In recent years, women's roles in industrialized society have changed radically. Women pursue formal education, work outside the home, and delayed child bearing. Not only do women work outside the home during early pregnancy, but the proportion of pregnant women continuing to work in the last trimester has also increased.1,2
It has been argued that work outside the home during pregnancy might be harmful to the pregnant woman's health by contributing to pregnancy complications and thus also would have negative effects for the outcomes of the pregnancy. It was assumed that pregnant women employed in physically demanding work that required, for instance, prolonged standing, long work hours, and shift work would be affected.3–5 Therefore, to support these women, many nations around the world provide paid maternity leave and health benefits by law, but some countries, including the United States, Australia, and New Zealand, do not.6
Sweden has generous benefit programs for pregnant women.7,8 Since 1974, Sweden has provided, besides paid sick leave, a general parental benefit program for all pregnant women. This benefit has been continuously extended to 450 days in 1989.9 In 1980, an additional benefit program, the pregnancy benefit program was introduced to help pregnant women employed in especially arduous or tiring occupations. Since the introduction of the pregnancy benefit in 1980, this benefit has been granted to almost 50% of the pregnant women who are occupied in industrial types of occupations or health and care type of occupations. The pregnancy benefit comprises paid leave from work for 50 days before delivery and is granted by the local social authorities after an application. The pregnancy benefit is not related to any pregnancy complication or illness.9–11 The benefits are economically provided in the same way that benefits would be provided if the pregnant woman were to be sick-listed. The aim of this study was therefore to investigate whether infants' weight showed any correlation with changes in the availability and the actual use of social benefits during pregnancy.
MATERIALS AND METHODS
To obtain a pregnant population representative of a regular Swedish health care region, we combined samples of delivered women in an already existing database on sickness absence from 2 hospitals in the southeastern region of Sweden, Linköping, a University Hospital (120,000 inhabitants) and Värnamo, a County Hospital (85,000 inhabitants). In Värnamo, all delivered women for the 4 years were included, but because Linköping had far more deliveries, every second consecutively delivered woman was included on the assumption that they were randomly collected.12 A total of 7 459 delivered women were initially included for the years 1978, 1986, 1992, and 1997 with regard to pregnancy and outcome as well as occupation and benefits paid by society. The study years were chosen because sickness absence in Sweden increased up to 1986, but then receded in 1992–1997. The material and results are presented in detail elsewhere.12
All information concerning the type of occupation and the degree of employment during pregnancy was based on the pregnant woman's own statement at her first visit to the antenatal care unit. Because close to 25 different occupations were registered, the occupations were merged into 4 main groups:
- Administrative work: Bank teller, postal clerk, secretary, bank officer, accountant, etc.
- Industrial work: Assembler, farming, weaver, manual worker, truck driver, etc.
- Service: Waitress, food counter, hairdresser, cleaner, etc.
- Health and care: Registered nurse, auxiliary nurse, child care worker, etc.
Information on sickness absence and the use of parental and the pregnancy benefits was collected from the delivered women's files at the local social insurance office. All kinds of rest or leave provided by and paid by society were merged in this study and presented as whole days to facilitate comparisons. Information from the files at the delivery ward on the status of the newborn infants was also added to the database. All information was then rendered anonymous. Due to missing values in some files on the children's birth weights (4.1%), gestational length (3.4%), as well as the mothers' smoking habits during pregnancy (6.5%), the number of observations included in the analyses was reduced to 6,850 women (91.8% of the original sample). The dropouts (8.2%), however, were evenly distributed for the study years and are not considered to have any effect on the results.
A national insurance program covers all Swedish residents aged older than 16 years. Housewives, students, and unemployed persons are also covered. The first 7 days of absence are self-certified. In this study, only days of sickness absence supported by a physician's certificate were included. The parental benefit was expanded from 270 days in 1978 to 450 days in 1989. Sixty days of this benefit may be used for rest before delivery; however, used days are then subtracted from the total number of days. The pregnancy benefit program was introduced in 1980 to provide extra support for employed pregnant women (50 days in total) whose occupations were considered particularly monotonous or strenuous and whose employer could not transfer her to a more suitable working position. The additional granting of the pregnancy benefit does not influence the available number of days in the parental benefit.9
The economic coverage in the various social benefit programs was practically 100% up to 1992 and was then, due to an economic recession, lowered to 80% of the pregnant woman's salary.
