Introduction

Pregnant women undergo important anatomical, physiological and psychological changes that often challenge job demands, especially at the end of pregnancy^1,2. The transition from active work to maternity license after delivery depends on a delicate balance where many factors intervene, both work and non-work related^3,4. In Spain, as in other countries with a consolidated social protection benefits scheme⁵, two optional types of social benefits have been developed to protect pregnant women and their foetus’s health. These benefits are aimed at guaranteeing the continuity of women’s labour relationship and to cover their salary if absences from work due to health issues are needed. One of them is the sickness absence (SA) benefit, applied when the absence is due to a health problem not related to working conditions. The other one is the pregnancy occupational risk (POR) benefit, available in only a few countries^6,7 and applied when absence from work is needed to prevent exposure during pregnancy to occupational risk factors that cannot be avoided by job adjustments or other measures.

Sick leave among pregnant workers is a frequently discussed issue as their SA rates are usually high^8,9. According to previous studies, three out of four pregnant women had at least one episode of SA^10,11,12. Other studies have tried to identify predictors of SA during pregnancy and factors that could increase the probability of returning to work after a SA episode, such as job adjustments, among others^13,14. Some studies have pointed out that there are not clear medical explanations for this phenomenon^15,16,17. However, health related problems such as musculoskeletal pain before¹⁸ or related to pregnancy¹⁹, overweight²⁰, and other factors including sedentary lifestyle²¹, tobacco, alcohol or other drugs consumption ^22,23,24, women’s "double burden"¹⁷ and attitudes towards SA²⁵ have been associated with SA during pregnancy. Furthermore, despite SA is a specific benefit to cope ill-health non-work related, exposure to some occupational risk factors, such as heavy load handling, forced postures or high psychosocial demands, have shown a significant influence on SA in workers during pregnancy ²⁶,^{27 28 29}.

In line with this situation, the study about sickness absence diagnoses and occupational risk factors influences on pregnant workers disability are needed to help women to stay at work in a sustainable, healthy and productive way.

Our hypothesis was that absences due to SA are not influenced by exposure to occupational risk factors but to factors not related to work, mainly to the pregnancy itself. The aim of the present analysis was to assess the association between SA and POR benefit according to information on medical diagnoses and exposure to occupational risk factors in a cohort of pregnant healthcare workers.

Methods

Results

A total of 655 episodes of SA were recorded among the 367 included pregnant workers, amounting a total of 28,313 AD and a median duration of 19 days (Table 1). The most frequent medical diagnosis group was health problems related to pregnancy (232 episodes, 35.4%), including mainly excessive vomiting (71 episodes, 2,641 AD), risk of abortive outcome (62 episodes, 3,644 AD) and risk related to delivery (39 episodes, 2,524 AD), accounting for a total of 12,039 AD (42.5%). The second most frequent group was musculoskeletal disorders (199 episodes, 30.4%), being low back pain the leading cause (178 episodes, 9,906 AD).

SA trajectories are shown in Figure 1. Trajectory 1 includes women with SA due to musculoskeletal medical diagnoses as the main cause of absence from work, whereas women in trajectory 2 were absent from work because of SA due to pregnancy-related disorders. POR was the main cause of work absence in those women included in trajectory 3, combined with some SA mainly due to pregnancy-related medical diagnoses.

Figure 1. Sickness absence (SA) trajectories among a cohort of pregnant workers with at least one sickness absence episode during pregnancy (N=367), 2010-2014 

Most women are fitted in trajectory 1, where 216 pregnant workers (58.9%) accounting for 14,942 days of SA (389 episodes) with a median duration of SA episodes of 15 days. Trajectory 2 included 92 pregnant workers (25.1%), with 12,414 days of SA (171 episodes) and a median duration of 24 days. There were only 59 women (16.1%) in trajectory 3 accounting for 6,403 days of absence due to POR (58 episodes) and only 957 days of SA (95 episodes), with a median duration of 112 and 4 days, respectively. Active work accounted for 75%, 48% and 55% of pregnancies time in trajectory 1, 2 and 3, respectively (Table 2).

