Association between endometriosis and working life among Danish women

Endometriosis is a systemic disease ( Taylor et al. , 2021 ) estimated to affect 10% of women of reproductive age ( Nnoaham et al. , 2011 ; Shafrir et al. , 2018 ; Zondervan et al. , 2020 ), although estimates are reported to range between 2% and 11% ( Shafrir et al. , 2018 ; Zondervan et al. , 2020 ; Rowlands et al. , 2022 ). The disease is characterized by growth of endometrium-like tissue outside the uterus ( Kennedy et al. , 2005 ; Vercellini et al. , 2014 ; Zondervan et al. , 2020 ) often causing chronic and severe pain in the pelvic region, deep dyspareunia, dysuria, dyschezia, dysmenorrhea, fatigue, and infertility ( Nnoaham et al. , 2011 ; Gallagher et al. , 2018 ; Rush et al. , 2019 ; Zondervan et al. , 2020 ). These debilitating symptoms affect everyday life and productivity including mental, physical, social, and sexual well-being for affected women ( Soliman et al. , 2017 ; Sperschneider et al. , 2019 ; Zondervan et al. , 2020 ). In addition, endometriosis is associated with diagnostic delay and underdiagnosis ( Nnoaham et al. , 2011 ; Sperschneider et al. , 2019 ; Zondervan et al. , 2020 ; Mundo-Lopez et al. , 2021 ) which challenge the estimation of impact and consequences of endometriosis both at individual and societal level. Previous studies have investigated the impact of endometriosis and related symptoms on work life, productivity, and/or sick leave ( Gilmour et al. , 2008 ; Fourquet et al. , 2011 ; Nnoaham et al. , 2011 ; Hansen et al. , 2013 ; Rowlands et al. , 2022 ; Soliman et al., 2017 ; Facchin et al. , 2019 ; Sperschneider et al. , 2019 ; Estes et al. , 2020 ; Mundo-Lopez et al. , 2021 ; Rossi et al. , 2021 ; Bell et al. , 2023 ). Among these, three studies from Denmark, Finland, and Australia found that endometriosis was associated with poorer work ability, work disturbances or absenteeism, and more sick days or disability days ( Bell et al. , 2023 ; Hansen et al ., 2013 ; Rossi et al. , 2021 ). The Danish study found an association between diagnostic delay and reduced work ability ( Hansen et al. , 2013 ). A study from the USA found that women with endometriosis had a higher risk of short- and long-term work loss (leave of absence, short- and long-term disability, and early retirement) ( Estes et al. , 2020 ). Another study from the USA found higher productivity loss (both for employment and household chores) with increasing patient-reported symptom severity and number of endometriosis symptoms ( Soliman et al. , 2017 ). Some studies quantified actual working time lost due to endometriosis during a week based on questionnaire data but with no reference group for comparison ( Fourquet et al. , 2011 ). One study across 10 countries found that women with endometriosis lost 10.8 h/week on average due to both absence from work and reduced productivity ( Nnoaham et al. , 2011 ). A study from Puerto Rico found that women with endometriosis lost 7.4 h/week on average during a week when symptoms were at their worst and 43% reported that productivity was extremely affected by endometriosis symptoms ( Fourquet et al. , 2011 ). These are important findings for a disease affecting women during their most productive years ( Fourquet et al. , 2011 ; Mundo-Lopez et al. , 2021 ) from early education and career building to retirement ( Rossi et al. , 2021 ). Previous studies have primarily been based on self-reported questionnaire data and/or smaller samples of women, and no studies have yet quantified the average reduction in working hours during the span of the entire working life caused by endometriosis. Thus, there is a need to estimate the consequences of endometriosis on working life at societal level based on nationwide representative data, and to quantify working years lost. The main aim of this study was to use Danish population-based registers to estimate the association between endometriosis and number of working years lost, defined as the number of years without education or work, and to describe specific types of absenteeism or workforce disruption on working years lost (unemployment, sick leave, disability pension, early voluntary retirement, or premature death) ( Plana-Ripoll et al. , 2023 ). In a supplementary analysis, the aim was to analyse recent Danish questionnaire data on women’s health and well-being to further investigate working life and productivity among women with self-reported endometriosis.

