Measurement of changes to the menstrual cycle: A transdisciplinary systematic review evaluating measure quality and utility for clinical trials.

Menstruation and the wider menstrual cycle play a notable role in the health, human rights, and sociocultural and economic wellbeing of people who menstruate [ 1 ]. In addition, although its significance should not be utilitarianly reduced to only reproductive function, continuity of the human species would not occur without the menstrual cycle. Despite its importance, the study of menstruation and the menstrual cycle continues to suffer from a historical lack of funding and research across disciplines, including within the biological, clinical, public health, and social sciences. Within biomedical research, for example, a publication reporting on a recent technical meeting on menstruation convened by the United States National Institutes of Health (NIH) decried a “lack of understanding of basic uterine and menstrual physiology” among researchers [ 2 ]. Indeed, many foundational, field-defining works have only recently emerged in the past five to ten years following increased attention to menstrual health, which the Global Menstrual Collective defined in 2021 as “a state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity, in relation to the menstrual cycle” [ 3 ]. The contemporary growth of the menstrual health field is—at least partly—due to grassroots menstrual activism, which resulted in 2015 being labeled as “the year of the period” in the lay press [ 4 ]. Other examples of recent fundamental work within menstrual health across disciplines include recommendations for the menstrual cycle to be considered a vital sign and the advent of the field of critical menstruation studies [ 5 , 6 ]. Despite these recent efforts, insufficient research on menstrual health persists. In addition, the study of menstrual health remains fractured across many fields and disciplines, many of which are siloed despite adjacent or even overlapping subject matters (e.g., menstrual health and hygiene within wider sexual and reproductive health; or gynecology, endocrinology, and many other specialties within medicine) [ 7 , 8 ]. As a result, we still lack a complete, systemic, and holistic understanding of menstruation and the wider menstrual cycle. The type of interdisciplinary, comprehensive global efforts needed to address such large gaps in menstrual health research can greatly benefit from standardization—of terminology, of measurement, of analysis, and of outcomes or indicators. The widest global effort at standardization to date has taken place within medicine; the International Federation of Gynecology and Obstetrics (FIGO) established clinical standards of normal and abnormal uterine bleeding occurring outside of pregnancy via a consensus-building process over a series of years [ 9 – 12 ]. These FIGO standards dictate four parameters for menstrual bleeding: the frequency, duration, volume, and regularity of bleeding. FIGO defines normal uterine bleeding as bleeding occurring every 24–38 days (frequency), bleeding lasting no more than 8 days (duration), bleeding of a ‘normal’ amount as defined by the patient that does not interfere “with physical, social, emotional, and/or material quality of life” (volume), and bleeding within a menstrual cycle that only varies in length by plus or minus 4 days (regularity). FIGO further defines bleeding outside these normal parameters as abnormal uterine bleeding, which is divided into standard categories based on whether it is acute or chronic and the source or etiology of the abnormality according to the acronym PALM-COEIN (i.e., P olyp, A denomyosis, L eiomyoma, M alignancy and hyperplasia, C oagulopathy, O vulatory dysfunction, E ndometrial disorders, I atrogenic, and N ot otherwise classified). Other examples of efforts at standardization include menstrual hygiene indicators within the Water, Sanitation and Hygiene (WASH) field and defining how contraception can impact the menstrual cycle and analyzing these data in contraceptive studies [ 13 – 18 ]. Related to terminology, this review uses the phrase, “people who menstruate”, which we define as those who can menstruate, do menstruate, or have menstruated. Although people who menstruate may or may not identify as women or girls, and not all women and girls menstruate [ 19 ], we do use the terms ‘women’ and ‘girls’ in some instances, especially when citing primary literature and because menstrual health cannot “be adequately addressed without attention to the gender norms and dynamics experienced by individuals in the cultures and communities in which they live” [ 7 ]. As much as possible, however, we use gender inclusive terms and other people-first language. To aid in efforts for standardized measurement across the study of menstruation and the menstrual cycle, we systematically reviewed approaches to measure four aspects of changes to the menstrual cycle: bleeding, blood, pain, and perceptions of bleeding, blood, or pain. We use the term ‘ menstrual changes ’ to refer to these four aspects for the remainder of the paper. We sought to include all types of measures or methods for assessing menstrual changes (e.g., quantitative assays, biomarkers, data reported by clinicians, researchers, or directly by the person who menstruates). We use the term ‘ instruments ’ to refer to any of these measures or methods for the remainder of the paper. Our aim was to identify any instruments that have been developed and validated within any field of study to measure menstrual changes, examine how these instruments measured menstrual