Measurement in the study of menstrual health and hygiene: A systematic review and audit.

Menstruation is a recurring experience in the lives of millions of women and adolescent girls across the globe. This natural process has gained increased attention for its role in female health and social participation in recent years, following a history of neglect and silence. Policy and programming have rapidly expanded, seeking to address unmet menstrual needs. This response has far outstripped the pace and funding of research to understand menstrual experiences and inform and evaluate interventions.[ 1 ] A growing body of qualitative research has highlighted the challenges faced by menstruating women and adolescent girls in low-resource settings and indicated negative effects on health, education, employment, and well-being. These studies have also highlighted a complex array of factors contributing to experiences of menstruation.[ 2 ] Quantitative studies testing risk factors or consequences hypothesized through qualitative research are rare, and systematic reviews of quantitative and trial research have highlighted the limited number and low quality of extant studies.[ 3 – 5 ] A significant challenge for quantitative research of menstrual health has been a lack of clarity around core concepts and a paucity of measurement tools to capture them. In 2016, Hennegan and Montgomery [ 3 ] highlighted this inconsistency in the measurement of outcomes across trials, as well as the absence of relevant menstrual health measures to capture experiences as a key barrier to improved trials of interventions. Similarly, in mapping the knowledge of menstrual health and hygiene across study designs, Chandra-Mouli and Patel [ 6 ] noted that “vague measures are often used to describe the menstrual experiences of girls , which impede data aggregation and direct comparisons . ” In proposing research agendas for menstrual health, many have called for improved clarity in the measures used, refinement of concept definitions, consistency in outcomes assessed, and the development and validation of new measures where needed [ 7 – 9 ]. While past reviews and research priority papers have highlighted measurement challenges as barriers, none have reviewed existing measures to provide an appraisal of current use and gaps. The present study seeks to inform improved measurement in the rapidly emerging field of menstrual health by auditing extant measurement. We aimed to describe the current concepts assessed in studies of interventions and identify the measures that have been used to capture core concepts. Two linked, simultaneous systematic reviews were undertaken. Review A included trials of menstrual health interventions in low- and middle-income countries (LMICs), as well as studies nested within those trials. The aim of the first review was to identify the concepts and outcomes measured in studies of interventions, supplemented by nested studies which may provide further explorations of menstrual experiences, measures, intervention theory of change or process evaluation. Review B collated and appraised studies developing, validating or testing tools to measure menstrual experiences across geographies. This second review aimed to identify the concepts for which measures have been developed and to highlight lessons learned from measure development and validation. By integrating the two reviews we were able to compare the measures used in trials to those developed and tested to date. Together, this work provides an overview and appraisal of current measurement in menstrual health and hygiene research and develops recommendations for pathways forward.

Searching and screening results are presented in Fig 1 . Characteristics of included studies are summarized in Tables 1 and 2 . NR: Not Reported; WASH: Water, Sanitation and Hygiene; SD: Standard Deviation. NR: Not Reported; WASH: Water, Sanitation and Hygiene; SD: Standard Deviation. Among the trial studies, 13 tested effects of education interventions, 8 tested product provision interventions or compared comfort and acceptability of different products (n = 4), and two tested combined education and product provision. Products tested included disposable and reusable pads (n = 6), menstrual cups (n = 2) and a combination of pads and cups (n = 2). Most studies focused on girls in schools or of school age (n = 18), with an additional 2 studies undertaken with university students. One study exclusively concerned girls who reported experiencing menstrual pain. One study sought to provide training for adolescents with mild intellectual disability (IQ 50–70) and their parents. No other studies reported including participants with disabilities. Nested studies explored intervention effects and menstrual experiences in more detail, with study aims reported in Table 1 . Of measure development studies, 11 developed a scale or item set to measure a defined latent construct, 6 investigated the performance, revalidation or a new language version of an existing measure, and 5 evaluated the accuracy of self-reported menstrual characteristics such as the date of