2024 Volume 46 Issue 4 Pages 283-292
Women’s health issues, including menstruation, childbirth, and menopause, significantly impact work performance and career progression, and pose physical and mental burdens on women. Despite their importance, research on how these events and conditions affect women’s quality of life and health in the workplace is limited. This study explores the relationship between living environment, health status, work factors, and the health and work performance of working women. It focuses on menstruation-associated symptoms (dysmenorrhea, PMS) and menopausal symptoms and their impact on working QOL and health. A cross-sectional study was conducted on the Internet from June 6 to 12, 2024 among working women in Japan aged 20–60. The sample came from a closed survey of panel monitors from a company commissioned to conduct the survey. Screening included gender, age, hours worked, place of residence, and type of work, with participants being women aged 20–60 and working at least 30 hours per week. Of 31,965 participants, 4,245 were excluded for reasons such as short response time, marital status inconsistencies, childcare inconsistencies, undergoing infertility treatment at age 55 or older, and incorrect responses, resulting in 27,720 participants. The survey consisted of 139 questions and was conducted on a relatively homogeneous sample across all strata. Understanding the impact of menstruation-related symptoms on labor productivity and enhancing knowledge of women’s health can lead to a better healthcare system and improved business performance. This study aimed to generate evidence to support these goals.
A social environment in which women can work securely and energetically is an important public health issue. In Japan, in particular, the working-age population (15–64) has been declining since its peak in 1995 [1], and women’s participation in the labor market is therefore an important factor underpinning Japan’s economic development [2]. Women’s work also contributes to securing the economic situation of families, and reduces the risk of family poverty [3]. Secure family economic conditions have a positive impact on family health and children’s education. Women’s participation in the labor market is important in achieving social diversity and gender equality [3], both of which are socially desirable in terms of both ethics and efficiency.
Marriage and childbearing, which many women experience in their lifetime, can be detrimental to women’s employment. In Japan and some other Asian countries, the age trend in women’s employment rate is known to follow an M-shaped curve [4]. Because of the availability of higher education, many young women enter the workforce after completing education in their 20s. In their 30s and 40s, however, many leave the labor market for the first time due to marriage, childbirth, and child-rearing, and the employment rate drops sharply during this period. From their late 40s to 50s, when they have reached the end of their child-rearing years, the number of women working part-time and in other non-regular jobs increases. These lifelong trends in women’s employment are due to social and cultural expectations and discriminatory practices against women, and to the institutional environment related to employment and childcare [5].
Meanwhile, the number of working women in Japan has been increasing since the Equal Employment Opportunity Law was enacted in 1986. The employment rate of women has gradually increased, reaching 78.6% in 2021 among women aged 15–64 [4, 6]. Moreover, the labor force participation rate of women in their late 20s to 40s, who are engaged in childbirth and childcare, has improved compared to the past, and the number of women working until retirement is increasing.
Women experience particular health-related events throughout their lives, such as menstruation, childbirth, and menopause. The impact of these women-specific health issues on the workforce has attracted research attention. These physical and mental burdens have an impact on work performance [7, 8]. Socioeconomic disadvantages may also be present, such as a decline in promotions and an increase in leaving the workforce due to prejudice [9], discrimination, and unfair treatment in the workplace [10]. In order for women workers to fully realize their full potential, it is necessary to examine these relationships and identify factors that may mitigate their effects.
While female workers are increasingly participating in the labor market throughout their lives, there is a paucity of research on how women’s specific life events and health affect their quality of life and health in the workplace. The present HERLIFE study (Study of Women’s Health, Employment, Resilience, and Lives) was conducted to examine the relationship between the living environment, health status, and work factors of working women and their health and work performance. In particular, we aimed to clarify the relationship between menstruation-associated symptoms such as dysmenorrhea and premenstrual syndrome (PMS), menopausal symptoms, and working QOL and health.