changes, and assess the measure quality of the identified instruments and their utility for the clinical trial context. Related reviews have been conducted: (a) within fields such as menstrual hygiene or the study of heavy menstrual bleeding (HMB) [ 20 , 21 ]; (b) to measure single parameters like volume of menstrual blood loss [ 22 ]; and (c) for specific approaches like pictorial methods to diagnose HMB [ 23 ]. However, given the gaps and silos within menstrual health research, our aim was to conduct an expansive and transdisciplinary review to inform more standardized measurement across menstrual health research and clinical trials. For this reason, we sought to include menstrual changes caused by any etiology or source. There are many factors that can result in menstrual changes, including those endogenous and exogenous to the person who menstruates. Examples of these etiologies or sources include menstrual or gynecologic disorders (e.g., endometriosis, uterine fibroids, or adenomyosis), use of hormonal or intrauterine contraceptives, use of other drugs or devices to treat or prevent disease, environmental exposures, infectious disease, injury, coagulation disorders, and diet and exercise. We are not aware of any previous efforts to look at menstrual changes across disciplines in this way. As mentioned, one area for which we intend our review to be quite relevant is for data collection in clinical trials, although our broad approach does not preclude the use of our results to inform the measurement of menstrual changes across other research contexts. The importance of data on menstrual changes in the clinical trial context was recently highlighted during the introduction of COVID-19 vaccinations. Because vaccine trials did not collect data on the impact to the menstrual cycle or menopausal uterine bleeding, there were concerns among vaccinated people who menstruate when they experienced these changes, which can erode trust in clinical research and public health interventions [ 24 – 28 ]. As authors working across various sexual and reproductive health spaces, our interest in conducting this review stemmed from a shared goal to improve and standardize the measurement of menstrual changes in contraceptive clinical trials; however, our broad methodological approach permits the utility of our findings across all clinical trials. Clinical trials, and the preclinical research that precedes them, collect data on key organ functioning and vital signs as part of standard toxicology and pharmacodynamics. Given the importance of the menstrual cycle, it may seem surprising that data on how investigational drugs may impact the menstrual cycle are not already routinely collected in clinical trials; however, research typically reflect the people, priorities, and purposes of those within the clinical trial ecosystem—that is, the individuals and systems that fund clinical research, conduct clinical trials, and regulate the drugs tested in trials, as well as the individuals who participate in trials. Historically, there has been an underrepresentation of people who menstruate within the clinical trial ecosystem [ 29 ]. This exclusion is true for much of the preclinical research that informs clinical trials across many biomedical fields as well, and even cell lines used in in vitro studies are predominantly derived from male animals [ 30 , 31 ]. Although proof-of-concept studies for drugs intended for use in women that are known to impact the menstrual cycle, such as hormonal contraceptives, do typically use female animals when the model organism has an estrous or menstrual cycle, other preclinical research disproportionately relies on only male animals. Using both female and male animals in the research that informs clinical trials, however, could provide early indications of any impacts on cycles, as well as many other sex-specific effects or differences. Despite decades of concrete efforts, sex and gender disparities persist in the clinical trial ecosystem [ 32 – 34 ]. Within the current clinical trial context, another element relevant to our review is how trials typically incorporate outcomes, like menstrual changes, that are reported by trial participants. The United States Food and Drug Administration (FDA) and NIH refer to these data as patient-reported outcomes (PROs), which they define as “a measurement based on a report that comes directly from the patient (i.e., study subject) about the status of a patient’s health condition without amendment or interpretation of the patient’s response by a clinician or anyone else.” PROs can include “symptoms or other unobservable concepts known only to the patient (e.g., pain severity or nausea) [that] can only be measured by PRO measures,” as well as “the patient perspective on functioning or activities that may also be observable by others” [ 35 ]. Unless an assay or biomarker are used, all outcomes on menstrual changes are reported by the person who menstruates and, therefore, are PROs. The FDA has a series of methodological guidance documents on the development, validation, and use of PROs in clinical trials as part of patient-focused drug development efforts [ 36 – 39 ]. Given the aim of the review, our review questions were: (a) What instruments have been developed to assess menstrual changes caused by any etiology or source? and (b) What is the quality of these instruments and their utility for clinical trials? The objective of our systematic review was to compile a complete list of validated instruments used to measure menstrual changes with an assessment of their quality and clinical trial utility.