menarche. Six studies were published before 2000, nine in the 2000’s and a further seven since 2016. More studies were conducted in North America than in other regions (n = 10), followed by Europe (n = 4). There were two studies each with samples from Iran and India, and single studies undertaken in Turkey, El-Salvador and the Philippines, Belize, Mexico and Israel. Many studies were undertaken with girls in schools (n = 9), and university students (n = 8). From Review A we included measures used as outcomes as well as those used descriptively to contextualise participant experiences and in nested studies. Concepts described by each study’s authors, their methods of assessment and a brief summary of measure development or validation efforts are reported in S2 Table . Most measures were designed by investigators for the purpose of their individual study, with some referring to external experts or past research to inform questions. Eight trials reported piloting measures prior to data collection or were themselves pilots to assess the feasibility of the measures tested. To describe and compare measures we iteratively grouped concepts measured across studies, displayed in Table 3 . From Review B we extracted the measures developed, their sub-scales and example items from each, reported in Table 4 . To compare developed measures from Review B with measures in trials (Review A) we highlight in Table 4 where measures, or sub-scales within measures, assess concepts corresponding to those in trials and nested studies. For studies developing or testing scales we extracted reported tests of dimensionality, reliability and validity and report these in S3 Table . Across scale development and validation studies most investigated measure factor structure, although many employed only exploratory factor analysis and did not include a confirmatory analysis. Internal consistency was reported for most measures as the sole test of reliability. Of the 11 studies testing new measures, only three reported test-retest reliability. Most studies appraised face and content validity through expert and participant input, or comparison to qualitative study findings. Far fewer studies included further quantitative validation such as tests of convergent, discriminant or predictive validity. This evidenced the lack of a clear nomological network, theory about the interrelationship between constructs and their measures, to guide the selection of related constructs. One measure of attitudes assessed relationships between the scale and mental health, self-esteem and locus of control [ 71 ]. While a second tested relationships with self-objectification [ 68 ]. A final instrument testing a broad menstrual health construct, assessed criterion validity against self-reported menstrual symptoms and quality of life measures [ 69 ]. A debilitating event (e.g., “Avoiding certain activities during menstruation is often very wise”) A natural event (e.g., “Menstruation provides a way for me to keep in touch with my body”) A bothersome event (e.g., “Menstruation is something I just have to put up with”) An event whose onset can be predicted and anticipated (e.g., “I have learned to anticipate my menstrual period by the mood changes which precede it”) An event that does not and should not affect one’s behaviour (e.g., “Premenstrual tension/irritability is all in a woman’s head”) Danger (e.g., “Women should not hold positions of power or authority because of mood changes during the menstrual cycle”) Stigma (e.g., “Menstruation should be kept secret”) Superstition (e.g., “Women should not cook while menstruating”) 4) Disability (e.g., “Menstruation is a significant cause of absence from work or school for women”) Positive Feelings (e.g., “I feel proud when I have my period/I will feel proud when I get my period”) Negative Feelings (e.g., “I worry a lot about periods starting unexpectedly/I worry a lot about my periods starting”) Living with Menstruation (e.g., “Girls with periods should avoid exercise”) Openness (e.g., “I like to talk about periods with my friends”) Acceptance of Menarche (e.g., “Coping with periods is easy”) Menstrual Symptoms (e.g., “Menstruating girls are grumpy and tense”) Secrecy (e.g., “It is important to discuss the topic of the period at school with boys and girls together”) Annoyance (e.g., “It is annoying for women to have the period every month”) Proscriptions and prescriptions (e.g., “Women must avoid exercising while they are having their periods”) Disability (e.g., “The period affects women’s abilities to do housework”) 5) Pleasant (e.g., “There are women who feel content to have their periods”) Menstruation as Bothersome (as appears in the MAQ) Menstruation as Disgusting or Shameful (e.g., “I would feel ashamed if I “leaked” menstrual blood on my clothes”) Menstruation as Enabling Awareness of One’s Body (e.g., “Menstruation is a reoccurring affirmation of womanhood”) Menstruation as Life Affirming (e.g., “The recurrent monthly flow of menstruation is an external indication of a woman’s general good health”) Adolescents and family interaction (e.g., “My father/mother welcome that I