During the period from 1 January 1992 to 31 December 2021, a total of 2 650 554 women were followed for a total of 42.8 million person-years. Of these, 42 741 (1.6%) had been diagnosed with endometriosis since 1977 with a median age at diagnosis (interquartile range (IQR)) of 34.4 years (28.2–41.4). The longest individual follow-up time was 30 years (1992–2021). A total of 652 867 (24.6%) women left the labour market prematurely before the age of 65 years; the number among women with endometriosis was 11 311 (26.5%). Median age (IQR) at premature exit was 60.5 years (50.5–62.3) in the endometriosis population and 60.7 years (52.5–62.3) in the general female population. The proportion leaving the labour market due to disability pension was higher among women with endometriosis compared to the general female population (11.6% vs 9.1%). In addition, the median age at exit for disability pension was a little lower among women with endometriosis compared to the general population (50.2 (43.3–56.4) vs 51.2 (42.5–57.4)) (see Table 1 ). Number of women with endometriosis and in the general population 1992–2021 aged 18–65 years, and causes of premature exit from the labour market. Note: Age in years, follow-up in person-years. Reported medians and interquartile ranges (IQR) are pseudo-estimates. Causes of premature exit from the labour market. On average, women with endometriosis lost an additional 0.26 years (95% CI: 0.17–0.37) of working life compared to the general population of women of same age. As expected, number of working years lost were highest for younger ages and decreased with older age. Additional working years lost among women with endometriosis were mostly due to sick leave and especially disability pension. In contrast, more working years were lost in the general population due to unemployment and voluntary early retirement. Only minor differences were seen for premature mortality. Lastly, the general population had more active years working or enrolled in education. Overall estimates across age and age-specific estimates are presented in Table 2 . Number of total working years lost (WYL), working years lost due to different causes, and active years for women diagnosed with endometriosis and the general population of same age (18–65 years), and the difference between the two groups, 1992–2021. 95% confidence intervals (CI) were used for the estimates related to women with endometriosis and the differences compared to the general population using non-parametric bootstrap with 1000 iterations. For the supplementary analysis, 35 490 women aged 26–51 years were invited to participate and 7298 women responded to the questionnaire (participation rate 20.6%). Due to different activations in the questionnaire (questions depending on previous responses) and non-response, a maximum of 6676 women responded to the questions used in these analyses. Among these, 4.0% reported to have endometriosis. We found that women with endometriosis reported to be less frequently working or enrolled in education (74.1% (95% CI 68.4%–79.2%) with endometriosis, 82.7% (95% CI 81.8%-83.7%) without). All women reported similar numbers of working hours per week (median of 37 h (IQR: 32–38) with endometriosis, median of 37 h (IQR: 33–37) without). Women with endometriosis also reported to have more sick days during the last 4 weeks (4–28 sick days: 16.2% (95% CI 11.6%-21.8%) with endometriosis, 7.9% (95% CI 7.2%-8.7%) without) and to have been less efficient at work or during other daily activities because of physical health during the last 4 weeks (all or most of the time: 18.2% (95% CI 13.8%-23.4%) with endometriosis, 10.5% (95% CI 9.8%-11.3%) without). In addition, more women with endometriosis reported that physical pain interfered with their work (both outside the home and housework) during the last 4 weeks (quite a bit or extremely: 17.5% (95% CI 13.2%-22.6%) with endometriosis, 8.4% (95% CI 7.7%-9.1%) without). When women were asked about their ability to work on a scale from 1 (lowest) to 10 (highest), the median was 8 (IQR 6–9.5) with endometriosis and 9 (IQR 7–10) without (see Table 3 ). Women’s responses across questions related to working life with numbers and percentages. Note: Questionnaire sent to women aged 26–51 years in Denmark in 2023. See full questions from questionnaire in Supplementary Data File S1 . CI = confidence interval. Maternity leave, stay-at-home, other, and don’t know. Question only given to women working, working alongside study, and flexible job arrangements. (pseudo) median (interquartile range (IQR)).