Across databases, our searches yielded 8,490 articles. We removed 376 duplicates, excluded 7,704 articles during title and abstract screening, and excluded 236 articles during full text review. In total, we identified 174 relevant full text articles. We present additional details on our search results and screening in the PRISMA diagram in Fig 1 . * Per Page et al ., 2021 [ 40 ]. We found some similarities across papers that we excluded for not meeting our inclusion criteria. For example, we excluded conference presentations that never became full papers, studies that focused on validating instruments among only menopausal populations (e.g., [ 54 , 55 ]), and studies that only validated surgical or treatment outcomes (e.g., [ 56 , 57 ]). In addition, there were two recent papers on core outcome sets for HMB and endometriosis relevant to the wider topic of measuring changes to the menstrual cycle, but we were unable to include them because there were no instrument details to extract [ 58 , 59 ]. Over 85% of the 174 articles were from either Europe (43%), North America (32%) or Asia (13%), and there were less than 15 articles from South America (n = 13), from the Middle East (n = 11), from Oceania (n = 8) and from Africa (n = 5). Just under half of articles were from only the United States (28%) or the United Kingdom (16%), although we did identify articles from a total of 50 countries. Nearly all articles were in English—even those reporting on instruments in other languages—except for two in Portuguese [ 60 , 61 ]. The most common study designs were cross-sectional or prospective cohort. We present details of all 174 included articles in S2 Table . From the 174 included articles, we extracted 94 instruments. Almost three quarters (72%, n = 68) were full instruments, collecting data on one or more menstrual change. Nearly a quarter (n = 21) were broader instruments that included sub-scales (9%, n = 8) or a small number of items (14%, n = 13) on menstrual changes. Five percent (n = 5) were general instruments validated in menstruating populations on one or more menstrual change. The instruments with the most articles in our review were the Endometriosis Health Profile-30 (EHP-30; 20 articles), the Pictorial Blood Loss Assessment Charts & Menstrual Pictograms (PBAC; 11 articles), the Uterine Fibroid Symptom and Quality of Life questionnaire (UFS-QOL; 9 articles), the Polycystic Ovary Syndrome Quality of Life scale (PCOS-QOL; 8 articles), and the Endometriosis Health Profile-5 (EHP-5), Menstrual Attitudes Questionnaire (MAQ), and menstrual collection (5 articles each). About a third (38%, n = 26) of full instruments used electronic data collection, and almost all full instruments (97%, n = 66) were completed by only the patient/participant who menstruated (i.e., they were PROs per the FDA and NIH definition). We present the list of the 68 full instruments and instrument characteristics in Table 4 . The remainder of results reported below are for these full instruments, with details on the sub-scales, items, and general instruments in S3 Table . * According to publications, "Yes" indicates either fully or partly electronic. ** Sum of quality and utility scores for studies conducted after 2006. † Evidence score based on only one study before 2006. †† Score provided for every aspect of quality (no scores of 0 in any category). ‡ Score provided for every aspect of utility (no scores of 0 in any category). Of the 68 full instruments, two-thirds were in English (66%, n = 45), followed by Spanish (13%, n = 9), French (9%, n = 6), and Portuguese (9%, n = 6); however, we identified instruments in 28 languages. About forty percent of instruments (41%, n = 28) were only in English, although about a quarter of instruments (26%) were in more than one language, and six instruments were in at least 4 languages. These instruments included the EHP-30 (13 languages), UFS-QOL (5 languages), MAQ (5 languages), PBAC (4 languages), Endometriosis Daily Diary (EDD; 4 languages), and the Daily Diary (4 languages). Nearly 60% (n = 40) of the 68 full instruments were developed and/or validated in populations with menstrual or gynecologic disorders or symptoms (i.e., 18 for endometriosis, 10 for HMB, 9 for dysmenorrhea, and 3 for uterine fibroids). Less than a quarter (24%, n = 16) of full instruments were developed for and validated with adolescents (mean ages less than 18, n = 10) or young people (mean ages early 20s, n = 6). Three full instruments were specifically developed for those in perimenopause. A few instruments were developed or validated in populations of athletes or people in the military. No instruments or articles indicated inclusion of trans and gender nonbinary people who menstruate. Among the 68 full instruments, half (49%, n = 33) measured bleeding, nearly half (47%, n = 32) measured uterine cramping or pain, and almost three quarters (74%, n = 50) measured perceptions. Only eight (12%) measured blood characteristics. Three instruments assessed all four of the parameters of bleeding—duration, volume, frequency, and regularity/predictability (i.e., the Aberdeen Menorrhagia Severity Scale [AMSS], the New Zealand Survey of Adolescent Girls’ Menstruation, and the World Endometriosis Research Foundation Endometriosis Phenome and Biobanking Harmonisation Project Standard Questionnaire [WERF EPHect EPQ-S]). No instrument assessed all three parameters of blood—color, consistency, and smell. Across the four aspects of menstrual changes (i.e., bleeding, blood, pain, and perceptions), no instrument measured all parameters for each aspect, and only seven instruments measured at least a single parameter of each aspect. These instruments were the AMSS; electronic Personal Assessment Questionnaire—Menstrual, Pain, and Hormonal (ePAQ-MPH); Endometriosis Self-Assessment Tool (ESAT); Fibroid Symptom Diary (FSD); Menstrual Bleeding Questionnaire (MBQ); Menstrual Insecurity Tool; and the New Zealand Survey of Adolescent Girls’ Menstruation. We present in Table 5 details on how the 68 full instruments measured bleeding (i.e., the four parameters of duration, volume, frequency, and regularity/predictability), blood (i.e., the three parameters of color, consistency, and smell), uterine cramping/pain, and perceptions. * There were 7 instruments with sub-scales that collected data on bleeding, 8 instruments with one to five items on bleeding, and two general instruments with items that asked about bleeding. Most sub-scales and items were for bleeding volume or regularity/predictability, often using terms not clearly defined or elaborated (e.g., ‘regular’ and ‘normal’). See S3 Table for details. ** One instrument had a subscale that collected data on blood color, consistency, and smell (i.e., the Menstrual Cycle-Related Signs and Symptoms Questionnaire subscale Section 1), and one other instrument had an item that asked about blood consistency (i.e., the Stellenbosch Endometriosis Quality of Life Measure). See S3 Table for details. † Four instruments with subscales, seven instruments with one to five items, and three general instruments asked about pain. See S3 Table for details. ‡ Two instruments with subscales, six instruments with one to five items, and four general instruments asked about perceptions. See S3 Table for details. When assessing measure quality, we found only five of the 68 full instruments (7%) had data on each of the six attributes of measure quality (i.e., conceptual or measurement model, reliability, content validity, construct validity, responsiveness, and sensitive nature of questions). These were the PBAC, EHP-30, Dysmenorrhea Daily Diary, MBQ, and a quantitative model for menstrual blood loss [ 85 ], each indicated by †† in Table 4 . All but three instruments (96%, n = 65) had evidence of a conceptual or measurement model and most also included evidence of content validity (81%, n = 55), construct validity (84%, n = 57) and reliability (66%, n = 45); however, less than a third of instruments had evidence on responsiveness (31%, n = 21), and less than a fifth (19%, n = 13) had evidence on question sensitivity ( Fig 2 ). Of the 68 full instruments, 18% (n = 12) had an overall good measure quality score, about three quarters (74%, n = 50) had a fair measure quality score, and 9% (n = 6) had a poor measure quality score ( Table 4 and Fig 2). When we looked at individual attributes of measure quality, over half of instruments had a good score for content validity (56%, n = 38), 47% had a good score for reliability (n = 32), 44% had a good score for conceptual or measurement model (n = 30), and