raise my issues with them”) The perception of female adolescents of premarital sexual relationships (e.g., “I think that having sexual relationships with a boyfriend is a sin”) Enabling factors for sexual and reproductive self-care (e.g., “Access to healthcare services such as visits to a gynaecologist or midwife, a psychologist and a nutritionist etc. enhances my ability for reproductive and sexual health self-care”) Understanding and behaviours of female adolescents of the interaction with the opposite sex (e.g., “Because of my adherence to family principles, I refrain from having a relationship with a boy”) Parent-adolescent communication barriers (e.g., “I do not want to talk with my mother/father about issues related to the opposite sex/puberty and menstrual cycle, so that our respect is preserved”) Reproductive and sexual knowledge (e.g., “Genital ulcers are a sign of STDs”) Self-care for reproductive health and menstruation (e.g., “If I cannot dry my underwear in the sun, I use a hot iron”) Attitudes towards menstrual health (e.g., “Menstruation causes difficulties in concentrating on some activities such as education”) Subjective norms (e.g., “My family believes that I should continue my social activities during menstruation”) Perceived behavioural control (e.g., “I can take a shower during my menstrual period”) Perceived parental control (e.g., “My parents determine how much I should read about the puberty health-related issues”) Behavioural intention (e.g., “I have decided to frequently change my menstrual pad during my menstrual period”) Menstrual health behaviours (e.g., “I don’t go to the sea and the pool during my menstrual period”; “I would use cotton underclothes during my menstrual period”) Menstrual health (e.g., “What kind of menstrual absorbent do you use?”) Menstrual hygiene (e.g., “Is there a separate bathing place at home?”) Affective symptoms (e.g, “I have mood swings during my period”) Somatic symptoms and school life (e.g., “I have lower abdominal pain or discomfort during my period”) Daily habits for menstrual health (e.g., “I have dietary habits of eating less salty food and taking less caffeine”) Menstrual cycle characteristics (e.g., “I have healthy menstrual cycles and periods”) Attitudes and perceptions on menstruation (e.g., “I think menstruation is an important indicator of women’s overall health”) School participation Stress Self-efficacy Audit of measures used across included studies are reported according to the groupings in Table 3 . Of 32 included trials and nested studies, 19 reported measuring different practices undertaken to manage menses, not all of these were considered ‘hygiene’ behaviours and so are referred to here and menstrual and hygiene practices. In studies, these were typically referred to as ‘menstrual practices’, ‘menstrual hygiene’, ‘menstrual behaviours’ or ‘hygiene behaviours.’ Eight trials included changes to menstrual practices as outcomes, making this the second most common trial outcome measure. Seven trials used menstrual and hygiene practice data to describe samples, while 5 nested studies reported on practices as part of their core research question. The list of behaviours and practices included in studies as menstrual and hygiene practices are reported in Table 5 for Review A and B studies. These do not include indexes from three studies which collapsed across an unknown set of items [ 20 , 43 , 46 ]. Further, in two studies, authors included attending school or university, and participating in religious practices during menstruation [ 16 ], and food restrictions and exercise [ 18 ] as menstrual practices. These fit poorly with those reported as hygiene behaviours in other studies and are not included in Table 5 . There was no consistent definition of menstrual or hygiene practices across studies, even among those using such concepts as trial outcomes. Djalalinia et al. [ 17 ] defined menstrual hygiene as “bathing and washing during the period of menstruation after each urination and defecation , and use of sanitary pad or cotton” , while Leventhal et al. [ 21 ] defined menstrual hygiene as the use of menstrual products and the frequency of changing products. For adolescents with mild intellectual disability, Sener and colleagues [ 41 ] evaluated menstrual hygiene as bathing practices, and observational assessment of demonstrating placing a menstrual product on a doll. Shesthra et al. [ 44 ] used menstrual practices to balance across intervention and control groups at baseline, implying hygiene as the menstrual materials used and if respondents had ever used sanitary pads. In nested studies, the menstrual and hygiene behaviours of focus were defined by the study objectives. Both Nyothach et al. [ 35 ] and Oduor et al. [ 36 ] investigated hand washing before and after changing menstrual cups as ‘handwashing for menstrual hygiene’ and the frequency of dropping menstrual products and subsequent cleaning or management of those products, respectively. In one nested study, [ 31 ] authors based the assessment of menstrual hygiene on a pre-existing definition developed by the Joint Monitoring Programme