Sick leave during pregnancy: a longitudinal study of rates and risk factors in a Norwegian population

Authors

  • SK Dørheim,

    Corresponding author
    • MoodNet Research Group, Division of Psychiatry, Stavanger University Hospital, Stavanger, Norway
    Search for more papers by this author
  • B Bjorvatn,

    1. Norwegian Competence Centre for Sleep Disorders, Haukeland University Hospital, Bergen, Norway
    2. Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway
    Search for more papers by this author
  • M Eberhard-Gran

    1. Health Services Research Centre, Akershus University Hospital, Lørenskog, Norway
    2. Division of Mental Health, Norwegian Institute of Public Health, Oslo, Norway
    Search for more papers by this author

Correspondence: SK Dørheim, Division of Psychiatry, Stavanger University Hospital, P.O. Box. 8100, 4068 Stavanger, Norway. Email sdhy@sus.no

Abstract

Objective

To describe the prevalence of, reasons given for, and factors associated with sick leave during pregnancy.

Design

Longitudinal, population-based descriptive study.

Setting

Akershus University Hospital, Norway.

Population

All women scheduled to give birth at the hospital (November 2008 to April 2010).

Methods

Consenting women were handed a questionnaire at the routine ultrasound check at 17 weeks of gestation. Women returning this questionnaire received a second questionnaire at 32 weeks of gestation. Multiple logistic regression analyses were performed to examine associations with somatic, psychiatric and social factors.

Main outcome measures

Rates and duration of sick leave.

Results

By 32 weeks of gestation, 63.2% of the 2918 women included were on sick leave, and 75.3% had been on sick leave at some point during their pregnancy. Pelvic girdle pain and fatigue/sleep problems were the main reasons given for sick leave. Being on sick leave in all trimesters was strongly associated with hyperemesis, exercising less than weekly, chronic pain before or during pregnancy, infertility treatment (all P < 0.001); younger maternal age, conflicts in the workplace (both P < 0.01); multiparity, previous depression, insomnia and lower education (all P < 0.05). Sick leave was associated with elective caesarean section and higher infant birthweight (P < 0.01). Adjustment of the work situation was associated with 1 week shorter duration of sick leave.

Conclusions

Most women receive sick leave during pregnancy, but sick leave might not be caused by pregnancy alone. Previous medical and psychiatric history, work conditions and socio-economic factors need to be addressed to understand sick leave during pregnancy.

Introduction

An increasing number of women have an active professional role during their childbearing years. Pregnancy is a normal physiological state, yet it often presents physical challenges as well as alterations in mental and social wellbeing. Some women may experience the normal bodily changes associated with pregnancy as more disabling than others, and need to take time off work to recover or to prevent further disability. Others develop complications or disorders associated with pregnancy that make further work activity hazardous both to their own and to the baby's health. In most cases, there is a continuum ranging from normal pregnancy-related bodily discomfort to serious disability, and it is often not possible to determine when the women should be encouraged to stay at work and when it is better to take time off. In addition, some women may have jobs where the exposure to environmental risk factors is dangerous to the pregnancy, e.g. radiation, risk of violence, or night shifts.

The proportion of women taking sick leave during pregnancy may increase when more women of childbearing age participate in the work force.[1] Different occupational groups may have different patterns of sick leave,[2] and occupational factors may contribute to sick leave in 50% of pregnancies.[3] Job adjustment, where possible, is associated with reduced sick leave during pregnancy.[4, 5] Not all differences in sick leave are related to physical work load, but may be affected by expectations in the workplace[6] or by social conditions and attitudes.[7] Some professionals may have the impression that pregnant women are sick-listed too easily, and this may have a negative effect on the patient–doctor relationship.[8]

Opportunities for obtaining pregnancy-related sick leave may vary according to how secure one's employment status is, the level of social benefits in the society, and collective economic agreement with the employer.[9] This may explain why small businesses and sole traders have less absenteeism, in spite of increased levels of stress and fatigue.[10]