Almost half of pregnancies in trajectory 3 (47.5%) had the highest score (12 to 18) of exposure to occupational risk factors. Conversely, this percentage was 22.7% and 33.7% for pregnancies in trajectories 1 and 2, respectively, and more than 40% of these women showed a low level of exposure to occupational risk factors (6 to 9). Statistically significant differences were observed between the three trajectories for the occupational risk score, occupation and working shift (Table 2).

The association between exposure to occupational risk factors and each trajectory separately is shown in Table 3. Pregnant workers in trajectory 1 had a negative association with safety and ergonomic risk factors (RR 0.56, 95% CI 0.35-0.90; RR 0.50, 95% CI 0.33-0.77, respectively), and with the highest global risk exposure index (score 12 to 18) (RR 0.68, 95% CI 0.49-0.96). Trajectory 3 was associated with safety (RR 2.75, 95% CI 1.59-4.75) and ergonomic risks (RR 3.64, 95% CI 2.18-6.06). The highest score of occupational risk factors exposure (score 12 to 18) was associated with trajectory 3 (RR 2.69, 95% CI 1.43-5.01). Trajectory 2 did not show any association with exposure to risk factors at work nor with any exposure score.

Regarding occupation and taking physicians as the reference category, nursing aides and nurses had a higher probability of belonging to trajectory 3 (RR 5.58, 95% CI 2.09-14.95 and RR 5.00, 95% CI 1.95-12.81, respectively) and less likely than physicians to belong to trajectory 1 (RR 0.55, 95% CI 0.36-0.86; RR 0.57, 95% CI 0.41-0.81, respectively). Split-shift, compared to morning shift, was associated with trajectory 1 (RR 1.42, 95% CI 1.03-1.97) and negatively associated with trajectory 2 (RR 0.47, 95% CI 0.28-0.79), and both the afternoon and the night shifts were associated with trajectory 3 (RR 2.26, 95% CI 1.14-4.45; RR 2.53, 95% CI 1.11-5.78, respectively). No differences were observed for maternal age, type of contract and number of weekly work hours.

After adjusting for maternal age, type of contract and working shift, most associations there are some changes (Table 4). In particular, trajectory 1 was negatively associated with exposure to safety (RR=0.67, 95% CI 0.46-0.97), ergonomic (RR=0.59, 95% CI 0.42-0.82) and psychosocial (RR=0.77, 95% CI 0.62-0.95) risks. Trajectory 2 did not show any association, except for exposure to psychosocial risks (RR=2.56, 95% CI 1.47-4.46).

The stratified analysis by occupation (Table 5) showed that for physicians trajectory 2 was significantly associated with exposure to chemical and safety risks (RR=3.40, 95% CI 1.39-8.34 and RR=2.59, 95% CI 1.03-6.49, respectively); for nursing aides trajectory 3 was significantly associated with exposure to ergonomic risks (RR=6.15, 95% CI 1.99-19.03); and for nurses trajectory 3 was significantly associated with exposure to physical (RR=3.10, 95% CI 1.11-8.66), safety (RR=1.99, 95% CI 1.12-3.54), ergonomic (RR=1.79, 95% CI 1.03-3.11) and psychosocial risks (RR=3.19, 95% CI 1.58-6.43).

Discussion

Pregnancy-related health problems and musculoskeletal disorders represented up to 80% of absence days due to SA in our cohort of pregnant women, being low back pain, excessive vomiting and risk of abortive outcome the most frequent particular medical diagnoses and showing a negative or no association with exposure to occupational risk factors, respectively. Logically, we also found that women who were highly exposed to occupational risk factors, mainly biological, ergonomic, safety, hygiene and/or psychosocial, had absences from work covered predominantly by POR benefit, with one or more previous SA episodes mainly attributed to pregnancy-related medical conditions early in the pregnancy.