of the World Health Organization and UNICEF in 2012 [ 74 ] to report the prevalence of the concept and its association with other outcomes. Very few studies reported the questions used to asses menstrual and hygiene practices. Thus, it was unclear what the recall periods, question structures, and response options were for most studies. Five studies disclosed the self-report questions used for core practices assessed [ 21 , 30 , 31 , 35 , 36 ]. Among these, ‘usual’ practice was most commonly assessed, with some asking for ordinal responses (‘always’, ‘sometimes’, ‘never’) to characterise their practice. Two studies used ‘this recent period’ as the recall period [ 35 , 36 ]. Two studies included insights on the reliability of menstrual and hygiene practices questions used. van Eijk 2018 [ 38 ] found poor agreement between self-reported menstrual cup use and cup use measured by observed change of the cup colour (kappa = 0.044). Womena 2018 [ 50 ] reported that quantitative survey responses in which respondents reported washing their materials with soap and water contradicted qualitative accounts wherein participants reported inadequate access to soap for washing. Among measure development studies, two focused on measuring menstrual hygiene [ 63 , 67 ]. An additional two measures included menstrual health or hygiene behaviours as subscales [ 51 , 58 ]. Where reported in studies, the practices measured in these scales are included in Table 5 . Menstrual hygiene . As part of the development of new measures for the Multiple Indicator Cluster Surveys (MICS) [ 75 ], Khan and colleagues [ 63 ] field tested questions on menstrual hygiene. They undertook cognitive interviews with respondents and focus group discussions with interviewers administering the surveys. Results indicated that for the three menstrual questions asked, in 38–52% of cases interviewers needed to clarify the questions or probe to elicit responses. Cognitive interviews suggested that the term ‘washing’ may have been interpreted by respondents to infer bathing, while the term ‘privacy’ was understood differently among respondents in interviews. In focus groups, interviewers suggested that ‘materials’ was not consistently perceived to mean a menstrual absorbent. Menstrual hygiene management scale . Ramaiya [ 67 ] developed a menstrual hygiene measure based on a past framework, and included domains of: use of menstrual absorbents (cloth, sanitary pad) and hygiene behaviours, which included genital washing, handwashing and bathing daily (see Table 5 ). Items were reported along with the categorisation of practices as ‘adequate’, ‘semi-adequate’ or ‘inadequate’ menstrual hygiene. Factors at baseline and endline differed with items all loading on individual factors in final analysis. The author suggested that revised understandings of the constructs were needed. In two broad measures of ‘reproductive self-care’ [ 51 ] and ‘menstrual health seeking behaviours’ [ 58 ], sub-scales included measures of menstrual self-care practices. The former included the subscale ‘Self-care for reproductive health and menstruation’, this included the menstrual practices listed in Table 5 , along with items that fit more poorly with menstrual hygiene practices reported in most Review A studies including: using iron pills, recording the dates of the menstrual period, using painkillers to manage menstrual pain, and monitoring menstrual blood loss. The latter measure assessed menstrual health seeking across constructs from the theory of planned behaviour (TPB) and included a subscale on behaviours. Alongside the practices listed in Table 5 , authors included items reported to capture menstrual health behaviours including; avoiding swimming in pools during menses, avoiding caffeine and reducing aggression during menstruation. These items evidenced inconsistencies in the boundaries of self-care behaviours, restrictions during menstruation, and hygiene practices. Knowledge about menstruation and puberty was the most frequent outcome assessed in trials (n = 10). A further two studies reported measuring menstrual knowledge for descriptive purposes. Knowledge was assessed through tests. The content of knowledge assessments varied, as did the level of detail provided by authors about the topics covered. Although noting expert or text-book input into knowledge measures, four studies provided no information on the content included in the tests [ 13 , 16 , 42 , 43 ]. For the studies that did provide example questions or indicate the coverage of knowledge content, there was a wide range. The indication that menstruation was a physical process, the age of onset of menarche, and the origin of menstrual blood as the female reproductive tract (uterus, through the vagina) were common across studies. Some studies broadened biological knowledge to include secondary sexual characteristic changes during puberty such as the development of breasts or hips, the timing of ovulation and links between menstrual cycle and reproduction. Others included knowledge of menstrual disorders e.g., the definition of dysmenorrhea, causes of pain or discomfort, and