Working conditions and compensation for sick leave vary widely between countries. The World Health Report from 2010 describes the situation in 165 countries, and reports that countries compensating 80% of wages (e.g. Sweden, Slovakia) had a higher prevalence of sick leave compared with countries refunding 100% of salary (e.g. Austria, France, Germany), whereas countries with little or no compensation for sick leave (e.g. USA and UK) had fewer days absent from work.[11] Sick leave patterns may therefore be linked to sick leave entitlements in each country, but they are not fully explained by this. In Norway, social welfare benefits enable workers to receive paid (100%) sick leave from day one and for up to 52 weeks if a medical condition demands it. However, the employer is responsible for adjusting the work situation so that pregnant women may stay at work throughout their pregnancy, as pregnancy is in itself not a medical disorder.[12] To obtain sick leave in Norway, a medical certificate from a doctor is needed, documenting that the reduced ability to work is due to sickness. However, expectations from the woman, her midwife or her employer for a sick leave certificate may influence this decision. All working parents in Norway are entitled to 100% paid parental leave from 3 weeks before term until 45 weeks after delivery. At least 12 of these weeks are reserved for the father.

In spite of an increasing number of women working throughout their reproductive years, there is a limited number of studies concerning the pattern of sickness absence during pregnancy, and prevalence estimates differ according to study methodology and populations studied.[4, 13, 14] In Norway, the rate of sick leave during pregnancy has increased substantially over the period 1995–2008.[15] To guide clinicians and policy makers in how to handle requests for and possibly prevent sick leave during pregnancy, we need more population-based information about the factors associated with sick leave in this period. The Akershus Birth Cohort study, a population-based, questionnaire study with the aim of examining fear of childbirth, depression and insomnia during pregnancy, provided the possibility of studying the prevalence of sick leave during pregnancy in a large population in Norway, and examining factors associated with sick leave in this period.

Methods

Study population and design

The Akershus Birth Cohort is a longitudinal questionnaire study targeted at all women scheduled to give birth at Akershus University Hospital. The hospital is located near Oslo, the capital of Norway, and serves a population of 350 000 from urban and rural areas. The women were recruited from November 2008 until April 2010. All women scheduled to give birth at the hospital were approached at 17 weeks of gestation, when they underwent routine fetal ultrasound (n = 6244). Women were included if they were able to complete a questionnaire in Norwegian (n = 5156) and gave consent to participate (n = 4662; 152 [2.9%] did not give consent, 342 were not asked because of a logistic failure). Women returning the first questionnaire from week 17 (3752, 80.5%) received a questionnaire in at 32 weeks of gestation. Participation rate was 81.3% (n = 2943 of 3620) for this second questionnaire; representing 63.1% of the women approached at 17 weeks of gestation. Women were included if they had completed both questionnaires, and the questionnaires were linked. The final study sample consisted of 2918 women, as 25 (0.8%) had missing data for the main outcome variable (sick leave).

Women not participating in our study were significantly younger (−1.3 years, P < 0.001), and median duration of pregnancy was 1.7 days shorter (P < 0.001). Women replying to the first questionnaire only were younger (−0.4 years, P = 0.03) and were more often multipara (57% versus 50%) or had lower education (44% versus 35%) than women replying to both questionnaires. The response rate was also lower among women scoring 10 or more on the Edinburgh Postnatal Depression Scale (EPDS) in week 17; response rate 70.3% compared with 79.4% among women scoring below 10 in week 32, all P < 0.001. There were no differences in marital status.

Variables

Sick leave was measured by the following questions at 32 weeks of gestation. Have you been on sick leave during pregnancy? If yes, in what trimester(s), for how many weeks, and for what reason(s)? Multiple answers were allowed, including fatigue/sleep problems, nausea/vomiting, back pain, pelvic girdle pain, complications related to pregnancy, anxiety/depression, and ‘other’. Furthermore, we asked whether their work situation had been adjusted to accommodate for their pregnancy and if not, why: not necessary because of nonstrenuous work, impossible or nearly impossible, it was difficult to ask, I have asked but the request was not granted, and ‘other’ (free text possible). Conflicts at work were measured by asking whether or not the women had experienced problems at work the previous year (coded yes/no).