Our results could be explained because the POR benefit was correctly used to prevent health problems that could be caused or aggravated by adverse working conditions, such as musculoskeletal disorders in women exposed to high exposure to ergonomic risk factors. These patterns were confirmed after adjusting by age, occupation, type of contract and working shift. Those women belonging to the trajectory with SA episodes caused mainly by musculoskeletal disorders where not associated to ergonomic and psicosocial risk factors, as it would be expected. A hypothesis is that these SA episodes due to musculoskeletal disorders could be directly consequence of the pregnancy or/and non-work ergonomic risk factors. No previous study has to our knowledge included this information.

An important result from our study is that absences from work covered by SA during pregnancy were not associated to exposure to occupational risk factors. However, for physicians, we found an association between exposure to chemical and safety risks with the SA trajectory with episodes due to health problems related to pregnancy. A possible explanation to these findings is that working conditions may play a role in worsening symptoms or diseases related to pregnancy.

The frequency of SA shown in our cohort (85%) was higher in comparison with other previous studies, where reported SA during pregnancy varies from 29% to 72%^33,34,35. A possible explanation may be due to differences in SA definition since some authors only consider SA episodes as those lasting more than one week or just consider those that occur only in certain weeks of pregnancy. We have registered all episodes from the first day and during the whole pregnancy period.

The main medical diagnosis of SA were pregnancy-related health problems, which confirms their important role contributing to the observed increase of SA during female reproductive age³⁶. Previous studies had also shown that other diagnosis such as pelvic girdle pain and fatigue/sleep problems as the main reasons of SA^11,37. Musculoskeletal complaints contributed substantially to SA during pregnancy in our study, being the second most frequent cause of absences from work. In fact, this is similar to the frequency of musculoskeletal disorders in the whole working population, for both men and women^38,39. As reported in several studies^11,40-42, low back pain was the most frequent reason of SA in our sample.

Another important implication from our results is the importance of the two current social benefits in Spain to cope with imbalances between work and health during pregnancy. On the one side, SA benefit is mainly used when pregnant workers suffer from health problems mainly related to pregnancy and/or other no-work related risk factors; on the other, POR benefit is applied when they are exposed to occupational risks before health effects occur. Our results show the suitable complementarity of the two benefits, and suggest that SA seems to be used also as a complementary way to assess the balance between work demands and physiologic changes.

This study has several strengths. First, it is based on a detailed follow-up throughout pregnancy with a precise daily register of active work and absences by SA, with or without POR. In addition, for each pregnancy we obtained detailed information on medical diagnoses of all SA from medical and SA records and a comprehensive occupational risk assessment. However, the interpretation of our results should be done with caution because of some limitations. The main one is related to the sample size and the characteristics of the study population, reducing the external validation of our results. Our sample corresponds to pregnant workers of a public university hospital who maintained their full salary during either of the two benefits as a result of an agreement with the unions. There is a need to reproduce the study in other companies with difference SA benefits schemes (usually only 70% for salary in SA benefits unless otherwise complemented, and 100% for POR benefits in Spain). In addition, most pregnant workers (76%) included in our study had a permanent contract. Nonetheless, we did not observe any differences of the magnitude of the association by type of contract in our sample. In any case, it would be interesting to analyse the same trajectories in companies of other sizes and different workforces and sectors, and with a higher proportion of temporary contracts.

In summary, a provisional conclusion is that current social benefits seem to be adequate for protecting pregnant workers from exposure to occupational risk factors, and to cope with imbalances between working conditions, health-related problems and pregnancy. Nevertheless, a great majority of pregnant women in our cohort had absences covered by SA benefit, some of them before initiating the POR benefit towards the third trimester of pregnancy. A great majority of women in our cohort had absences covered by SA benefit that could not be caused by pregnancy itself, but also, as we have seen, to working conditions⁴³. Maybe a better management of occupational risk assessment and information provided to pregnant workers, stressing the need of POR benefit when occupational exposures are high, could reduce episodes of SA. Further research therefore needs to focus on the effects of working conditions improvement and the role of specialized preventive interventions that may help pregnant workers to ensure an adequate balance between motherhood and active work.