included questions about self-care for pain during menses [ 18 , 20 ]. Four studies reported including questions about hygiene practices during menstruation as part of knowledge assessments such as the types of materials to use as absorbents, and the frequency with which one should change materials [ 20 , 24 , 44 , 46 ]. In their list of examples, Blake 2018 [ 15 ] included “Girls should stay home from school when they are menstruating” as part of their knowledge test as a reflection of the content of the puberty book provided as part of the intervention. A second study [ 46 ] also stated including questions about avoiding foods or physical activity practices to be undertaken during menses in knowledge assessments. These questions were similar to those used to capture menstrual restrictions in other studies. We did not identify any eligible studies developing measures of menstrual knowledge. One broad measure of reproductive self-care included a subscale on knowledge and attitudes towards open discussion of sexual and reproductive health topics [ 51 ]. In this sub-scale, some items concerned freedom to discuss menstrual and reproductive health topics with parents, although many items were more focused on sexual knowledge such as sexually transmitted diseases. In our iterative groupings, eight studies included measures aiming to capture the acceptability of the tested interventions. In three studies comparing different menstrual products, satisfaction, comfort and product preferences or willingness to continue use were primary outcomes [ 14 , 26 , 45 ]. Mohamed and colleagues 2014, [ 26 ] also compared physiological responses to wearing sanitary pads of different thicknesses to further assess product performance. Shrestha 2018 [ 44 ] compared product demand in response to awareness raising and discounts, using coupon redemption for products as an objective measure of demand. Two other pilot trials sought to understand the acceptability of interventions to their target recipients using qualitative feedback and self-report [ 48 , 50 ]. One nested study compared ratings of product reliability, comfort and satisfaction between participants given a reusable sanitary pad as part of the trial to those using their existing materials, and assessed participant willingness to continue use of the provided pad [ 30 ]. Questions used across studies varied specific to research questions or intervention. No included studies developed measures of menstrual product or intervention acceptability. Five trials assessed intervention impacts on menstrual attitudes, and one additional trial reported assessing ‘menarche experience’ comparing self-reported attitudes at menarche. Two studies used the MAQ [ 55 ] and AMAQ [ 65 , 66 ]. Three studies used self-created attitudes questionnaires. Blake et al. [ 15 ] described a combined knowledge and attitudes questionnaire, although it was unclear what attitude questions were included and no examples were provided. Additionally, the study included two items capturing fear and shame in association with menstruating, with the items reported in full: “Does the idea of menstruating make you feel afraid or shameful?” and “Is talking about menstruation something that is shameful for you?”. Similar items were described as psychosocial well-being outcomes in another program of work [ 27 , 29 ]. Fetohy [ 20 ] provided little description of the menstrual attitude scale used, noting only that it assessed “attitude toward healthy and unhealthy practices during menstruation”. Afsari and colleagues [ 47 ] developed a 15-question attitude scale which included items regarding menstruation, nutrition, exercise, physical activity and psychological and mental health, but no further detail was provided. Djalalinia [ 17 ] reported comparing whether participants reported feeling good (happy or proud, and thankful) at the moment of menarche in methods, and in results presented if participants felt confused, scared, uncomfortable, or good. One study assessed the impact of a menstrual intervention on beliefs regarding menstruation and sexual behaviour. In the study, this was separated from menstrual knowledge and included ten items capturing: whether girls should continue education after menarche, marriageability at menarche, and if there was a relationship between eating sweets and menstrual bleeding. A second study [ 44 ] assessed stigma and norms about menstruation, although it did not feature in the main text reporting of results. Stigma questions were described and asked women to report whether they were allowed in the kitchen, holy places, secluded to a shed or considered untouchable during menstruation; these were more similar to items capturing restrictions in another study. As noted in the section on menstrual and hygiene practices, some studies included avoiding physical activity, education or other activities during menstruation as practices. One nested study [ 30 ] compared the impact of menstrual materials used by participants on whether they avoided daily activities such as cooking, or doing sports during menstruation, and if there were activities