Demographic information collected included maternal age, marital status, level of education, number of previous children, twins in current pregnancy, previous pregnancies, miscarriages or terminations of pregnancy, and whether the current pregnancy was a result of infertility treatment. Hyperemesis was measured at 32 weeks of gestation by the question: Are you or have you been bothered by nausea during pregnancy (the answers ‘no’ or ‘a little’ were coded as ‘no’, and ‘very much’ was coded as ‘yes’). Additionally, we asked about the experience of ten specific stressful life events during the last year and collected information regarding smoking (coded yes/no/occasionally) and amount of regular exercise in early pregnancy (before 17 weeks of gestation) (never, less than weekly, once, twice, three times a week, or more). This was later dichotomised into ‘less than weekly’ or ‘weekly or more’.

The following information regarding outcome of delivery was collected from the birth files at the hospital: mode of delivery (vaginal, elective or emergency caesarean section), start of labour (spontaneous, induction, operative), duration of pregnancy, duration of delivery, postpartum haemorrhage, birthweight of the child and Apgar scores at 1 and 5 minutes after delivery.

The Bergen Insomnia Scale (BIS) was used to assess for insomnia according to the Diagnostic and statistical manual of mental disorders, fourth edition text revision criteria.[16, 17] Women were scored as having insomnia if they had experienced at least one criterion A (delayed sleep onset, poor sleep maintenance, early morning awakening or not feeling adequately rested after sleep) and one criterion B (daytime impairment due to sleepiness or dissatisfaction with sleep) for 3 days or more per week during the last month.

The EPDS[18, 19] was used to measure depressive symptoms at 17 and 32 weeks of gestation. A cut-off of ten or above has been used in previous studies of pregnant women in Norway to indicate probable depression,[20] and was therefore used in this study. Information on previous depression was measured by the Lifetime Major Depression Scale.[21] This scale consists of five questions (concerning sadness, appetite changes, lack of energy, self-blame and concentration). Previous depression was defined as having had at least three symptoms at the same time with a duration of at least 2 weeks. Fear of childbirth was measured by the Wijma Delivery Expectancy/Experience Questionnaire version A (W-DEQ).[22, 23] This is a 33-item self-assessment scale where each item ranges from 0 to 5, total sum ranging from 0 to 165. Serious fear of childbirth was defined as a W-DEQ sum score ≥85.[24]

Information about pain was obtained in the following way. At 17 weeks of gestation we asked about the experience of frequently recurring or chronic pain that lasted for more than 3 months before the pregnancy. At 32 weeks of gestation the questions included 1) the experience of long-lasting or frequently recurring pain during the last 2 weeks (coded yes/no); 2) current lower back pain (yes/no); 3) pain in the pelvic girdle, and if so, the location of the pain (frontal part of the pelvis, one side of the rear part, or both sides of the rear part of the pelvis). Pelvic girdle syndrome was defined as pain in all three locations combined.[25] Furthermore, we asked about the number of days the previous month during which the woman had experienced headache (0, 1, 2–6, 7–14 and >14 days). This was later dichotomised into ‘0 to 1 day’ and ‘>1 day’.

Statistical methods

Normal distributions of numerical data were tested by p–p charts. For normally distributed data, means and standard deviations were calculated, whereas for non-normally distributed data, median and interquartile ranges were used. Differences in means were tested by Student's independent t-test and differences in medians were investigated with the Mann–Whitney U test. For categorical data, statistical differences in proportions were examined with chi-square tests and logistic regression analyses. Multiple logistic regression analyses (forward and backward, testing all variables individually associated with sick leave) were performed to examine factors associated with sick leave for each trimester. All analyses were performed in spss 18.0 for Windows (SPSS Inc., Chicago, IL, USA). Significance level was set to a P value <0.05.

Ethical considerations

All women asked to participate were given written information explaining the purpose of the study and told that their participation was voluntary. Informed consent was obtained from all participants. The study was approved by the Regional Committee for Ethics in Medical Research in Norway, approval number S-08013a.

Results

Table 1 shows that 75.3% of the women reported being on sick leave at some point during the pregnancy, whereas 19.6% remained at work up until 32 weeks of gestation, and 5.1% were not in a work situation. In week 32, 63.2% of the women were receiving sick leave. Nearly 14% were on sick leave in all trimesters. The median duration of sick leave was 8 weeks, interquartile range 4–16 (total range 0–50). The main reasons given for sick leave were fatigue or sleep problems (34.7%), pelvic girdle pain (31.8%) and nausea/vomiting (23.1%) (Table 2). Very few, only 2.1% of the women, reported depression/anxiety as a reason for their sick leave, but these women had the longest median duration of sick leave (20 weeks).

