Social Determinants of Health and Dysmenorrhea: A Systematic Review.

A protocol was developed by the research team who had expertise in systematic reviews, health disparities, dysmenorrhea, and women’s health. The authors registered the review protocol on the international prospective register of systematic reviews PROSPERO (Registration number: CRD42023391715) 18 prior to the full-text screening of the results from the initial literature searches. The conduct and reporting of this review followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines. 19 Inclusion criteria were (1) English language, (2) full-length publications (including journal articles, book chapters, and dissertation theses), (3) examined at least one dysmenorrhea outcome, (4) reported at least one SDoH variable, (5) reported findings from the statistical analysis of the relationship between ≥1 SDoH and ≥1 dysmenorrhea outcome variables, and (6) were observational studies, baseline analysis of interventional studies, or mixed methods research with a quantitative component. Dysmenorrhea outcomes included presence/prevalence, menstrual pain severity, menstrual pain frequency, symptom interference with daily life (e.g., absence from school or work), needs for dysmenorrhea treatments, use of or access to dysmenorrhea treatments, and the risk of other chronic pain conditions (e.g., irritable bowel syndrome, non-cyclic pelvic pain) among people with dysmenorrhea. SDoH factors were identified a priori. The identification of SDoH factors was guided by a report from the WHO Commission on SDoH, 13 Healthy People 2030, 20 and the PROGRESS framework (a framework that specifies factors that contribute to health equity: P lace of residence, R ace/ethnicity/culture/language, O ccupation, G ender/sex, R eligion, E ducation, S ocioeconomic status, S ocial capital). 21 As specified in the protocol, SDoH factors we planned to include were socioeconomic status (including education, parent’s education, and income), race/ethnicity, gender identity, sexual orientation, insurance, healthcare access, social support and social capital, social stress, traumatic experiences, religion, health literacy, cultural factors, discrimination, employment, occupational factors, geographical locations, rurality, and environmental exposures. Excluded articles were editorials, letters, comments, preprints, erratum, news, case reports, review papers without primary data collection, inaccessible full-length articles, studies only focusing on measurement/scale development, qualitative research, and duplicated reports of findings from a given dataset. We conducted searches of the following databases with no publication date limits: Medline (Ovid), EMBASE (Elsevier), CINAHL (EBSCO), PsycINFO (EBSCO), Scopus (Elsevier), and Google Scholar . The first 100 articles retrieved by Google Scholar were downloaded using the Publish or Perish tool. 22 Searches reflected articles published as of February 4, 2024. The search strategies were developed by a health sciences librarian in collaboration with coauthors. The search terms incorporated subject headings and keywords associated with dysmenorrhea and SDoH. Example dysmenorrhea terms, which were informed by previous systematic reviews on dysmenorrhea, 23 included dysmenorrh*, menstrual cramps, menstrual pain, and period pain. The SDoH search terms were informed by the list of SDoH factors identified a priori. In addition, the comprehensive health equity search strategy developed and validated by Prady and colleagues assisted us in identifying SDoH search terms to include in our search strategy. 24 When possible, articles indexed as editorials, preprints, erratum, news, letters, and case reports were excluded using the Boolean operator NOT. The search was limited to English-language articles. Appendix 1 outlines the full search strategies for all information sources. Literature search results were imported into the Covidence systematic review management platform. 25 After deduplication, the literature was screened in two stages. First, two reviewers independently screened titles and abstracts. Then, full texts were pulled, if available. The university interlibrary loan service attempted to locate the missing full-text articles until all options were exhausted. At least two reviewers independently screened all available full texts to assess eligibility. For both stages of screening, two reviewers reviewed and discussed discrepancies and if they could not reach a consensus, then consulted with a third reviewer. Additionally, authors reviewed the reference lists of included studies as well as relevant review papers to identify potential additional studies, a strategy known as spooling. We developed a standard data extraction form on the Covidence platform. 