The results are presented using percent or the average with corresponding 95% confidence intervals (95 % CI). In Tables 1 and 2 and the corresponding text, the categorical variables were analyzed by means of the χ2 test, whereas quantitative variables such as the women's age and the children's birth weight and gestational length were analyzed by means of 1-way analysis of variance. An error rate below 0.05 was considered as statistically significant. Analysis of covariance was used to estimate the effect of all the studied variables on the children's birth weight during the study period (Table 3). In these analyses, the categorical variables were included as factors, whereas the women's age as well as the children's gestational length were included as covariates. Because some of the factors included, such as the proportion of smokers, were not constant during the period of study, interactions between these variables and the year of birth were tested in the model. However, none of these interactions were significant (the lowest P value was observed for the interaction between social benefits and year of study, P = .094) and were thus not included in the model presented.
Approval of the project's aims and methods was obtained from the Regional Ethics Committee on Human Research Projects at the Linköping University.
During the years of the study (ie, 1978, 1986, 1992, and 1997), the proportion of primiparas among the studied pregnant women remained virtually constant (χ2 = 5.533, P = .137) (Table 1), as did the proportion of twin pregnancies. There was, however, a significant decline of smokers during the study period; from 32.5% in 1978 to 14.1% in 1997 (χ2 = 201.383, P < .001). The mean age of the pregnant women increased significantly during the study period from 27.2 in 1978 to 29.3 in 1997 (F = 54.614, P < .001), whereas the percentage of employed women increased between 1978 and 1986, but then receded between 1986 and 1997 (χ2 = 132.401, P < .001). The main changes observed were due to the increase in the percentage of women who by definition could not be included in any of the 4 occupational groups.
The proportion of women who received social benefits during pregnancy (i.e. who received the parental benefit and/or the pregnancy benefit, and/or who were sick-listed) increased between 1978 and 1986, but then decreased from 1986 and onwards (χ2 = 245.350, P < .001) (Table 1). The proportion of women sick-listed during the studied period increased from 35.6% in 1978 to 58.5% in 1986 and then receded to 26.1% in 1997 (χ2 = 424.160, P < .001), whereas the percentage of women who used days of their parental benefit during pregnancy decreased from 56.0% in 1978 to 39.0% in 1997 (χ2 = 120.605, P < .001). In contrast, the proportion of women granted the pregnancy benefit increased from 14.4% in 1986 to 20.7% in 1997 (χ2 = 25.253, P < .001). In 1978, no pregnancy benefit was available. For all pregnant and employed women, the total number of days of benefits per woman covered by the society (sickness absence, the parental benefit, and the pregnancy benefit) remained approximately constant in 1978, 1992 and 1997 (35.8, 40.6 and 36.9 days respectively). Only in 1986 was there a significant increase to 54.8 days, caused by the much higher average number of days of sickness absence during this year (P < .001).
During the study period, a continuous and significant increase in birth weight took place among the children to the women studied (Table 2 and Fig. 1). The average increase in birth weight between 1978 and 1997 was approximately 82 g for all women studied, but the increase was higher among working pregnant women than among women not employed during pregnancy. The mean increase in birth weight between 1978 and 1997 among working pregnant women was 101 g (F = 10.027, P < .001) compared with 15 g among the unemployed (F = 0.766, P = .533). The duration of pregnancy decreased significantly during the study period (F = 31.023, P < .001; all data not shown). However, this decrease was only evident between the years 1978 and 1986, during which time the average gestational length decreased from 39.7 to 39.3 weeks. From 1986 onward the mean gestational length remained constant. The decrease in gestational age between the first 2 study years is probably explained by the change of the pregnancy dating method. In 1978 pregnancies were still dated with reference to the date of the last menstrual period, but from 1986 onward all pregnancies were dated by the more reliable method of ultrasonography.13
Although the model presented in Table 3 was able to explain some of the variance in birth weight during the study period (adjusted R2 = 0.312), Figure 1 shows that social benefits during pregnancy as well as other previously known influencing factors do not explain the observed trend of increasing birth weights during the study period. Table 3 shows that receiving social benefits during pregnancy was not of significant influence on the children's birth weight (P = .107), even though adjustments were made for several influencing factors. As expected, women with no previous children gave birth to children with lower average birth weights than women with previous children, and mothers who smoked during pregnancy tended to give birth to children with lower birth weights. In addition, the mother's age when giving birth as well as the gestational length was related to the birth weight of the studied children. However, the mother's occupational status during pregnancy did not seem to influence the children's birth weights (P = .563). The results did not change substantially when restricting the analyses to women working during pregnancy, with the exception that social benefits were less significantly associated with the outcome than in the analysis presented (P = .507; data not shown).