Table 1. Sick leave during pregnancy
  n %
  1. Total n = 2918.

None 57219.6
Sick leave any time in pregnancy 219775.3
No paid work 1495.1
Semester of sick leave
First trimester, total79029.0
Second trimester, total105338.7
Third trimester, total172063.2
All trimesters37913.9
Table 2. Reasons given for and duration of sick leave in pregnancy
Reason for sick leave (multiple answers possible) n % of total population% of sick leaveMedian (weeks)Interquartile rangeTotal weeks (n × mean)
  1. a

    P < 0.01, Mann–Whitney U test, difference from total sample.

  2. b

    P < 0.001, Mann–Whitney U test, difference from total sample.

  3. c

    A total of 252 women with sick leave had missing values for duration.

  4. Total n = 2769.

Pelvic girdle pain88131.845.312b6–2011 629
Fatigue/sleep problems95934.749.384–1610 549
Nausea/vomiting63823.132.816b8–2410 208
Other65923.833.994–187776
Lower back pain48917.725.110a5–185965
Pregnancy complications2549.213.110a5–203327
Anxiety or depression582.13.020b12–201148
Total sample 1945c70.210084–16 

Factors associated with sick leave varied according to trimester of pregnancy (Table 3). Being on sick leave in all trimesters was strongly associated with hyperemesis, exercising less than weekly, chronic pain before or during pregnancy, and infertility treatment (all P < 0.001), younger maternal age, and conflicts in the work place (both P < 0.01), multiparity, previous depression, insomnia and lower education (all P < 0.05). Smoking in pregnancy, sum of stress factors during the previous year, or an EPDS score of ten or more at 32 weeks of gestation did not remain associated with sick leave in any trimester when adjusted for the other factors in the model.

Table 3. Factors associated with sick leave according to trimester of pregnancy, results from multiple logistic regression analyses
 Total 2769 n (%)Missing (n)All trimesters (= 318/2318) Adj. OR* (95% CI)First trimester (= 729/2522) Adj. OR* (95% CI)Second trimester (= 879/2421) Adj. OR* (95% CI)Third trimester (= 1440/2287) Adj. OR* (95% CI)
  1. *P < 0.05; **P ≤ 0.01; ***P ≤ 0.001.

  2. Odds ratios are adjusted for all variables significant in the column.

  3. –, the factor did not remain significantly associated in the backward logistic regression analysis; n.a., not in the analysis because the data related to third trimester of pregnancy.

Missing, n  451247348482
Maternal age 1250.96 (0.93–0.99)**0.96 (0.94–0.98)***0.96 (0.94–0.99)**
Hyperemesis (self-report)479 (17.3)01.93 (1.45–2.56)***2.91 (2.34–3.61)***1.75 (1.41–2.17)***
No weekly exercise832 (30.4)331.79 (1.69–2.31)***1.57 (1.30–1.90)***1.30 (1.08–1.56)**1.40 (1.15–1.72)***
Chronic pain before pregnancy302 (11.0)221.79 (1.28–2.51)***1.58 (1.20–2.08)**1.60 (1.22–2.10)**
Infertility treatment406 (14.9)391.77 (1.27–2.46)***1.40 (1.09–1.80)**1.40 (1.10–1.79)**
Chronic pain, third trimester1027 (37.8)531.66 (1.28–2.15)***n.a.n.a.1.80 (1.47–2.21)***
Conflicts at work place540 (19.5)01.47 (1.10–1.96)**1.31 (1.05–1.63)*1.34 (1.08–1.66)**1.32 (1.03–1.68)*
Multipara1373 (49.7)61.40 (1.06–1.86)*1.53 (1.27–1.86)***1.43 (1.19–1.71)**
Previous depression924 (33.6)181.37 (1.05–1.77)*1.33 (1.10–1.61)**1.39 (1.15–1.68)**
Insomnia third trimester1667 (61.9)751.35 (1.02–1.78)*n.a.n.a.1.53 (1.27–1.84)***
No high education859 (33.8)2231.32 (1.01–1.72)*1.34 (1.11–1.62)**1.26 (1.03–1.53)*
Previous miscarriage638 (23.1)61.34 (1.09–1.66)**
EPDS ≥10 at week 17301 (10.9)17n.a.1.79 (1.35–2.37)***
Pelvic girdle syndrome249 (9.0)0n.a.n.a.2.40 (1.59–3.61)***
Headache last month >1 day911 (33.3)35n.a.n.a.1.40 (1.15–1.72)***
Back pain third trimester1411 (51.2)11n.a.n.a.1.35 (1.11–1.63)**
Single status63 (2.4)1462.50 (1.24–5.02)**
Fear of childbirth216 (8.0)51n.a.n.a.1.49 (1.03–2.15)*