25 The form was tested and revised accordingly before use. The following data were extracted: author, publication year, study design, country and recruitment setting, participant characteristics, sample size, dysmenorrhea outcomes and measures, SDoH factors and measures, methods for statistical analysis, and strength and significance of the relationships of interest. Four reviewers participated in data extraction. At least two reviewers independently extracted data from each article. Disagreements were resolved by discussion and consensus between two reviewers or consultation with a third reviewer. Where missing information was detected, or clarification was needed, authors attempted to contact study authors via email. At least two out of four reviewers independently assessed RoB using the RoB in Non-randomized Studies - of Exposure (ROBINS-E) tool. 26 There is no existing RoB tool designed to specifically assess studies on social determinants of health. We chose to use the ROBINS-E tool, as it is broadly applicable to RoB assessment in non-randomized studies of environmental exposures. As different study designs were included in this review, using a design-specific tool would make it challenging to synthesize study results across studies. For each included study, authors evaluated RoB arising from (1) confounding, (2) measurement of the exposure (i.e., SDoH), (3) selection of participants into the study, (4) post-exposure interventions, (5) missing data, (6) measurement of the dysmenorrhea outcome, and (7) selection of the reported result. RoB for each domain was judged as “low RoB,” “some RoB concerns,” “high RoB,” and “very high RoB” based on ROBINS-E tool instructions. 26 Disagreements were resolved through discussion and consensus with assistance from a third reviewer. After assessing all seven domains, an overall RoB rating (low, some concerns, high, very high overall RoB) was generated for each article based on the ROBINS-E algorithm. Specifically, “low overall RoB” meant the RoB for all domains was low except for concerns about confounding. “Some overall RoB concerns” meant at least one domain was at some concerns, but not all domains were at high or very high RoB. “High overall RoB” meant at least one domain was at high RoB, but no domains were at very high RoB, or several domains were at some concerns. “Very high RoB” meant at least one domain was at very high RoB, or several domains were at high RoB. 26 Key study characteristics (e.g., sample characteristics, study design, SDoH factors examined, dysmenorrhea outcomes) were summarized. Relationships within articles were synthesized according to SDoH factors. Articles reporting relationships with more than one factor were included in more than one section of the results (e.g., race/ethnicity and income). Findings were synthesized narratively, focusing on associations between SDoH factors and dysmenorrhea outcomes. Lower dysmenorrhea presence/prevalence, lower severity, lower frequency, lower interference, lower needs for treatment, lower risk of other chronic pain conditions, and higher treatment use/access were considered positive outcomes. Conversely, higher dysmenorrhea presence, higher severity, higher frequency, higher interference, higher needs for treatment, higher risk of other chronic pain conditions, and lower treatment use/access were considered negative outcomes. We further categorized SDoH factors into risk factors, protective factors, and equivocal relationships based on the dominant evidence pattern in association with dysmenorrhea. If an SDoH factor was associated with positive dysmenorrhea outcomes in >60% of articles reporting that factor and in at least 5 articles, it was considered a protective factor. If an SDoH factor was associated with negative outcomes in >60% articles reporting on that factor and in at least 5 articles was considered a risk factor. If an SDoH factor did not show a dominant pattern, it was classified as having an equivocal relationship with dysmenorrhea outcomes. A quantitative meta-analysis was not conducted due to the pronounced heterogeneity in study samples (recruitment settings, eligibility criteria, and sample characteristics), dysmenorrhea outcome measures, SDoH measures, design, and analytical approaches across studies. This was in alignment with our registered protocol. Consequently, our systematic review aimed to offer a comprehensive account of the current literature regarding SDoH and dysmenorrhea.

Figure 1 provides a flow diagram of the articles identified, screened, included, and excluded. There were 4702 records identified – 4577 through database searching and 125 through citation searching, of which 2108 were excluded as duplicates, resulting in 2594 unique records screened. Three articles reported results from one study; 27 – 29 the one that reported the most comprehensive and non-duplicated results was retained and the other two were excluded. 27 Another pair of articles from the same study reported different SDoH factors, and both were retained. 30 , 31 A total of 166 articles were included in the review (See Supplementary Table 1 ). Table 1 summarizes the characteristics of the included articles. Articles included participants from 47 countries or territories, with half of the studies conducted in Asia. Most articles did not clearly indicate the type of dysmenorrhea examined, included samples of more than 100 participants, did not include participants under the age of 18, did not report the race/ethnicity of participants, recruited via communities rather than clinics, and were cross-sectional. Among SDoH factors, the top three most studied were income, participant education/literacy level, and geographical location. No articles examined relationships between dysmenorrhea outcomes and the following SDoH factors: gender identity, sexual orientation, healthcare access, insurance, or discrimination. For a detailed description of study characteristics, please see Supplementary Table 1 . For dysmenorrhea outcomes, the two most frequently studied were dysmenorrhea presence/prevalence (75.3%) and severity (24.1%). Menstrual pain frequency (3.6%), symptom interference with daily life (6.6%), treatment access/use (5.4%), and needs for treatment (0.6%) were studied much less often. None of the