Here we report the protocol of the HERLIFE study in accordance with the requirements of the Checklist for Reporting Results of Internet Electronic Surveys (CHERRIES) [11], a set of guidelines designed to improve the quality and transparency of web-based survey research. CHERRIES ensures standardization of the methodology and reporting of results of Internet surveys to ensure reliability, reproducibility, and validity. We also report on the basic characteristics of the participants.
Study protocol
This was a cross-sectional study of working women in Japan, aged 20–60, who were employed on a regular basis. The survey was conducted on the Internet from June 6 to 12, 2024. The sample was a closed survey of panel monitors managed by a company commissioned to conduct the survey. The company sent an e-mail invitation to participate in the study to 572,861 of the 593,523 subjects who had pre-registered as monitors. Of these 572,861 subjects, 75,782 participated in the screening process.
The screening asked for gender, age, hours worked, place of residence, and type of work (desk/non-desk work) without indicating inclusion criteria. Only women aged 16–60 years and working at least 30 hours per week were considered study participants. Regarding the type of work, participants were asked to choose the one that most closely matched their situation from the three options presented: mainly desk work (clerical or computer work) ; mainly talking to people (e.g., customer service and sales); mainly labor (e.g., construction, physical labor, nursing work).
We accounted for the regional and occupational maldistribution of the sample participating in the survey by controlling for geographic and occupational conditions to adjust for the availability of survey participation. Specifically, the sample was geographically divided into seven regions (Hokkaido, Tohoku, Kanto, Chubu, Kinki, Chugoku/Shikoku, and Kyushu), and in each region, the number of occupations was capped at 1:1 for desk and non-desk workers. In other words, 14 sampling blocks were created with 7 regions and 2 occupations.
The sample size was estimated according to a random survey, although this study is not a random survey. Assuming a sample error of 2.5% in each block, the required sample size was 1,500. Thus, a total of 28,000 would be needed for the 14 blocks. Given that inappropriate responses to this would be approximately 10%, we planned to collect a final sample size of 30,000.
A total sample of 30,000 persons was planned, with a maximum participation limit of 2,140 persons in each block. After the survey began, sampling was monitored, and it was observed that there were four regions where the sample was not collected up to the target number. In response, the upper limit for the blocks was eventually removed and the sample was added regardless of region or occupation. Finally, 31,965 individuals participated in the study.
The study was approved by the Ethics Committee of the University of Occupational and Environmental Health Sciences (ER24-002). Informed consent was obtained from the participants.
Data retrieval
Of the 31,965 participants in the survey, the following respondents were excluded: those who responded for less than 4 minutes (n=101); those who indicated that they were not married but lived with their partner (n=463); those who answered that they used a day-care center even though they did not live together with their preschool children (n=1,120); those who reported using a babysitter even though they did not live with their preschool child (n=511); those aged 55 years or older who answered that they were undergoing infertility treatment (n=33); and those who responded incorrectly to a question that asked them to select the third-largest of a series of five numbers (n=2,734).
Those who met any of these criteria were excluded (n=4,245). Finally, 27,720 persons were included in the analysis.
Measurement
In total, the survey asked 139 questions about socioeconomic status, daily life, work-related matters, and health status. Participants were required to answer all of the questions, so there were no missing values.
Socioeconomic status included questions on marital status, family structure, family living together, educational history, income, and area of residence. Regarding daily life, participants were asked about sleep, exercise, hobbies, commuting, household chores, child care, and time spent with family. Regarding work, questions included employment status, occupation, hours worked, and company size. With regard to health, they were asked about their medical history and any illnesses they were receiving treatment for. They were also asked about menstrual status, symptoms associated with menstruation, and menopause-related symptoms.
In addition, the following standard batteries were employed: the Brief Job Stress Questionnaire (BJSQ), the Work Functioning Impairment Scale (WFun), the Work and Family Conflict Scale (WAFCS), the UCLA Loneliness Scale, the Menopause Rating Scale (MRS), the SAMANTA Questionnaire (Sangrado Menstrual Abundante en Ginecología), and the Kessler Psychological Distress Scale (K6).