Table 4 displays the rates of work adjustment and reasons for work not being adjusted. A majority, 60.1% of the women, reported having had some adjustment of their work situation because of pregnancy. Women without sick leave were significantly less likely to have had their work adjusted, but this significance disappeared when coding women replying ‘Not working due to sick leave’ (n = 149) as having had no work adjustment. Most of the women who stayed in work throughout pregnancy without job adjustments responded that their work was not strenuous (73.9%). The main reason given for work not being adjusted among women with sick leave was that it was impossible or nearly impossible (44.2%). ‘Other reasons’ stated in the open-ended question included information about being on sick leave already, having twin pregnancy, being partly on sick leave as an adjustment, or that work had been partly adjusted. Women who reported job adjustments had 1 week (median value) (2 weeks when including women replying ‘already on sick leave’) shorter total duration of sick leave compared with women without job adjustments. The duration of the sick leave was longest for women who replied already being on sick leave, median values 16–17 weeks.

Table 4. Sick leave during pregnancy: work situation
 Total (= 2754)* n (%)Well (= 572) n (%)Sick leave (= 2182)* n (%)Weeks of sick leave Median (IQR)
  1. *Data were missing for 15 women.

  2. **b, chi-square test; c, Mann–Whitney U-test.

  3. ***P < 0.05; ****P ≤ 0.01; *****P ≤ 0.001.

  4. IQR, interquartile range; Y/N, – Yes/No.

Adjustment of work situation **  bbc
No950 (34.5)241 (42.1)*****709 (32.5)9 (4–18)
Yes1655 (60.1)331 (57.9)1324 (60.7)8 (4–15)***
Not working due to sick leave149 (5.4)0149 (6.8)16 (10–26)*****
Why not adjusted work? (Y/N)  = 230= 692 
Not necessary, work not strenuous303 (32.9)170 (73.9)*****133 (19.2)4 (2–7)*****
Impossible or nearly impossible339 (36.8)33 (14.3)306 (44.2)*****11 (4–20)****
It was difficult to ask55 (6.0)11 (4.8)44 (6.4)6 (3–9)****
I have asked, but not granted128 (13.9)3 (1.3)125 (18.1)*****12 (5–19)***
Other39 (4.4)13 (5.7)28 (4.0)10 (5–17)
Other—on sick leave already56 (6.1)056 (8.1)17 (12–27)*****
Conflicts at workplace 540 (19.5)70 (12.2)470 (21.4)*****12 (6–20)*****

Women being on leave in the third trimester were more likely to have elective caesarean section (P = 0.001), and less likely to have vaginal delivery (P = 0.007). Median duration of pregnancy was 1.2 days shorter (P = 0.01) for women with sick leave in the last trimester, and their infants were slightly heavier when adjusted for gestational age and maternal body mass index (P = 0.002) (Table 5).