menstruation caused them to miss. These were compared according to the menstrual materials used suggesting they reflected limitations due to material quality, but these also may reflect attitudes towards appropriate behaviours during menstruation and did not fit easily into any iterative groupings in the review. Most Review B measure development studies sought to capture attitudes towards menstruation ( Table 4 ). The Menstrual Attitude Questionnaire (MAQ) . The MAQ was developed in 1980. The initial questionnaire was tested among both male and female university students. Concurrently, developers proposed a shorter, simplified version for adolescent individuals pre and post menarche. As displayed in Table 3 , the scale has five factors. Attempts to revalidate the MAQ produced mixed results. Stubbs and colleagues (1988) found poor internal consistency of subscales and proposed an alternate factor structure, capturing ‘affirmations’ of menstruation, and ‘worry’ or dislike of menstruation. Authors assess validation through relationships with depression and anxiety symptoms, self-esteem, locus of control, and body satisfaction. Over a decade later, Bramwell and colleagues (2002) found the MAQ factor structure was not an acceptable fit in British or Indian samples. Similarly, a Turkish version of the MAQ exhibited poor fit for the original factor structure. Authors undertook follow-up exploratory analyses to propose an alternate factor structure broadly like the original MAQ, with some items loaded differently. The original factor structure was not supported in a Greek version of the MAQ, although similar to the Turkish version, follow-up EFA determined a similar five-factors with items re-distributed. With the hypothesis that the framing of menstruation in a positive rather than negative light might cause respondents to view their menstrual experiences differently, two studies [ 52 , 56 ] used a parody Menstrual Joy Questionnaire (MJQ) (not eligible for this review due to a focus on symptoms) to examine the effect of priming positive views of menstruation on responses to the MAQ, as well as a questionnaire assessing the severity of menstrual symptoms not eligible for this review (the Menstrual Distress Questionnaire, MDQ). Chrisler et al.[ 56 ] found exposure to the MJQ resulted in more positive responses to menstruation on the MAQ, while in a second study Aubeeluck et al. [ 52 ] found that the MAQ primarily increased scores on the ‘menstruation as a natural event’ subscale of the MAQ. The Adolescent Menstrual Attitude Questionnaire (AMAQ) . In 1993, Morse and colleagues developed an alternate attitude measure for adolescents with a form for premenarcheal and postmenarcheal girls. Developed from the results of qualitative studies rather than the MAQ, the AMAQ covers a different range of topics (see Table 4 ). Authors sought to identify differences between girls before and after menarche and develop a meaningful measure for both groups. Stereotypic Beliefs about Menstruation Scale (SBAM) . This measure was reported only in a conference abstract and focused on the assessment of negative stereotypical beliefs about menstruation [ 61 ]. Authors compared subscale scores between men and women (see Table 4 ). Beliefs about and attitudes towards menstruation questionnaire (BATM) . Marvan and colleagues [ 64 ] developed the BATM to capture Mexican adults’ attitudes towards menstruation ( Table 4 ). The measure was tested among Mexican and North American populations. Authors noted differences between samples with the Mexican sample reporting higher expectations for avoiding swimming, carrying heavy items and avoiding certain foods. Men also endorsed higher proscriptive attitudes than women. Menstrual self-evaluation scale . This measure was developed in 2004 and uses two sub-scales from the MAQ (‘bothersome’ and ‘menstruation as a natural event’) in addition to six new items [ 68 ]. Self-evaluation scale scores were associated with self-objectification and objectified body consciousness, consistent with author hypotheses as predictive validity. One trial (Abedian) measured menstrual pain and blood loss using a variety of measures beyond the scope of this review to compare between intervention education conditions and controls. Four other trials used a variety of menstrual characteristics to provide a picture of the sample [ 16 , 17 , 34 , 50 ]. All described the proportion of the population experiencing menstrual pain or dysmenorrhea. This was self-reported, although no studies described the questions used. Three studies also reported the duration of menses, either in the average number of days [ 34 , 50 ] or the proportion of the sample experiencing ‘longer periods’ defined as more than 5 days of bleeding [ 16 ]. Age at menarche was reported by Phillips-Howard et al. [ 34 ], and used in one nested study to compare menstrual cup uptake according to the time since menarche [ 38 ]. Five measure development studies focused on the accuracy of self-reported menstrual characteristics. These studies evaluated the timing and modality of surveys and their impact on recall accuracy. Two