The BJSQ is a questionnaire widely used in Japanese companies to investigate job stress and psychological responses; the WFun is a tool developed to measure the degree of work functioning impairment due to presenteeism; the WAFCS is a tool to assess conflicts that arise between work and home roles, which primarily measures both the conflicts that arise in the home role due to the demands of the job and the conflicts that arise in the work role due to the demands of the home; the SAMANTA Questionnaire is a self-report tool to assess menstrual hypermenorrhea; and the K6 is a tool to measure psychological distress.
The CHERRIES checklist is shown in Appendix 1.
Table 1 shows the sampling results for each of the planned blocks. The target was to collect 4,280 people in each area, but this target was not reached for four areas, so additional recruitment was conducted until an overall total of 30,000 people was reached, resulting in the overall participation of 31,965 people.
| Area | Participants who completed their answers | Analysis Subject* | |||||
|---|---|---|---|---|---|---|---|
| desk worker | non-desk worker | Total | desk worker | non-desk worker | Total | ||
| n=15,848 | n=16,117 | n=31,965 | n=13,964 | n=13,756 | n=27,720 | ||
| Hokkaido | 867 | 1,004 | 1,871 | 749 | 859 | 1,608 | |
| Tohoku | 1,057 | 1,450 | 2,507 | 904 | 1,216 | 2,120 | |
| Kanto | 6,078 | 4,418 | 10,496 | 5,390 | 3,771 | 9,161 | |
| Chubu | 2,416 | 2,956 | 5,372 | 2,111 | 2,526 | 4,637 | |
| Kinki | 2,680 | 2,728 | 5,408 | 2,379 | 2,335 | 4,714 | |
| Chugoku/Shikoku | 1,286 | 1,706 | 2,992 | 1,113 | 1,465 | 2,578 | |
| Kyushu | 1,464 | 1,855 | 3,319 | 1,318 | 1,584 | 2,902 | |
*Excluded 4,245 people who were deemed to have answered inappropriately
Table 2 shows the distribution of the analysis subjects by age group. There were no blocks with a lack of participants by either region or job category, and the sample size was sufficient in all blocks. The age distribution was 12.9% in their 20s, 26.6% in their 30s, 31.5% in their 40s, and 29.0% in their 50s. This age distribution was approximately consistent across the areas and job categories.
| age | ||||||
|---|---|---|---|---|---|---|
| 20 | 30 | 40 | 50 | Total | ||
| Total | n | 3,577 | 7,379 | 8,720 | 8,044 | 27,720 |
| row % | 12.9 | 26.6 | 31.5 | 29.0 | 100 | |
| Area | ||||||
| Hokkaido | n | 167 | 391 | 526 | 524 | 1,608 |
| row % | 10.4 | 24.3 | 32.7 | 32.6 | 100 | |
| Tohoku | n | 248 | 564 | 732 | 576 | 2,120 |
| row % | 11.7 | 26.6 | 34.5 | 27.2 | 100 | |
| Kanto | n | 1,325 | 2,373 | 2,704 | 2,759 | 9,161 |
| row % | 14.5 | 25.9 | 29.5 | 30.1 | 100 | |
| Chubu | n | 642 | 1,286 | 1,410 | 1,299 | 4,637 |
| row % | 13.9 | 27.7 | 30.4 | 28.0 | 100 | |
| Kinki | n | 598 | 1,251 | 1,552 | 1,313 | 4,714 |
| row % | 12.7 | 26.5 | 32.9 | 27.9 | 100 | |
| Chugoku/Shikoku | n | 275 | 682 | 826 | 795 | 2,578 |
| row % | 10.7 | 26.5 | 32.0 | 30.8 | 100 | |
| Kyushu | n | 322 | 832 | 970 | 778 | 2,902 |
| row % | 11.1 | 28.7 | 33.4 | 26.8 | 100 | |
| Job type | ||||||
| Desk worker | n | 1,636 | 3,622 | 4,535 | 4,171 | 13,964 |
| row % | 11.7 | 25.9 | 32.5 | 29.9 | 100 | |
| Non-desk worker | n | 1,941 | 3,757 | 4,185 | 3,873 | 13,756 |
| row % | 14.1 | 27.3 | 30.4 | 28.2 | 100 | |
Table 3 shows the basic attributes of the analysis subjects: 41% were married; 74% had a college education or higher; and 44% had an annual income of less than 3 million Japanese yen. Regarding menstrual status, 50% reported regular menstruation. A further 20% reported menopause, and 6% reported no menstruation due to surgery or treatment.