Table 5. Pregnancy outcome for women with and without sick leave in the last trimester of pregnancy
 At work, third trimester (= 1003) 36.8%Sick leave, third trimester (= 1720)a 63.2%P valueChi-square test
Start of labour, n (%)
Spontaneous764 (76.2)1242 (72.2)0.014
Induction151 (15.1)278 (16.2)n.s.
Operative43 (4.3)124 (7.2)0.002
Mode of delivery, n (%)
Vaginal883 (88.0)1448 (84.2)0.007
Normal vaginal773 (77.0)1269 (73.8)0.06
Instrumental vaginal110 (11.0)179 (10.4)n.s.
Elective caesarean section34 (3.4)107 (6.2)0.001
Emergency caesarean section86 (8.6)164 (9.6)n.s.
Preterm (<259 days)32 (3.3)71 (4.3)n.s.
Post-term (≥294 days)98 (10.2)154 (9.4)n.s.
    Mann–Whitney U test
Duration of pregnancy (days), median (IQR) 283.2 (276.0 -288.7)282.0 (274.5 -288.1)0.01
Duration of delivery (hours), median (IQR) 5.58 (3.41-8.87)5.58 (3.25-8.53)n.s.
Postpartum haemorrhage (ml), median (IQR) 300 (200-400)300 (200-400)n.s.
    t-test
  1. a

    Data for sick leave in third trimester were missing for 46 women.

  2. b

    Data were missing for body mass index of 688 women.

  3. c

    Adjusted for maternal prepregnant body mass index and gestational age. Multiple linear regression. Unadjusted P = 0.06.

  4. IQR, interquartile range.

Birthweight (g), mean (SD) 3528 (505)3569 (556)0.002c
Maternal prepregnant body mass index, mean (SD)b 24.1 (4.5)24.7 (4.8)0.004
Apgar score 5 minutes, mean (SD) 9.58 (1.04)9.56 (1.00)n.s.

Discussion

Three out of four women had been on sick leave at some point during pregnancy, and more than six out of ten were currently on sick leave in the last trimester. Pelvic girdle pain, fatigue/sleep problems and nausea were the largest contributors to sick leave measured as total weeks away from work at the population level, whereas factors such as chronic pain and previous depression were associated with the longest duration of sick leave for the individual. Adjustments to the work situation were made in the majority of the pregnancies, and seemed to shorten the total duration of sick leave by 1–2 weeks.

Strengths and limitations

This is a large study carried out on a representative population with a high participation rate, which evaluates the women's perception of their somatic and mental health, as well as exploring other factors associated with sick leave. The nonresponders may have a different prevalence of and risk factors for sick leave than the responders. Women with depression during pregnancy were less likely to respond to the second questionnaire. This may have led to an underestimate of sick leave, especially in the second trimester. Younger women, multipara and women with lower education were also less likely to respond. However, compared with a previous Norwegian study[4] we obtained a higher response rate, including more multipara, younger women and women with less education, which enabled us to obtain comprehensive information about sick leave during pregnancy. As 17.4% of the women approached were not eligible for the study because of language difficulties, women with different ethnic or cultural backgrounds were underrepresented.

The study is based on self-reports, which may lead to bias both with respect to duration of and with respect to reasons for sick leave, as well as with respect to reported health problems and work conflicts. On the other hand, information regarding insomnia, depression, previous depression and pelvic girdle pain was obtained by previously used and reliable study scales.[16, 19, 25] The focus of the main study was fear of delivery and mental and physical symptoms during pregnancy,[26] and restrictions of space in the questionnaire limited the list of explanatory variables that could be examined with regard to sick leave. The study does not distinguish between part-time and full-time sick leave, a distinction that should be drawn in future studies. Furthermore, it does not examine whether the current sick leave was due to causes unrelated to pregnancy. Most of the associations found are cross-sectional in nature, as they were based upon information recorded at 32 weeks of gestation, and as a consequence we cannot draw conclusions about cause and effect.

Prevalence of sick leave

Most pregnant women included in the study were employed during pregnancy. We found the prevalence of sick leave considerably higher than the 51% previously reported from Norway[4] and 31% from Denmark.[14] Analogously with the previous Norwegian study we found increased rates of sick leave among younger mothers, mothers with less education and multipara. The higher rate of sick leave may in part be explained by the fact that more women from these groups were included in the present study. In contrast to the studies mentioned above we also included women with shorter durations of sick leave (<2–4 weeks), but on the other hand we did not collect data from the last weeks of pregnancy. A Swedish study found an even higher rate than the present study, with 82% of the pregnant women not working in late pregnancy, either because of sick leave (68%) or because of maternity leave (14%).[13]