The study design for the main analysis was a population-based cohort study. We included information on all women from minimum age of 18 years to a maximum age of 65 years (born 1935–2002) living in Denmark between 1 January 1992 and 31 December 2021, who had not retired before 1 January 1992. Women were followed from 18 years of age, immigration, or 1 January 1992, whichever occurred later. Follow-up ended at 65 years, emigration, death, disability pension, voluntary early retirement, or 31 December 2021, whichever occurred first. From Statistics Denmark, we had information on date of birth, immigration and emigration, occupation and retirement, and date of death. From the National Patient Registry, we had information on all hospital-based diagnoses of endometriosis regardless of diagnostic method (ICD-8 codes 62530 and 62532–62539 until 1994 and ICD-10 codes DN801–809 from 1994, both A and B diagnoses, where A diagnoses were the main reason for hospital contacts and B diagnoses were present, but not the main reason for contact) from all public and private hospitals (since 1977 and 2002, respectively) ( Schmidt et al. , 2015 ). Approximately 60% of hospital-based endometriosis diagnoses were histologically verified ( Illum et al. , 2022 ; Melgaard et al. , 2023 ). All data were anonymized and not identifiable at individual level. The supplementary analysis was conducted as a cross-sectional study based on questionnaire data from the CYKLUS questionnaire. We included data from CYKLUS from a random sample of women aged 26 to 51 years (born 1972–1996) living in Denmark in 2023. Women were invited to participate in the study and complete a questionnaire through the official digital mail system in Denmark. CYKLUS includes questions on self-reported endometriosis, occupation, working hours, sick days, and different questions related to work ability and productivity. Questions on work ability ( de Zwart et al. , 2002 ; Hansen et al. , 2013 ; Ilmarinen, 2006 ) and daily activities ( Ware, 2005 ; Larsen et al. , 2021 ; Christensen et al. , 2022 ) were based on validated questionnaires (see Supplementary Data File S1 ). For the main analysis, all women diagnosed with endometriosis since 1977 were included and treated as time-varying factors, i.e. women with a diagnosis of endometriosis were considered to be exposed only from the date of diagnosis. We compared women with endometriosis to the general female population (including women with endometriosis and other diseases) to estimate working years lost. This primary outcome had two components: One was related to premature permanent exit from the labour market due to disability pension, early voluntary retirement, or death; and the other was related to delayed entry to or temporary absence from the labour market due to unemployment or sick leave. Estimation of working years lost due to premature permanent exit from the labour market was based on an adaption of the life-years lost method ( Andersen, 2017 ), which has been used to estimate working years lost in individuals with mental disorders ( Plana-Ripoll et al. , 2023 ). For the first component, premature permanent exit from the labour market, we estimated age-specific rates of retirement for all women with endometriosis and for all women in the general population. These retirement rates were constructed based on birth cohorts as no one in the study population could be followed all years from 18 to 65 years of age. We could then estimate cumulative incidence curves starting at every single age (18, 19, 20, etc) and estimate the expected residual working life at each single age as the area under the curve. Next, we used these to estimate average years on the labour market before retirement both for those with endometriosis (considering the number of incident cases of endometriosis at each age) and for the general population of the same age. These averages were weighted based on the number of individuals diagnosed at each age. For both groups, we estimated working years lost as years lost on the labour market compared with a standard retirement age of 65 years. Next, we decomposed working years lost into specific types of workforce disruption (disability pension, voluntary early retirement, or death) using a competing risks model ( Andersen, 2013 ). To estimate temporary absence from the labour market, expected residual working life before retirement was divided into active periods (working or enrolled in education) and temporary inactive periods (unemployed or on sick leave) based on the proportion of individuals in each category at each age. Lastly, total working years lost was estimated as the average number of years women with endometriosis were not working or enrolled in education compared to the general female population of the same age at the time of diagnosis. 95% CIs were used for the estimates related to women with endometriosis and the differences compared to the general population using non-parametric bootstrap with 1000 iterations. R version 4.3.2. was used for the analysis. Additional details of this method are available in previous publications ( Plana-Ripoll et al. , 2023 ). The supplementary analysis was based on questionnaire data and descriptive analyses were conducted. Stata version 18 was used for the analyses.

At the time of writing, this is the largest known register-based study on working years lost during the span of work life among women with endometriosis. We found that women with endometriosis had slightly more working years lost primarily due to disability pension and secondarily to sick leave. Questionnaire data showed that women with endometriosis were less frequently working or enrolled in education and had more sick days. They were also less productive and had lower work ability. These results highlight the need to increase knowledge and improve diagnostics in women with endometriosis to improve treatment and work life participation for these women. This is important at individual level and for the costs related to productivity loss for society.