| Characteristic | n | % |
|---|---|---|
|
|
||
|
|
12,277 | 44.3 |
|
|
11,564 | 41.7 |
|
|
3,879 | 14.0 |
| Educational history | ||
|
|
386 | 1.4 |
|
|
6,944 | 25.1 |
|
|
4,593 | 16.6 |
|
|
4,509 | 16.3 |
|
|
11,199 | 40.4 |
|
|
89 | 0.3 |
| Annual income (JPY) | ||
|
|
5,761 | 20.8 |
|
|
6,622 | 23.9 |
|
|
6,589 | 23.8 |
|
|
4,003 | 14.4 |
|
|
2,038 | 7.4 |
|
|
1,120 | 4.0 |
|
|
633 | 2.3 |
|
|
361 | 1.3 |
|
|
141 | 0.5 |
|
|
244 | 0.9 |
|
|
208 | 0.8 |
| Living with pre-school children | 3,252 | 11.7 |
| Living with school-age children | 2,691 | 9.7 |
| Menstrual Status | ||
|
|
13,904 | 50.2 |
|
|
5,547 | 20.0 |
|
|
5,629 | 20.3 |
|
|
883 | 3.2 |
|
|
1,757 | 6.3 |
| Treatment history related to menstruation | ||
|
|
20,547 | 74.1 |
|
|
5,007 | 18.1 |
|
|
2,166 | 7.8 |
| Treatment history related to menopause | ||
|
|
25,167 | 90.8 |
|
|
1,601 | 5.8 |
|
|
952 | 3.4 |
Here we report details of the HERLIFE study in accordance with CHERRIES. The disclosure of protocols in accordance with CHERRIES will improve the transparency of Internet research. This report will serve as the basis for future testing of various hypotheses arising from this project.
The HERLIFE study aims to extensively examine the impact of menstruation-related health problems and menopausal symptoms on the lives and work of female workers. It is therefore desirable that a diverse range of people who experience menstruation and menopausal symptoms be included as study participants. About 50% of the subjects reported regular menstruation, 20% reported irregular menstruation, and 20% reported menopause. In addition, 25% of the subjects were undergoing or had undergone treatment with respect to menstruation, and 9% were undergoing or had undergone treatment with respect to menopausal symptoms. These numbers are considered to be balanced and adequate for the purpose of this study.
This study intended to examine the association between exposure and outcomes, including several pre-defined central hypotheses, e.g., menstruation-related symptoms and work performance. Accordingly, it did not intend to obtain a representative sample of Japan, as the study does not aim to simply make claims about the prevalence of menstrual symptoms, etc. However, we controlled for region and occupation in our sampling since the evaluation of associations must take account of regional and occupational maldistribution. We thereby obtained a relatively homogeneous sample in all the strata, with no strata unrepresentative in region, occupation, or age. This is an advantage for future testing of hypotheses.
The study obtained information on factors that affect work and health; in other words, exposure factors such as menstrual complications, dysmenorrhea, PMS, hypermenorrhea, support from the company, financial situation, loneliness, and family-work conflict. The study also obtained information about outcomes affected by these factors, such as employment status, psychological distress, stress, sleep, work dysfunction, and unfair treatment. These variables can be interchangeable as exposures and outcomes.