Factors associated with sick leave

Pelvic girdle syndrome was the factor most strongly associated with sick leave in the last trimester when adjusted for a range of confounders such as previous chronic pain, lower back pain, insomnia, depression, socio-economic status, and conflicts in the workplace. In line with our study, Bjelland et al.[25] found that pelvic girdle syndrome is common, and is associated with disability. Compared with that study, we found a lower prevalence of pelvic girdle syndrome (9% versus 15%), but 32% of the women nevertheless reported pelvic back pain as a reason for their sick leave. These women may have unilateral pelvic pain rather than the full syndrome, and sick leave may be given to prevent further deterioration. Pelvic girdle pain and fatigue/sleep problems were the factors that contributed most to the total weeks away from work at the population level; so interventions that lead to a small decrease in prevalence or duration of sick leave among women with pelvic girdle pain or fatigue/sleep problems may still have a large impact on the total burden of sick leave among pregnant women.

Insomnia has been found to predict work disability in the general population,[27, 28] but has not been studied among pregnant women until now. High cumulative work fatigue has been identified as a strong risk factor for preterm birth.[29] Hence, sick leave may be needed for women who experience severe work-related fatigue. Both pelvic girdle pain and insomnia are associated with depression during late pregnancy.[30] Previous depression was among the factors associated with sick leave during all trimesters, and women reporting depression as the main reason for their sick leave also had the longest duration of sick leave. Psychiatric disorders, especially depression and anxiety, are among the most common reasons for long-term sick leave in the population.[31] This study therefore underscores the serious impact of insomnia and depressive disorder on the individual and on society, because these factors cause a long-lasting impairment of daily function.[32] Current depressive symptoms, however, were not related to sick leave in the last trimester, whereas fear of childbirth was. Previously, we have shown that the majority of women with fear of childbirth had neither anxiety nor depression.[33] Women with sick leave may have more reasons to fear childbirth as a result of poorer health status, more pain and less energy. However, the association remained also when adjusting for pain and insomnia.

Self-reported severe nausea was associated with sick leave not only in the first trimester, as expected,[34] but also had the highest odds ratio for sick leave in all three trimesters combined. Nausea as reason for sick leave ranked as the third largest factor contributing to total sum of weeks away from work in our study population. In a review article, Erick[35] described nausea during pregnancy as causing significant disabilities and loss of income. In addition to being a possible cause of sick leave itself, nausea could be a marker of other health impairments that occur later in pregnancy, or be a marker of stress or lack of psychosocial support. Our study underscores the importance of addressing nausea as a serious health problem during pregnancy in further studies. Infertility treatment was also strongly associated with sick leave in all trimesters, and previous miscarriage was associated with sick leave in the first trimester. Sick leave may have been used as a preventive measure among these women and be related to a fear of losing the child. However, studies evaluating the effect of bed rest on threatened miscarriage are inconclusive or have not shown this practice to be beneficial.[36] Similarly, Henriksen et al.[37] found no detrimental or beneficial effects of work on the risk of preterm delivery or the risk of the infant being small for gestational age. We found slightly lower birthweight of infants, but longer duration of pregnancy of mothers working in the third trimester of pregnancy. In western countries there is now concern about newborn infants being too heavy because of obese mothers.[38] These differences in our sample may therefore reflect that women remaining in work were healthier and had fewer risk factors for elective or preterm delivery.

We found that women reporting chronic pain before pregnancy or a low level of exercise in early pregnancy had a high odds ratio for sick leave during all trimesters. Previous sick leave related to nonpregnancy-related causes, especially chronic back pain, have been shown to be associated with long-term sick leave also during pregnancy.[14] Haakstad et al.[39] found that a low level of exercise in the third trimester was associated with pre-pregnant low levels of exercise and having no social role models with respect to exercise; this was independent of any pelvic girdle pain. That study, however, did not find any association between exercise levels in the last trimester and sick leave. Kuoppala et al.[40] found that exercise did not decrease the level of sick leave. During pregnancy, however, treatment with water gymnastics may help women with pelvic girdle pain to remain working.[41]