There seems to be a common opinion that working pregnant women benefit from a period of rest.6,10 As more and more Swedish women work outside the home, a concomitant increase in sickness absence has been observed reaching a peak in 1989.14 The frequency of taking sick leave has been found to be especially high among pregnant women, but was observed to decrease as among all insured persons when the level of economic compensation provided by the social security department was cut in the 1990s due to an economic recession.12,15
A general assumption is that social benefits in connection with pregnancy will help to prolong gestation and therefore lead to an increase in infant's weight.16,17 However, home is not necessarily a place for rest.18 If the outcomes of pregnancy were correlated or sensitive to social benefits among working pregnant women, then, for instance, birth weight data should show changes parallel to changes in the use of such benefits. We found a steady and significant increase in the mean birth weight of the infants and a rather stable duration of the pregnancy after 1986, independent of the amount of social support used.
Although many studies underline the importance of lightening the burden of a pregnant woman, few actually deal with the outcomes of such welfare or social laws even if such assessment of welfare policies is considered as essential, as proposed for instance by Wise et al19 in 1999. Only 1 study examined sickness absence in connection with the outcomes of pregnancy, a study of the perinatal outcome of twin pregnancies.20 The authors found no obvious association and concluded that the effect of sickness absence on prenatal outcome may be less important than previously thought.20 Our results are also in accordance with official statistics. Although sickness absence has both increased and decreased among pregnant women over the past 2 decades in Sweden, a concomitant increase in birth weight among the newborn has taken place.14 Recent studies also report an increase in birth weight of neonates in other Nordic countries21 that do have an ample social welfare system, although not as generous in several aspects as in Sweden.8,22
We found that the type of occupation did not have any explanatory value in the analysis performed. When looking at risks of working during pregnancy it must be taken into account that the character of occupational hazards may change with time, and some hazards may even gradually disappear, as modern technology successively is introduced in a workplace.23 In Sweden, a modern western welfare country, few occupations probably seem to be so extreme as to cause a threat to the pregnant women or her child. In a survey of 3,451 infants on education, socioeconomic status, and work environment, it was found that practically all social differences in birth weight are related to maternal age, parity, height, and smoking habits but not to the work environment.24
In our study, factors like smoking, age, and parity correlated with the observed increase in birth weight. The proportion of smokers in the Swedish pregnant population has been successively reduced and is used as 1 of the main explanatory factors for the general increase in birth weight in Sweden.25 In this study, a major reduction of reported smokers took place during the study period. We also found an increase in age, most clearly expressed among the women in the administrative type of occupation, but no difference in parity over time.
Obesity has grown to an alarming problem in the western hemisphere, and studies have been carried out to examine the relation between the mother's weight and the weight of the infant.26 We have not incorporated maternal weight for all of the women in this study, because this initially was not the aim of this study. However, an increase in body mass index has been observed in the same region up to the year 2001 (Brynhildsen J, Sydsjo A, Norinder E, Selling KE, Sydsjo G. Trends in body mass index during early pregnancy. In press). Thus, the influence of a general increase in body mass index among pregnant women on the results in this study therefore cannot be excluded.
Our study is based upon all consecutively delivered women at 2 separate delivery wards during 4 years in the interval 1987–1997. Although some women (8.2%) could not be included because we lacked information on some of the measures required for inclusion in the calculations in this study, the combination of benefits and birth weight still yielded valuable information. In this context, the well-organized Swedish Antenatal Care must be mentioned. It is free of charge and attended by practically all pregnant women in Sweden. Positive influence on the outcomes of pregnancy may well be a result of this.27 We have not tried to separate the parental and the pregnancy benefits from the sickness absence benefit, because sickness absence during pregnancy in many cases can result from a desire to get rest rather than from actual illness, and therefore, we simply focused on the total number of days of absence covered by the society in the separate benefits systems.28,29
A factor that may influence birth weight or pregnancy length is that obstetric handling of pregnancies may have changed over time. In vitro fertilization pregnancies, for instance, may lead to an increased proportion of twins, delivered prematurely.30 Because the outcomes of pregnancies are more favorable due to improved care at neonatal wards, obstetricians may also choose to deliver pregnancies at risk for a premature child, thus making any comparisons over time difficult. The high ratio of elective caesarean delivery and induced deliveries on pregnant women's request also contributes to changes in pregnancy length and, indirectly, the infant's weight.30 These factors, however, work toward a general lowering of both pregnancy length and the infant's weight and may underestimate the real results in this study.
We found no evidence of a relationship between the social benefits granted and the infant's birth weight. Thus, the effects of social benefit programs on the outcomes of pregnancy may be overrated in a western society. Although the mother's parity, age, and smoking habits were significantly related to the birth weight of her offspring, maternal weight most probably is also of importance. It is not yet clear whether this is an advantage for the mother or her offspring, because weight gain per se is not entirely favorable from either an obstetric or a pediatric point of view.
1.Seguin RE. Pregnancy and the working woman: a review. J Ark Med Soc 1998;95:115–8.
2.Sorenson DL, Tschetter L. Reasons for employment or non-employment during pregnancy. Health Care for Women Int 1994;15:453–63.