Examining the impact of female-specific health events on women’s lives, health, and work is a worthwhile endeavor. Understanding the impact of menstruation-related symptoms on labor productivity can aid in the design and provision of an appropriate health care system and, ultimately, strengthen business performance. A better understanding of women’s health issues in the workplace will provide a better working environment for all. Awareness and provision of necessary support regarding women’s health issues also promotes a gender-equal society. We hope that this study will generate evidence in the future that will contribute to these efforts.
This study was supported and partly funded by research grants from the University of Occupational and Environmental Health, Japan (Research Grant for Promotion of Occupational Health, no grant number); the Japanese Ministry of Health, Labour and Welfare (grant numbers 22FB1001, 23JA1005).
The authors declare no conflicts of interest associated with this manuscript.
Data not available due to ethical restrictions.
Conceptualization: YF
Funding acquisition: YF, MO, TI, SM
Investigation: YF, KH, MO, TI, NL
Supervision: YF, SM
Writing-original draft: YF
Writing-review & editing: YF, KH, MO, TI, NL and SM
| Item category & checklist item | Explanation |
|---|---|
| Design | |
| Target population | The survey was conducted among working women in Japan who were in regular employment and aged 16–60. The sample was a closed survey of panel monitors recruited by a company commissioned to conduct the Internet survey. The survey was not designed to be a representative sample of the Japanese population. |
| Sampling frame |
The company that commissioned the study asked screening questions to subjects who were pre-registered as monitors and met the inclusion criteria. The screening step asked about gender, age, working hours, place of residence, and type of work (desk/non-desk work) without indicating inclusion criteria; only women aged 20–60 years and working at least 30 hours per week were considered study participants. To control for regional and occupational maldistribution, we adjusted the availability of survey participation by geographic and occupational conditions. Specifically, we divided participants geographically into seven regions (Hokkaido/Tohoku, Kanto, Chubu, Kinki, Chugoku/Shikoku, and Kyushu), and in each region, we set an upper limit for the number of persons in each occupation so that the ratio of desk and non-desk workers was 1:1. In other words, 14 sampling blocks were created with 7 regions and 2 occupations. A sample of 30,000 persons was planned, with a maximum participation limit of 2,140 persons in each block. After the survey began, we monitored the number of participants and identified four regions where the sample did not reach the target number, so we eventually lifted the upper limit for the blocks and added additional participants, regardless of region or occupation. Ultimately, 31,965 participants were recruited. |
| Sample | The company commissioned to conduct the survey sent an invitation to participate in the survey to 572,861 of the 593,523 subjects who were pre-registered as monitors by e-mail. Of these, 75,782 people participated in the screening. Ultimately, 31,965 people participated. |
| IRB (Institutional Review Board) approval and informed consent process | |
| IRB approval | This study was approved by the Ethics Committee of the University of Occupational and Environmental Health Sciences (ER24-002). |
| Informed consent | Subjects received a brief guide and completed a screening during the first phase of the survey. If they met the inclusion criteria by responding to the screening, they were presented with a detailed survey description screen. Only those who clicked the Consent to Participate button on the screen and gave the appropriate consent were directed to the survey screen. |
| Data protection | No personal information was obtained in this study. Information on those who did not meet the inclusion criteria for participation in the study and those who did not give their consent to participate was immediately deleted. |
| Development and pretesting | |
| Development and testing |
The survey was developed by a team of approximately 10 individuals, including the authors and their colleagues. The team included public health and epidemiology specialists, occupational physicians, obstetricians, gynecologists, pediatricians, breast surgeons, and psychiatrists, as well as non-medical clerical personnel. The development team proposed several hypotheses and selected the questions needed to test them. Some of the questionnaire items were newly created, while others were derived from an existing battery. The survey items were developed after repeated review and revision for consistency, clarity, wording, and accuracy. The web screens for the survey were developed by a dedicated team from an outsourced firm. The screen was tested repeatedly by a team of ICT experts and the authors’ team to ensure that it worked and was easy to use. |