over a third of instruments (35%, n = 24) had a good score for construct validity; however, only a quarter had a good score for responsiveness (25%, n = 17), and only 4 instruments (6%) had a good score for question sensitivity. When assessing clinical trial utility, we found 11 full instruments (16%) had data on each of the five attributes of utility (i.e., interpretability of results, transferability, participant burden, and investigator burden), each indicated by ‡ in Table 4 . All but three instruments (96%) had information on participant burden, 84% (n = 57) had evidence of the interpretability of the instrument results, and slightly less than two thirds (60%, n = 41) had documented investigator burden; however, only just over one third (37%, n = 25) had evidence of transferability ( Fig 3 ). Of the 68 full instruments, 22% (n = 15) had an overall good clinical trial utility score, almost two thirds (62%, n = 42) had a fair score, and 13% (n = 9) had a poor score ( Table 4 and Fig 3). When we looked at individual attributes of clinical trial utility, almost half of instruments (49%, n = 33) had a good score for the interpretability of results, about 40% had good scores for participant burden (41%, n = 28) or investigator burden (40%, n = 27), but only 8 instruments (12%) had good scores for transferability. Only the PBAC had evidence on all attributes of measure quality and all attributes of clinical trial utility, and only three instruments had both a good measure quality score and a good clinical trial utility score: EHP-5, the Spanish Society of Contraception Quality-of-Life (SEC-QOL), and the SAMANTA Questionnaire. Thirteen instruments had both measure quality scores and clinical trial utility scores greater than 2.5. Only one instrument, the Squeezing Pain Bulb, had both poor measure quality and poor clinical trial utility. Full instrument total evidence scores ranged from 4 for the World Health Organization Disability Assessment Schedule 2.0 to 332 for the EHP-30, with an overall median score across instruments of 16 and mean score of 27 ( Table 4 ). Overall, the following instruments had the five highest scores across measure quality, clinical trial utility, and total evidence: EHP-30, EHP-5, UFS-QOL, PBAC, and MBQ.

Despite the novel, broad, and transdisciplinary approach to our systematic review, the current instrument landscape, limitations in the literature, and gaps in evidence on measure quality and clinical trial utility indicate there is a need to examine changes to the menstrual cycle in a more complete, inclusive, and standardized way. Rigorous formative research—across sociocultural contexts—that is focused on how all people who menstruate experience and understand their menstrual cycles and more fully addresses menstrual stigma can inform the development of new or modified instruments to meet this need. We also identified a need for greater evidence of the validity for existing and new instruments. For the clinical trial context, FDA guidance on selecting, developing, or modifying patient-reported outcomes like menstrual changes indicate there must be evidence to support the use of an instrument for the specific concepts of interest and context of use [ 37 ]. At a minimum, per this guidance, evidence would be needed to support the use of the instruments identified and assessed in this review in the clinical trial context with a broader patient population (i.e., context of use) and to measure the full scope of menstrual changes that people experience (i.e., concept of interest). In addition, the recent emergence of core outcome sets within areas like HMB and endometriosis will be useful to promote standardization of validated instruments, especially if these efforts are interdisciplinary and coordinated across research areas. The findings of our review will be helpful in developing new or modified instruments that assess menstrual changes in a validated, comprehensive way. If used across the many fields that study menstrual health, data from these standardized instruments can contribute to an interdisciplinary, systemic, and holistic understanding of menstruation and the menstrual cycle. In turn, this improved understanding can be translated into ways to enhance the health and wellbeing of people who menstruate.