articles examined the risk of chronic pain conditions among people with dysmenorrhea as an outcome. For a study-by-study accounting of dysmenorrhea outcomes, see Supplementary Table 1 . Figure 2 summarizes RoB findings. Most articles had high or very high RoB related to confounding, missing data, measurement of the dysmenorrhea outcome, and/or selective reporting of results. For overall RoB, none of the identified articles were classified as low RoB, one was classified as some concerns, 128 were classified as high RoB, and 37 were classified as very high Rob. As almost all articles (99.4%) had an overall high or very high RoB, RoB was not further summarized within the results pertaining to each SDoH. Across 21 articles, a large majority (n = 19, 90.5%) reported statistically significant relationships between traumatic experiences and worse dysmenorrhea outcomes, including higher dysmenorrhea presence, 31 – 43 severity, 44 – 47 and frequency. 38 – 41 , 48 , 49 Specifically, risk factors for higher dysmenorrhea presence, severity, or frequency were sexual assault/abuse/harassment, 31 , 35 , 36 , 39 – 41 , 44 – 46 , 49 sexual coercion, 40 gender-based harassment, 40 domestic violence by a partner or other family member, 33 maternal abuse, 37 cumulative childhood trauma, including general, emotional, physical, and sexual trauma, 41 war-related concussion, 42 and post-traumatic stress disorders. 38 , 46 , 47 Analyses that combined or separated types of abuse showed important differences. In 5 articles that combined different types of abuse (sexual, physical, or emotional) in analyses, 4 identified combined abuse as a risk factor. 33 , 34 , 43 , 50 In contrast, physical violence alone (domestic or non-domestic) was not associated with dysmenorrhea presence or frequency. 39 , 49 In the 2 articles that separated a history of childhood and adult sexual abuse, 31 , 48 one found childhood but not adult abuse was a risk factor for dysmenorrhea presence, 31 and the other showed childhood abuse was a more salient risk factor than adult abuse, though both were significant predictors for higher dysmenorrhea presence. 48 Across 13 articles, a majority (n = 8, 61.5%) showed that exposures to environmental toxicants were associated with worse dysmenorrhea outcomes, including higher presence, 51 – 57 severity, 51 and frequency. 51 Risk factors for worse dysmenorrhea outcomes included exposures to (1) passive smoking (linked to higher presence), 27 , 52 , 54 , 57 (2) toluene vapor (linked to higher dysmenorrhea presence and severity) 51 , 55 or handling military fuel (high toluene levels, linked to high dysmenorrhea presence), 57 (3) mercury vapor (linked to higher dysmenorrhea presence), 56 and (4) air pollutants (NO, NO 2 , NO x , CO, and PM2.5, linked to high presence). 53 However, 5 studies did not find statistically significant relationships between dysmenorrhea outcomes and environmental exposures, including passive smoking, 58 , 59 styrene in plastics through contact or ambient air, 60 phthalate metabolites (in urine and follicular fluid), 61 or cooking oil contaminated with polychlorinated biphenyls and dibenzofurans. 62 Across 8 articles, 5 (62.5%) reported statically significant relationships between female genital mutilation and worse dysmenorrhea outcomes, including higher dysmenorrhea presence 63 – 66 and severity. 67 The remaining articles (n = 3, 37.5%) showed no association. 68 – 70 Out of 7 articles on job-related stress, 6 (85.7%) suggested high job-related stress was associated with worse dysmenorrhea outcomes (See Table 2 ). Specifically, low job control was consistently linked to worse dysmenorrhea outcomes, including higher severity 71 and higher interference. 72 Low job security was consistently linked to worse dysmenorrhea outcomes, including higher frequency 73 and higher interference. 72 Except for one study where no significant relationship was reported, 57 high job strain was linked to worse dysmenorrhea outcomes, including higher presence, 74 – 76 higher interference, 76 and higher frequency. 73 In the 11 articles exploring the relationship between social support and dysmenorrhea outcomes, 7 (63.6%) reported that better social support was associated with better dysmenorrhea outcomes. 77 – 83 Specifically, better social support was linked to lower dysmenorrhea presence 79 , 81 , 83 and severity, 78 and disruptions in or absence of social support were related to higher dysmenorrhea presence 77 , 80 or severity. 82 The remaining articles (n = 4, 36.4%) showed no relationship. 45 , 84 – 86 Among the 8 articles that examined the relationship between prior menstrual education and dysmenorrhea outcomes, 87 – 92 5 (62.5%) linked prior menstrual education or menstrual literacy to better dysmenorrhea outcomes, including lower presence 88 , 90 , 93 and severity. 89 , 92 The remaining 3 articles either showed an opposite relationship 94 or did not find statistically significant associations. 87 , 91 About half of the articles (n = 18, 51.4%) showed significant relationships. 30 , 44 , 80 , 83 , 93 , 95 – 107 However, the direction of the relationships was inconsistent. Across 14 articles, lower income was associated with worse dysmenorrhea outcomes, including higher presence, 30 , 44 , 80 , 93 , 96 , 97 , 102 , 104 , 105 higher severity, 95 , 108 greater needs for treatment, 99 and lower likelihood to take pain medications 98 and oral contraceptive pills 107 for dysmenorrhea. In contrast, across 4 articles, higher income was associated with higher dysmenorrhea prevalence. 