We found that women with endometriosis generally left the labour market a little earlier than the general female population and more often this was due to disability pension. In addition, women with endometriosis had slightly more working years lost compared to the general population (3.1 months for all ages and 4.8 months for women diagnosed at age 30 years). This was primarily due to disability pension and secondarily due to sick leave. In the general population, more working years were lost due to unemployment and voluntary early retirement compared to women with endometriosis. However, the estimated differences in working years lost were relatively low from a societal perspective. Nonetheless, individual women might lose many years in the labour market even though the average estimates were low. Results from the supplementary analysis of questionnaire data confirmed that working life was negatively affected among women with endometriosis. We found that women with endometriosis were less frequently working or enrolled in education and had more sick days compared to women without endometriosis. More women with endometriosis also responded that they had been less efficient because of physical health and that physical pain interfered with their daily work. Lastly, women with endometriosis reported that their ability to work was lower on a scale from 1 to 10 (median value one point lower). In the main analysis, more working years lost due to unemployment in the general population could be due to competing risks, i.e. when women with endometriosis lose more years due to sick leave, then they cannot be unemployed during that time. It could also be because women in the general population are more flexible and have more courage to quit their job. In contrast, women with endometriosis could be more worried about keeping their jobs and doing a satisfactory job ( Hansen et al. , 2013 ). More working years lost in the general population due to voluntary early retirement could also be due to more time with disability pension among women with endometriosis reducing the time and/or need for voluntary early retirement. In the supplementary analysis, results on shorter periods of sick leave and lost productivity for women with endometriosis are an important contribution to the main analysis as this information was missing in the registers. Thus, it adds important information on the consequences of endometriosis and costs for society. Furthermore, a difference in work ability of one point may be an important finding because studies in patients with chronic low back pain found that improvements of 1.5 points (in mean score) on the single item Work Ability Score (WAS) may be interpreted as a clinically important change ( Boekel et al. , 2022 ). This could suggest that a one-point difference on the WAS might also be a relevant difference in both women with and without endometriosis. To the best of our knowledge, no previous studies have quantified the average reduction in working years during the span of the entire working life. In addition, other studies were generally based on self-reported questionnaire data. However, in line with our findings, previous studies from Denmark, Finland, Australia, and the USA found that endometriosis was associated with poorer work ability, work disturbances or absenteeism, and more sick days or disability days ( Hansen, 2013 ; Soliman et al. , 2017 ; Rossi et al. , 2021 ; Bell et al. , 2023 ) or associated with a higher risk of short- and long-term work losses ( Estes et al. , 2020 ). Other studies quantified working time lost and found that women with endometriosis lost between 7.4 and 10.8 h/week (no reference for comparison) ( Fourquet et al. , 2011 ; Nnoaham et al. , 2011 ). In a Danish context, this would amount to 1 day/week. If this amount of working time is lost during an entire working life, this is much more than we estimated based on register data. It is an important strength in the main analysis that we used administrative registers for both workforce participation and diagnoses of endometriosis covering all of Denmark. It is a limitation that endometriosis is underdiagnosed due to use of only hospital-based diagnoses ( Illum et al. , 2022 ). All diagnoses from both public and private hospitals (since 1977 and 2002, respectively) were included ( Schmidt et al. , 2015 ). We had no information on diagnoses from private practicing gynaecologists or general practitioners. Additionally, we had no information on diagnostic method (surgery, imaging, or symptoms), severity of endometriosis, or prior surgical or medical therapy and therefore did not have the possibility to adjust for this or investigate the potential impact on work years lost. If only the most severe cases of endometriosis are diagnosed, the association could be overestimated in this study. On the other hand, as having a diagnosis of endometriosis has been found to be associated with higher socioeconomic status and possibly better health ( Parazzini et al. , 2017 ), this type of selection could result in an underestimation of the estimated association. Furthermore, misclassification of endometriosis due to a diagnostic delay of several years can also result in an underestimation of the association. It is likely that socioeconomic status could also have impacted on the length of the diagnostic delay. Another limitation is that we did not include time trends in hospital-based diagnoses of endometriosis or employment policies in Denmark. However, incidence of hospital diagnoses in Denmark increased over time ( Illum et al. , 2022 ) and employment policies were implemented and changed over time. From 2013, the disability pension policy was changed to increase employment, possibly decreasing and/or delaying disability pensions ( Borup et al. , 2019 ; The Danish Agency for Labour Market and Recruitment). The underestimation of sick leave is a limitation as the variable used for employment is categorized based on primary type of income each year. Thus, shorter periods of sick leave are not registered. In line with this, when endometriosis might be unknown or unaccepted, it could be more difficult for women with endometriosis to get longer periods of sick leave or receive disability pension as both are based on an external assessment from general practitioner or social worker. However, by adding self-reported information from the CYKLUS questionnaire in the supplementary analysis, we can include information on shorter periods of sick leave (sick days during the last 4 weeks) productivity and ability to work which strengthens the study. In the supplementary analysis based on questionnaire data, more endometriosis diagnoses were included with self-reported data compared to the register data. However, similar challenges with underdiagnosis of endometriosis as mentioned above exist. Furthermore, women responding to the questionnaire could be a selected population. With a questionnaire about women’s health, more women with diseases and health problems affecting women are likely to respond. Lastly, the findings from this study might only be generalizable to a Danish and possibly Nordic context. Denmark has a welfare system with economic security during unemployment and shorter and longer periods with illness or reduced ability to work. In addition, people in flexible jobs due to reduced ability to work ( OECD Better Life Index, 2024 ) are most likely categorized as employed in the registers (which might result in an overestimation of employment) but could be categorized as unemployed in other countries.