women (Holben & Pheley, 2006;Martin & Ferris, 2007) and, consuming low-energy foods with high energy density but low value of micronutrients is more in people with food insecurity, leading to obesity and overweight (Hill et al., 2013). One of the strengths of current study is to consider several factors of social determinants of health for the purpose of the relation assessment. Also, food insecurity is one of the most important factors that little studies have been done in this regard during pregnancy so far. Therefore, the results of present study can be used as a basis for other studies.

| CON CLUS ION
Given the results, half of women had enough weight gain, it seems that food security, high social support and increased the number of prenatal care are associated with a normal weight gain in pregnancy. Other variables, such as symptoms of stress, anxiety, depression, pregnancy-specific stress and violence, had an indirect effect on weight gain in pregnancy. Therefore, the need to educate mothers about weight gain in pregnancy and its effects on inappropriate outcomes of pregnancy should be considered and its related factors such as nutritional status of pregnant women, level of stress, depression and other related factors should be investigated.

| Limitations
The limitations of the study can be attributed to a large number of questionnaires that were taking a long time to complete the questionnaires. Using some data that contained in the prenatal care file such as weight gain in each prenatal care.

E TH I C A L A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
This study was approved by the Ethics Committee of SBMU, Grant no: SBMU.REC.1394.112 as a PhD dissertation. Written and informed consent was obtained from the participants, and they were told they could withdraw at any time during the study.