83 , 103 , 106 , 109 The rest of the studies (n = 17, 48.6%) showed no relationships between income and dysmenorrhea outcomes. Over half of the articles (n = 20, 58.8%) showed no relationships, and the rest (n = 14, 41.2%) showed significant relationships. Of the 14 showing significant relationships, the direction of the relationships was inconsistent. Across 6 articles, lower education/literacy was associated with worse dysmenorrhea outcomes, including higher presence, 39 , 110 , 111 higher severity, 108 higher interference, 72 and lower oral contraceptive use for dysmenorrhea. 107 In contrast, across 7 articles, lower education/literacy was associated with better dysmenorrhea outcomes, including lower presence 87 , 96 , 112 , 113 and severity. 114 – 116 In the remaining article, education was significantly correlated with analgesics use for dysmenorrhea, but the direction of the relationship was unclear. 117 Over half showed no relationships (n = 10, 52.6%). 70 , 76 , 94 , 97 , 104 , 118 – 122 The rest (n = 9, 47.4%) showed significant relationships, but the direction of the relationships was inconsistent. In 6 articles, lower parent education was associated with worse dysmenorrhea outcomes, including higher dysmenorrhea presence 83 , 123 – 125 and severity. 44 , 124 , 126 However, in 2 other articles, lower parent education was associated with lower dysmenorrhea presence. 91 , 127 In the remaining article, the direction of the relationship was unclear. 128 In 16 articles, miscellaneous socioeconomic status (SES) factors were examined using undefined/unspecified measures (n = 10), 79 , 81 , 86 , 114 , 115 , 129 – 133 a one-factor proxy measure (n = 3, e.g., housing ownership, 39 managing to make ends meet, 39 access to tap water/clean toilet facility/ hygienic menstrual products, 110 possession of costly goods 134 ), or by combining multiple factors to derive a composite score (n = 3, e.g., mother’s education and employment situation, 90 parent education, house type, and possession of costly goods, 135 place of residence, parent education, and number of siblings 127 ). Across these 16 articles, the relationship between SES and dysmenorrhea outcomes was equivocal. Slightly more than half (n = 9, 56%) did not show significant relationships. 79 , 81 , 86 , 90 , 130 , 131 , 133 – 135 In 6 articles, lower SES was associated with worse dysmenorrhea outcomes including higher presence, 39 , 110 , 127 , 129 , 132 and severity. 115 in univariable or multivariable 110 analyses. In the remaining article, higher SES was associated with higher dysmenorrhea severity. 114 Nine articles (30.0%) showed significant relationships. 87 , 105 , 110 , 120 , 129 , 136 – 140 However, the direction of the relationships was inconsistent. In 7 articles, rural living was associated with worse dysmenorrhea outcomes including higher dysmenorrhea presence 87 , 105 , 110 , 129 , 136 , 139 and severity. 138 In 2 articles, urban girls had higher dysmenorrhea presence 136 , 141 and severity; 141 however, in these two articles, rural girls were more likely to miss school due to dysmenorrhea 136 and less likely to take medications than urban girls. 141 The remaining 21 articles comparing rural and urban locations did not find any statistically significant associations with dysmenorrhea outcomes. 70 , 118 , 120 , 121 , 124 , 126 , 127 , 142 – 155 In one study, high altitude was associated with higher dysmenorrhea presence. 137 Table 2 summarizes results on other occupational factors (e.g., occupational exposures, types of occupations, work schedule, and physical aspects of the job). Across 19 articles, relationships between occupational factors and dysmenorrhea outcomes were equivocal. In terms of occupational exposures, evidence was scarce, null, or equivocal. A small number of studies linked occupational chemical exposures (including hair products, 156 chemicals used in the semiconductor industry, 157 disinfectants, 75 or anti-cancer agents 75 ) to higher dysmenorrhea presence or severity. Neither exposure to vibration 27 nor exposure to noise 27 , 75 , 157 was associated with dysmenorrhea presence. mong 3 studies on occupational exposure to low temperatures, 2 reported associations with higher dysmenorrhea presence. 158 , 159 Regarding job types/settings, the evidence was inconclusive (n=12). In 8 articles, job types/settings were associated with dysmenorrhea outcomes, including presence 39 , 75 , 158 , 160 – 164 and severity. 165 However, the relationships were not significant in the rest of the 4 studies. 71 , 73 , 157 , 166 While most articles (n=8, 66.7%) show significant relationship, the job types/settings studied were too heterogeneous to make a meaningful conclusion. In terms of work schedule, both long work hours and shift work were examined in relationship to dysmenorrhea outcomes. Except in one study where long work hours were associated with higher dysmenorrhea presence, 158 associations were not found in the other 5 studies. 73 , 75 , 112 , 167 , 168 Similarly, except in one study where shift work was associated with higher dysmenorrhea presence, 75 shift work was not associated with dysmenorrhea outcomes in the other 4 studies. 27 , 58 , 73 , 169 For physical aspects of the job, both sedentary work and physical demands of the job were examined. The evidence was limited or equivocal. Sedentary work was not linked to dysmenorrhea presence in the one study on this relationship. 170 High physical demand was associated with higher dysmenorrhea presence in one study, 75 but not with dysmenorrhea presence in other studies. 