3.Brown MA. Employment during pregnancy: influences on women's health and social support. In: Swanson-Kaussman C, editor. Women's work, families, and health. Washington (DC): Hemisphere Publishing Corporation; 1987. p. 151–67.
4.Gabbe SG, Turner LP. Reproductive hazards of the American lifestyle: work during pregnancy. Am J Obstet Gynecol 1997;176:826–32.
5.Saurel-Cubizolles MJ, Kaminski M. Work in pregnancy: its evolving relationship with perinatal outcome (a review). Soc Sci Med 1986;22:431–42.
6.Commission on Family and Medical Leave. A workable balance: report to Congress on family and medical leave policies. Washington, DC: U.S. Department of Labor, Women's Bureau; 1996.
7.Saurel-Cubizolles MJ, Romito P, Garcia J. Description of maternity rights for working women in France, Italy and in the United Kingdom. Eur J Public Health 1993;3:48–53.
8.Hakansson A. Different benefits among pregnant women in Europe [in Swedish]. Läkartidningen 1996;93:1654–9.
9.Riksförsäkringsverket. Föräldraförsäkringen 1987-1989. Stockholm (Sweden); 1991.
10.Wergeland E, Strand K. Need for job adjustment in pregnancy. Early prediction based on work history. Scand J Prim Health Care 1998;16:90–4.
11.McGovern PM, Gjerdingen DK, Froberg DG. The parental leave debate: implications for policy relevant research. Women Health 1992;18:97–118.
12.Sydsjo G, Sydsjo A, Wijma B. Variations in sickness absence and use of social benefits among pregnant women in a Swedish community 1978–1997. Acta Obstet Gynecol Scand 1999;78:383–7.
13.Selbing A, Fjällbrant B. Accuracy of conceptual age estimation from fetal crown-rump length. J Clin Ultrasound 1984;12:343–6.
14.The National Board of Health and Welfare. Sweden's public health report 1997. Stockholm (Sweden); 1998.
15.Dembe AE. Changes in Swedish Workers' Compensation: An American Perspective. Int J Occup Environ Health 1997;3:144–49.
16.Chavkin W. Work and pregnancy: Review of the literature and policy discussion. Obstet Gynecol Surv 1986;41:467–72.
17.Launer LJ, Villar J, Kestler E, de Onis M. The effect of maternal work on fetal growth and duration of pregnancy: a prospective study. Br J Obstet Gynaecol 1990;97:62–70.
18.Romito P. Woman's paid and unpaid work during pregnancy. A psycho-social analysis. London (UK): Thomas Coram Research Unit; 1989.
19.Wise P, Chavkin W, Romero D. Assessing the effects of welfare reform policies on reproductive and infant health. Am J Public Health 1999;89:1514–21.
20.Rydhstroem H, Bakketeig L, Magnus P, Knudsen LB, Wedel H. Perinatal outcome after leave of absence from work for twin-pregnant women. Acta Genet Med Gemellol (Roma) 1997;46:175–83.
21.Orskou J, Kesmodel U, Henriksen TB, Secher NJ. An increasing proportion of infants weigh more than 4000 grams at birth. Acta Obstet Gynecol Scand 2001;80:931–6.
22.Sydsjo A, Sydsjo G, Kjessler B. Sick leave and social benefits during pregnancy–a Swedish-Norwegian comparison. Acta Obstet Gynecol Scand 1997;76:748–54.
23.The back pain epidemic. Acta Orthop Scand 1989;60:633–4
24.Nordstrom ML, Cnattingius S. Effects on birthweights of maternal education, socio-economic status, and work-related characteristics. Scand J Soc Med 1996;24:55–61.
25.Cnattingius S, Lindmark G, Meirik O. Who continues to smoke while pregnant? J Epidemiol Community Health 1992;46:218–21.
26.Rössner S, Ohlin A. Maternal body weight and relation to birth weight. Acta Obstet Gynecol Scand 1990;69:475–8.
27.Lindmark G. Hälsovård före, under och efter graviditeten. The Swedish Medical Board of Health. Stockholm, Sweden: SoS Rapport; 1996. p. 7.
28.Grünfeld B, Qvigstad E. Disease during pregnancy. Sick-listing among pregnant women in Oslo [in Norwegian]. Tidsskr Nor Lægeforen 1991;111:1269–72.
29.Lindmark G. [Women should be provided with other leave possibilities during pregnancy than sick-listing]. Läkartidningen 1992;89:1882–4.
30.Daltveit AK, Vollset SE, Skjaerven R, Irgens LM. Impact of multiple births and elective deliveries on the trends in low birth weight in Norway, 1967-1995. Am J Epidemiol 1999;149:1128–33.