Work adjustment

Adjustment of the work situation was made in six out of ten women. Even so, work adjustment did not seem to be sufficient to eliminate sick leave at some point, although it was associated with a shorter duration of sick leave. We do not know if these women were also more likely to be partly away from work as opposed to having 100% sick leave. Work adjustment was often not possible. Women with sick leave during pregnancy may have more strenuous positions or exposure to risk in the work place, as Kaerlev et al.[14] found among hospital employees. In our study, we found that women with higher education were less likely to be on sick leave; this may partly be explained by these women having less physically strenuous work. They may also have jobs with more flexibility to adapt their work situation to the demands of pregnancy. We did not ask the women about type of job or type of employment (private, public, self-employed, part-time). This may have influenced the risk of sick leave and should be included in future studies. Alexanderson et al.[6] found that women in office-related occupations had lower rates of sick leave than women in occupations involving manual work. Others have found that women with heavy physical work or night shifts had an increased risk of sick leave.[4] Sick leave during pregnancy may be more strongly related to type of work than to type of symptoms.

Women staying in work throughout pregnancy reported a lower rate of conflicts in the work place. Increased risk of sick leave has been found to be related to factors such as low job control, high work level and little practical support from supervisors et al.,[14] as well as to a hectic work pace and low decision latitude.[4, 5] Wergeland and Strand[42] found that women with power to control their own work pace had better health in pregnancy than women without such powers.

Interpretation

We found a very high rate of sick leave during pregnancy. This may be a cause for concern, or it may reflect a necessary adaptation between pregnancy/family life and the demands of modern work life. A recent qualitative study from Norway found that pregnant women wanted to stay at work, but found that being pregnant while working demanded more energy than anticipated.[43] Gatrell[44] warned that a one-sided focus on decreasing sick leave during pregnancy may lead to some employed women denying their own ill health during pregnancy to avoid being associated with ‘female weakness’, placing themselves and their baby at risk. On the other hand, Sydsjö et al.[7] found that 75% of women on sick leave during pregnancy still rated their health as good or excellent. They suggested that social conditions and attitudes may explain some of the increase in sick leave among pregnant women. In Norway, the cost of sick leave benefits during pregnancy is refunded by the national welfare system. As a consequence, instead of adjusting the work to pregnancy, employers may prefer pregnant women to take sick leave so as to employ a healthy person in her vacancy. Further studies should explore the healthcare providers and employers' attitudes towards and experiences with sick leave among pregnant women.

A report from the World Health Organization found that paid sick leave was frequently linked to a high economic productivity, and that staying at work while sick may result in lower productivity for the work place and worse health for the woman.[11] Earle et al.[45] conclude that work–family policies that ensure parents to be able to care for their children's healthy development do not prevent a country from being highly economically competitive. Norwegian women do not have an increased rate of sick leave in the years after delivery for other causes than can be explained by subsequent pregnancies.[15] Women with sick leave during pregnancy are also less likely later to receive disability pension compared with women with sick leave at other times in their lives.[46] The long-term costs of sick leave during pregnancy may therefore be low, and should be kept separate from the general trends of sick leave.[47] Most women in Norway participate in the work force during their reproductive phases, and an increased rate of sick leave during pregnancy may be a necessary price to pay for this.[11, 48]

Conclusions

A great majority of women receive sick leave during pregnancy, but sick leave may not be caused by pregnancy alone. The major factors associated with sick leave were chronic pain before pregnancy, pelvic back pain and insomnia, but nausea and lack of regular exercise were also important factors. Further studies are needed to evaluate whether treatment of these conditions, both by the health services and by work adjustment may lead to less sick leave and better quality of life among pregnant women.

Disclosure of interests

The authors report no financial affiliations or other conflict of interest related to the subject of this article.

Contribution to authorship

SD, BB and MEG participated in the conception of the study and in developing the questionnaires, SD analysed the data and drafted the main text of the article, whereas BB and MEG revised the analyses and text critically for content. MEG planned and supervised the whole study, the inclusion of women, and the collection of data. The final manuscript has been approved by all authors.

Details of ethics approval

The study was approved by the Regional Committee for Ethics in Medical Research in Norway, approval number S-08013a.

Funding

The study was funded by the Norwegian Research Council, project number 191098.

Acknowledgements

The authors thank the women who volunteered their time to participate in this study. We also thank Tone Breines Simonsen, Wenche Leithe and Ishtiaq Khushi for assistance in the data collection.

Ancillary