27 , 158 , 161 Across 21 articles, the direction of associations between race/ethnicity and dysmenorrhea outcomes was equivocal. Nine articles (42.9%) showed no relationship between racial/ethnic groups and dysmenorrhea outcomes. 30 , 57 , 102 , 107 , 154 , 171 – 174 Seven articles (33.3%) suggested minoritized race/ethnicity was significantly associated with worse dysmenorrhea outcomes, including higher presence, 39 , 104 , 123 , 175 higher severity, 123 , 176 , 177 and higher interference. 91 These articles suggested worse dysmenorrhea outcomes in racially/ethnically minoritized groups including African/Black Americans, 91 , 175 , 178 , 179 Hispanic/Latino Americans, 179 Ethnic Minority Chinese in China, 104 Migrant Indians in India, 39 and girls of Asian/African ethnic origin living in Israel. 123 Three articles (14.3%) showed that minoritized race/ethnicity was associated with better dysmenorrhea outcomes including lower presence, 100 severity, 108 , 180 and interference. 100 These articles showed lower dysmenorrhea severity in Black American women compared with White, 108 lower dysmenorrhea severity in Afro-Caribbean individuals compared with Caucasian individuals in England, 181 and lower dysmenorrhea presence and interference among non-Malay adolescent girls than Malay adolescent girls in Malaysia. 100 The remaining 2 articles (22.2%) showed minoritized race/ethnicity as both risk and protective factor for dysmenorrhea depending on the racial group and outcome variable. 182 , 183 A study conducted in Canada showed that South Asian Canadians had a higher prevalence of dysmenorrhea, while East Asian Canadians had lower dysmenorrhea prevalence, both in comparison to Caucasians. 183 A US-based study on female soldiers suggested African American soldiers had better dysmenorrhea outcomes than Caucasian American soldiers. 182 In the same study, Asian American soldiers with severe dysmenorrhea sought healthcare and treatment at a significantly lower rate than American soldiers in other racial/ethnic groups with severe dysmenorrhea. 182 In 12 articles, the relationship between employment status and dysmenorrhea outcomes was equivocal; 7 showed no relationship, 60 , 93 , 108 , 120 , 151 , 174 , 184 and 5 showed statistically significant but inconsistent findings. Among these 5 articles, one suggested unemployment was associated with higher dysmenorrhea severity, 185 3 showed employment was associated with higher dysmenorrhea presence, 149 , 159 , 186 and one reported higher dysmenorrhea presence among both employed and unemployed women compared to homemakers. 187 Seven (63.6%) did not show any significant relationships. 87 , 92 , 94 , 103 , 123 , 146 , 188 Among 4 articles (36.4%) reporting significant results, 3 suggested Muslim girls in India had higher dysmenorrhea presence 39 , 110 and severity 134 compared to Indian girls with other religious affiliations, and the remaining article reported a higher presence of dysmenorrhea among religious Vietnamese women than their counterparts who practiced no religion. 113 Across 8 articles that examined students’ living situations (e.g., living in dorms versus at home with family), 5 showed no relationship, 94 , 97 , 119 , 122 , 124 and 3 showed statistically significant relationships with dysmenorrhea outcomes. However, the direction of the association was inconsistent. Two articles suggested worse dysmenorrhea outcomes (i.e., higher dysmenorrhea presence) among those living in dorms or living away from parents, 189 , 190 while the other article suggested better outcomes (i.e., lower dysmenorrhea severity) among those living in dorms. 126 There was limited evidence for a number of factors. Four articles examined family culture and dysmenorrhea outcomes, 39 , 83 , 96 , 191 but the results were equivocal. Family conflicts were positively associated with dysmenorrhea severity in one study, 191 but was not associated with dysmenorrhea presence in another study. 96 Low family autonomy in decision-making was associated with high dysmenorrhea presence. 39 Family psychosocial coherence was not associated with dysmenorrhea presence. 83 Two articles examined birthplace, though neither study analyzed the data by race/ethnicity. 192 , 193 The articles had conflicting findings. An older Australia-based study suggested dysmenorrhea presence differed based on birth country, 187 while a Canada-based study suggested dysmenorrhea presence did not differ based on being foreign-born or not. 192 Two studies examined relationships between country of residence and dysmenorrhea outcomes. 113 , 194 In the two studies, higher dysmenorrhea prevalence was seen in Vietnamese immigrants in Korea versus native Vietnamese, 113 and higher dysmenorrhea severity was seen in people living in Australia versus China. 194 One study examined registered city residency status. 73 A China-based study showed that registered residents of a large city were less likely to have dysmenorrhea than non-registered residents working in the city. 73 One study examined the relationship between menarche initiation rites and dysmenorrhea outcomes and reported that women completing the rites (e.g., followed a prescribed diet, spent time in isolation, cared for during the ceremony, and emesis) were less likely to report dysmenorrhea. 195 One study examining stress related to life events found that it was significantly associated with dysmenorrhea presence. 57 Lastly, one study examined the relationship between e-health literacy and dysmenorrhea severity and reported a significant negative relationship (i.e., high e-health literacy, lower severity). 196