| Recruitment process and description of the sample having access to the questionnaire | |
| Open survey versus closed survey | This survey was closed to the public. Invitations to participate in the survey were only extended to those who were registered as monitors in advance with the company conducting the survey. Participants could access the survey screen with an ID and password. |
| Contact mode | Participants received a request to participate in the survey via email from the survey firm. |
| Advertising the survey | Respondents were recruited from a panel of pre-registered monitors. |
| Survey administration | |
| Web/E-mail | Participants clicked to answer questions posted on the website. The survey screen consisted of multiple pages, and responses were automatically saved when the subject moved to the next page. |
| Context | The survey website was hosted exclusively for the survey and was not linked to any other website. Subjects accessed it directly from the survey URL indicated on the e-mail through which they were recruited. |
| Mandatory/voluntary | The survey was voluntary. |
| Incentives | Those who completed the survey were given coupons with monetary value as a reward. |
| Time/Date | The survey was conducted between June 6 and 12, 2024 |
| Randomization of items or questionnaires | None of the questionnaire items were randomized. |
| Adaptive questioning | Adaptive questioning was employed depending on the content. |
| Number of Items | There were 33 sections with 139 questions |
| Number of screens (pages) | The survey was presented on 25 pages. The average number of items presented on a page was 5.5. |
| Completeness check | When a participant finished answering a question on one page, he/she was redirected to the next page, but was able to return to the previous page at any time to revise his/her answers until he/she made a final confirmation. If there were any unanswered items, a warning was displayed and the participant was unable to transition to the next page. All the questions in this survey were required items. This was explained in the informed consent process prior to the start of the survey. Participants could withdraw from the survey at any time, even after starting to answer the questions. |
| Review step | Respondents were able to review and change their answers at any point, up until they formally submitted their survey. At the end of the survey, the respondent was prompted to click the submit button in order to submit their completed responses. |
| Response rates | |
| Unique site visitor | Unique site visitors were not counted as all potential respondents were invited to complete the survey using a unique web link. |
| View rate (Ratio of unique survey visitors/unique site visitors) | Unique site visitors were not counted. |
| Participation rate (Ratio of unique visitors who agreed to participate/unique first survey page visitors) | Participation rates could not be calculated. Only those invited to participate in the survey who met the inclusion criteria on screening were progressively included in the study. |
| Completion rate (Ratio of users who finished the survey/users who agreed to participate) | The completion rate was not supplemented. |
| Preventing multiple entries from the same individual | |
| Cookies used | Cookies were not used to identify multiple entries. |
| IP check | Duplicate participants were identified by IP address, and samples with duplicate IP addresses were excluded. |
| Log file analysis | The log file was not used to identify multiple entries. |
| Registration | Participants were registered in advance as monitors. |
| Analysis | |
| Handling of incomplete questionnaires | Only completed questionnaires were analyzed. Since this survey required responses to all items, there were no missing values. |
| Questionnaires submitted with an atypical timestamp | The time from the start to the end of the response was recorded. The median response time was 9 minutes (7–13 minutes for the 25–75%). Subjects with response times of 3 minutes or less were excluded (101 subjects). |
| Statistical correction | Statistical corrections such as weighting and propensity scores were not used, as the primary objective was not surveillance to investigate simple prevalence rates, etc. Rather, the study aimed to examine several a priori hypotheses regarding the association between exposure and outcome. The hypotheses will be tested using multivariate analysis and propensity scores to adjust for confounding factors. |
記事関連データはJ-STAGE Dataで利用できます。（ 産業医科大学雑誌 J-STAGE Data） / The data analysis file and all annotator data files are available in J-STAGE Data. ( J-STAGE Data from Journal of UOEH)