We conducted our systematic review in alignment with Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [ 40 – 42 ], including a protocol registered in PROSPERO (Protocol ID: CRD42023420358) [ 43 ]. A completed PRISMA checklist for this review is in S1 Table , and S1 Appendix includes additional details on the search strategy, inclusion/exclusion criteria, title/abstract screening, full text review, data extraction, and data analysis. We searched for peer reviewed articles in the MEDLINE and Embase literature databases and for any relevant instruments measuring menstrual changes in four instrument databases: (a) the NIH Common Data Element (CDE) Repository [ 44 ], (b) the COSMIN database of systematic reviews of outcome measurement instruments [ 45 ], (c) the Core Outcome Measures in Effectiveness Trials (COMET) Database [ 46 ], and (c) ePROVIDE databases [ 47 ]. Table 1 shows the final search strategy for MEDLINE, and S1 Appendix includes search strategies for other databases. We uploaded articles from the literature database searches and articles for any relevant instruments identified via the instrument databases into Covidence [ 48 ]. Following screening and review of these articles in Covidence, we identified relevant review articles and extracted primary articles published since 1980 from those reviews. During data extraction, we identified any original articles for instruments developed before 2006. We uploaded these primary articles and original development articles into Covidence for screening. We included all peer-reviewed articles—including those with prospective, retrospective, or cross-sectional study designs, and review papers—that met our inclusion and did not meet our exclusion criteria, listed in Table 2 . *Examples are: (a) an increase in how long bleeding lasts (bleeding duration), (b) a reduction of clotting (blood consistency), (c) a decrease in dysmenorrhea (pain), and (d) an impact on quality of life or attitudes (perceptions of changes). We held weekly author meetings to discuss progress, questions, and discordance, and to document decisions in a shared Word document. We began title/abstract screening with an ‘inter-reviewer reliability’ meeting where all authors completed title/abstract screening on the same 50 articles to establish and confirm group standards. Then, two authors independently screened each remaining title/abstract and two authors independently reviewed each relevant full text in Covidence. We resolved any discordance during weekly meetings via consensus conversations. We conducted data extraction in Excel using a template data extraction form that collected article information, study design, sample information, details on the instrument, measure quality attributes, and clinical trial utility attributes. For articles not in English, we used the text translation feature of Google Translate to review titles and abstracts, we used the document translation feature of Google Translate and/or consulted a fluent colleague to review full text articles, and we completed data extraction with a fluent colleague for included articles. For assessing measure quality and clinical trial utility, we followed the recent Patient-Reported Outcomes Tools: Engaging Users and Stakeholders (PROTEUS) Consortium recommendations to use the International Society for Quality of Life Research (ISOQOL) standards for PRO measures [ 50 , 51 ]. We made two adjustments to the ISOQOL standards: (a) we added an attribute on sensitivity of questions given the topic of menstruation has a noted amount of stigma surrounding it [ 52 ]; and (b) we separated out participant burden from investigator burden given these two can differ greatly for instruments measuring menstrual changes. We categorized six attributes as related primarily to the quality of the instrument (i.e., measure quality : conceptual/measurement model, reliability, content validity, construct validity, responsiveness, and sensitive nature of questions) and four attributes as related primarily to the utility of the instrument in clinical trials (i.e., clinical trial utility : interpretability of results, the transferability of the instrument, participant burden, and investigator burden). We scored each attribute of measure quality and clinical trial utility on a scale from 0 to 3 (0 = no data, 1 = poor , 2 = fair , and 3 = good ) based on criteria in line with ISOQOL standards [ 51 ] that were reviewed by measurement and clinical experts at FHI 360 and within a related global task force. We show the measure quality and clinical trial utility attributes and scoring criteria in Table 3 , and S1 Appendix includes details on the other fields of the data extraction form. * Attributes and definitions from Reeve et al . 2013 [ 51 ] per PROTEUS-Trials Consortium guidance [ 50 ], with modified as specified in the text. † Crossnohere et al ., 2021 [ 53 ]. We conducted data analysis in Excel and included counts and frequencies, as well as specific analyses to assess instrument measure quality and clinical trial utility. For the measure quality score and clinical trial utility score of an instrument, we used an average of the highest score for each attribute of measure quality or clinical trial utility across all articles on an instrument. Because instruments could have more than one article providing data on measure quality and/or clinical trial utility and not every article evaluated all attributes of an instrument, we did not include scores of zero (i.e., no data reported) in these averages. To reflect these differences in the number of articles and attributes reported in the article(s), we also calculated a total evidence score , which was the total of all scores—including zeros—across all attributes of measure quality and clinical trial utility. The total evidence scores, therefore, ‘penalize’ instruments for a lower level of evidence due to fewer articles or less attribute data and vice versa. These three scores—measure quality (ranging from 1–3), clinical trial utility (ranging from 1–3), and total evidence (ranging 0+)—reflect different dimensions of an instrument. For example, two instruments might both have a score of 2.5 for measure quality, but one instrument might have an evidence score of 10 and the other, 100, indicating the latter has considerably more evidence and likely more certainty in the measure quality score. Alternately, two instruments may have similar measure quality and evidence scores, but one may have a clinical trial utility score of 1 and the other a score of 3, indicating the latter is likely better suited for use in clinical trials despite the similar levels of measure quality and evidence.

(PDF) (PDF) (PDF) (PDF)