This systematic review synthesizes literature on relationships between a comprehensive list of SDoH factors and dysmenorrhea outcomes. Despite limitations in the existing evidence, we identified SDoH variables as risk and protective factors for dysmenorrhea outcomes. SDoH factors associated with worse dysmenorrhea outcomes included traumatic experiences, harmful environmental exposures (especially passive smoking), female genital mutilation, and job-related stress. SDoH factors associated with better dysmenorrhea outcomes were prior menstrual education and social support. The mechanisms behind these associations are largely unknown but can be inferred. Chronic stress from trauma, jobs, or environmental exposures may induce epigenetic, hormonal, and neuroimmune changes (e.g., central sensitization, allostatic load, HPA activation), which increase vulnerability to pain. 197 , 198 Genital mutilation may result in injury, infection, and psychosocial consequences, 199 increasing the risk for chronic pain. Menstrual health education may provide people with knowledge, skills, and resources to manage dysmenorrhea. 200 Social support has been shown to decrease the adverse influence of pain-related stress and facilitate coping with pain. 201 Some of the identified risk factors are consistent with existing literature. First, a systematic review focusing on childhood adversity linked sexual abuse and post-traumatic stress disorders with dysmenorrhea. 202 A meta-analysis pooling 4 studies showed that sexual assault increased the risk for dysmenorrhea with a pooled odds ratio of 1.6. 203 Additionally, lifetime traumatic events and chronic stress have been shown to increase the risk for chronic pain. 204 , 205 Second, environmental exposures have also been linked to the risks of uterine fibroids, 206 a condition that can cause menstrual pain. Third, among our included articles on female genital mutilation, a majority showed its associations with higher dysmenorrhea presence or severity, providing updated evidence on previously inconsistent associations between female genital mutilation and dysmenorrhea. 202 , 207 Compared to previous reviews, we comprehensively examined diverse SDoH factors, included broader and more updated articles, and incorporated a rigorous RoB assessment. Our systematic review highlighted equivocal findings regarding the associations of SES and race/ethnicity with dysmenorrhea outcomes. The inconsistent findings may be due to several factors. First, there were significant heterogeneities across studies in terms of study populations, settings, designs, and SDoH and dysmenorrhea outcome measures. Second, some smaller studies might have been under-powered to detect significant associations. Third, inconsistent findings might also indicate that the influence of race/ethnicity and SES on dysmenorrhea is context-specific, intertwined, or intersected with other SDoH. On a global level, race and ethnicity are not uniformly defined or applied, and depending on the comparison group, dysmenorrhea outcomes could be better or worse. It will be important to focus on underlying mechanisms (e.g., discrimination, racial bias, access to care) that contribute to dysmenorrhea disparities when conceptualizing future research questions and hypotheses. One overarching finding of this review was the lack of comprehensive, rigorous investigations of SDoH in relation to dysmenorrhea outcomes. The quality of evidence was limited for all SDoH factors as evident in the high and very high RoB for most included studies. There is a complete lack of evidence for SDoH factors including healthcare access, insurance, gender identity, sexual orientation, and discrimination. Additionally, how SDoH factors may be related to long-term dysmenorrhea outcomes, such as risk of other chronic pain conditions, remains unknown. Although the biopsychosocial model 208 is the long-standing accepted model for studying pain, 209 , 210 relatively little is known about how social and environmental factors affect dysmenorrhea. High and very high RoB stemmed from several significant methodological issues in the reviewed articles. First, there was RoB arising from issues in measuring dysmenorrhea outcomes. Over 80% of articles had a high or very high RoB related to outcome measurement. Researchers often used poorly developed and/or unvalidated tools, threatening the validity of their findings. Moreover, important outcomes such as symptom severity and interference were infrequently assessed. Second, there was high RoB for over 80% of articles from confounders not being considered in study design and analyses. Dysmenorrhea risk factors such as age, family history, and parity, along with confounders for specific SDoH factors should be considered in study design and analyses. Third, there was RoB because studies were often poorly reported. Descriptions were insufficient on sampling approaches, sample characteristics (e.g., race/ethnicity), handling of missing data, SDoH and dysmenorrhea measurement, and analytic approaches. The case definition of dysmenorrhea and the types of dysmenorrhea studies were often unclear, making it challenging to compare studies and generalize findings. While it is often infeasible to confirmatively differentiate primary and secondary dysmenorrhea, researchers should thoughtfully determine eligibility criteria, be transparent about screening approaches, collect and report information on gynecological conditions, and follow established guidelines to ensure studies are clearly reported. This review has important implications for future research. First, more attention to social and environmental factors is needed in dysmenorrhea research. The lack of systematic examination of SDoH in dysmenorrhea creates barriers to developing public health strategies to improve dysmenorrhea outcomes and reduce health inequities. Future research needs to consider a broad range of SDoH since people with dysmenorrhea are heterogeneous with diverse lived experiences, backgrounds, and socioenvironmental contexts. 211 Second, to facilitate large-scale research on SDoH and dysmenorrhea outcomes, dysmenorrhea outcomes need to be included in large epidemiologic studies, which already collect a wealth of SDoH information. Despite its high prevalence and impact, dysmenorrhea outcomes have been historically excluded in large population-based or clinical studies. 212 Third, in examining the relationship between SDoH and dysmenorrhea, researchers need to rigorously design studies that can investigate the effects of SDoH on dysmenorrhea, account for confounders in study design and analysis, enhance the size and representation of study samples, use validated dysmenorrhea outcome and SDoH measures, apply sound approaches for missing data, and allow examination of longitudinal relationships between SDoH and longer-term dysmenorrhea outcomes (e.g., risk for future chronic pain). Fourth, studies on SDoH and dysmenorrhea should include a wide developmental range. Less than one-quarter of included studies enrolled adolescents younger than 18 years, the age group that undergoes significant biopsychosocial changes. A better understanding of the impact of SDoH on dysmenorrhea across the reproductive lifespan is needed to develop developmentally tailored interventions. 213 Fifth, future research needs to examine potential social and environmental barriers and facilitators to effective dysmenorrhea management. Our review also has important implications for clinical practice. Clinicians need to be aware of SDoH risk and protective factors related to dysmenorrhea outcomes. Multidisciplinary approaches are needed for holistic treatment. SDoH are often considered not addressable in clinical practice, and clinicians often feel helpless and frustrated when caring for patients faced with intertwined health and social challenges. 214 However, some SDoH factors are clinically addressable (e.g., trauma, menstrual education, and occupational stress), and recommendations are available for clinicians to influence SDoH factors at the patient, practice, and community levels. 214 For example, practical strategies include asking about social challenges sensitively, collecting history of trauma exposures, adopting trauma-informed care, evaluating potential environmental exposures and offering harm reduction resources, providing menstrual education (particularly for pediatric populations), and addressing barriers to care. These strategies may help acknowledge patients’ psychosocial concerns, promote a holistic person-centered approach, and improve equity in dysmenorrhea care. This systematic review has some limitations. First, we only included English-language publications. Second, some studies containing relevant data might have been missed due to a lack of standard search terms on many SDoH factors and variations in how researchers referred to SDoH factors. Third, in this review, we treated higher treatment use/access as a positive outcome. However, we acknowledge that non-socioeconomic factors such as symptom severity and treatment preferences can affect treatment access and use. 215 Fourth, there is a lack of standardized tools to evaluate studies on SDoH. While the ROBINS-E tool is broadly applicable to ROB assessment in non-randomized studies-of environmental exposure, its item concerning post-exposure intervention was not as relevant to some SDoH variables such as race/ethnicity. Fifth, because of the extensive scope of the review, the overall high RoB of the included studies, and the significant heterogeneities across studies in terms of study populations, SDoH measurement, dysmenorrhea outcome measurement, study design, and statistical tests, a meta-analysis was not conducted. Sixth, the relationship between SDoH and dysmenorrhea outcomes may be context dependent. Future research should assess how contextual factors affect the relationship between SDoH and dysmenorrhea outcomes. This systematic review has several strengths. First, it offers a comprehensive picture of how various SDoH and dysmenorrhea outcomes are related. We included a wide range of SDoH factors and dysmenorrhea outcomes. Second, the review provides a global landscape of the social determinants of dysmenorrhea. Over 160 articles from over 40 countries were included. Third, we followed rigorous review methods including pre-registration of the protocol, multiple search engines, comprehensive search terms, and independent parallel screening/full-text review/data extraction. This systematic review identified SDoH factors likely associated with dysmenorrhea outcomes, thus helping to inform the assessment, prevention, and reduction of disparities related to dysmenorrhea. Studies had a high RoB due to methodological issues. This review suggests a lack of systematic examination of SDoH in dysmenorrhea and calls for more rigorous research to understand the impact of SDoH on dysmenorrhea and thereby inform clinical practice and health policy.

Dysmenorrhea is the most common gynecological pain condition in the world. 1 Clinically, there are two types of dysmenorrhea: primary and secondary. 2 Primary dysmenorrhea occurs without demonstrable anatomical or pathological abnormalities. 2 It is highly common among females of reproductive age with a reported prevalence of 45% to 95%. 1 , 3 Key elements of the pathogenesis of primary dysmenorrhea have been described. 2 Effective treatments based on modification of etiologic pathways largely support the validity of the tenets of its pathogenesis. 2 Dysmenorrhea can also be secondary to an underlying pathology such as endometriosis, uterine fibroids, and pelvic inflammatory disease (PID). 2 The estimated prevalence of endometriosis is about 11% among reproductive-age females; 4 the reported prevalence of uterine fibroids was 4.5%-68.6% among females; 5 and the prevalence of self-reported lifetime PID among sexually experienced females in the United States is 4.4%. 6 Not all people with these conditions experience dysmenorrhea symptoms, and the exact prevalence of dysmenorrhea secondary to these conditions has not been reported in population-based studies. The distinction of primary and secondary dysmenorrhea may not always occur, as the confirmatory diagnosis of secondary dysmenorrhea often requires diagnostic technology or surgical procedures that are unavailable in many settings. 7 Whatever its etiology, dysmenorrhea is a population health problem that contributes to educational, occupational, and health disparities of affected people. Over 20% of female students across the world reported absence from grade school or university due to dysmenorrhea. 3 Dysmenorrhea negatively interferes with physical activities, sleep, and quality of life. 1 It also significantly increases females’ risk of developing both noncyclic pelvic pain and non-pelvic pain conditions. 8 , 9 While it is clear that biological and psychological factors contribute to individual differences in dysmenorrhea outcomes, 2 , 10 – 12 the ubiquity of dysmenorrhea among people who menstruate belies the substantial variation in prevalence and severity despite reasonably well-established models of pathophysiology. 2 This suggests that dysmenorrhea may be differentially experienced across populations through the influence by factors that interact with underlying biological processes. These factors, known as “social determinants of health (SDoH),” include social, cultural, and environmental exposures during the life course that affect health, functioning, and quality of life outcomes. 13 SDoH are now well-established drivers of health disparities, defined as differences in health outcomes linked with social, economic, or environmental conditions. 14 , 15 As examples suggesting the relevance of SDoH to dysmenorrhea, racial and socioeconomic disparities in chronic pain 16 and gynecological disorders (e.g., uterine fibroids) 17 have been reported. Given the large proportion of people with dysmenorrhea around the world, understanding SDoH in relation to dysmenorrhea could inform targets for intervention through research, policies, and clinical practice to lower the negative health impacts associated with dysmenorrhea. To our knowledge, no reviews have systematically synthesized data related to risk and protective social determinants of dysmenorrhea. Therefore, the purpose of this systematic review was to synthesize the literature on SDoH and dysmenorrhea. Our goal was to create a thorough summary of SDoH associated with dysmenorrhea and describe the valence associated with dysmenorrhea (i.e., increased or decreased risk). Synthesis of existing research can guide health policy and public health efforts and direct clinical attention to issues that affect dysmenorrhea outcomes. Critical assessment of research evidence can further direct rigorous new research on issues not yet understood.