studies examined women’s self-reported cycle length, comparing respondents cycle recorded in daily diaries to self-reported average cycle length [ 62 , 70 ]. Jukic [ 62 ] reported that diary recorded cycle length over 6 months showed moderate agreement with self-reported cycle length (kappa = 0.45) and reported that women overestimated their cycle length by an average 0.7 days. In this study, almost 35% of women reported having a 28-day cycle, but the diary observed cycles suggested a less peaked distribution with between 15 and 20% of women recording cycles 26–28 days. In contrast, Small [ 70 ] found a similar near 40% of women reporting a 28-day cycle, while diary data suggested a less peaked distribution. In this study women underestimated their cycle length by 1.5 days. Weller [ 73 ] assessed women’s concepts of menstrual regularity and irregularity, comparing self-reported irregularity to diary records of menses. Authors coded irregularity as those with one third of cycles over a six-month period as more than 35 days or less than 21. Approximately 70% were classified as regular in both methods. However, only 44% who reported irregularity were coded as irregular using diary data, while some of the women who considered their periods to be regular were coded to have irregular periods (18%). Wegienka [ 72 ] found that 56% of women accurately reported the date of their last menstrual period and that 81% reported it accurately within +/- two days. Authors also noted that a duration of 3 weeks or longer since the last period was associated with overestimation of the time since the last period. In their US sample, education level was not associated with recall accuracy. Cooper [ 57 ] found that 85% of women at age 48 accurately recalled their age at menarche (reported when 14–15 years old) within 1 year.

Multiple trials assessed the impact of MH interventions on education. School attendance measured through a combination of school registers [ 27 , 29 , 40 ], spot checks [ 29 , 40 ], diaries [ 32 , 34 ] and self-report surveys [ 48 ] was the main education outcome measured in all studies assessing this concept. Phillips-Howard et al. [ 34 ] also assessed school drop out as a primary outcome. Attendance across all days was used in both Montgomery 2012 and 2016 [ 27 , 29 ], while Oster 2011 used a combination of attendance registers, absenteeism diaries and menstrual records to assess absences during menstruation [ 32 ]. Phillips-Howard et al. [ 34 ] found that diaries were unreliable and were unable to compare conditions on attendance using this method and so did not report findings in the primary trial report. In 2019, Benshaul-Tolonen and colleagues used the school register and spot-check data collected as part of that trial to compare intervention effects [ 40 ]. This study also used additional register and spot-check data on non-trial students and boys in the trial schools to compare attendance as recorded by these different methods, finding non-random inconsistencies between spot-checks and attendance records. Wilson et al. [ 48 ] reported descriptively the days missed due to menstruation but compared conditions based on the total number of days girls self-reported missing in the preceding month. Educational engagement was inferred in a pilot trial and one nested study [ 27 , 31 ]. In surveys, girls self-reported if they were able to concentrate in school during their period. Identified through Review B, Haver and colleagues described efforts to develop a measure including menstrual related self-efficacy, stress and school participation [ 60 ]. They did not report tests of a final measure, discussing difficulties in initial piloting and validation. Authors noted that items used to measure the three pre-defined latent constructs overlapped in girls’ experiences. Two trials compared scores on the Strengths and Difficulties Questionnaire [ 76 ] as secondary psychosocial well-being outcome [ 27 , 29 ], and this score was also included in study nested in the second trial [ 31 ]. In addition, these trials included items asking girls to report on shame and insecurity during menstruation compared to when they were not menstruating. These items are similar to those assessed by Blake et al. [ 15 ] which assessed girls fear and shame associated with menstruation. Leventhal et al. [ 21 ] included a range of psychosocial measures as outcomes, although linked these to the resiliency components of the intervention rather than any attention to menstruation [ 22 ]. These measures included emotional resiliency, general self-efficacy, social-emotional assets, depression, general anxiety, positive psychological well-being and social well-being; all using previously developed and established measures [ 23 ]. Phillips-Howard et al. [ 34 ] reported assessing wellbeing through the Paediatric Quality of Life Inventory (PEDSQL; [ 77 ]) and used scores to describe balance across trial conditions at baseline. Well-being was not compared as an outcome in trial reports, although was listed among secondary outcomes in trial registration. The inclusion criteria for this review were designed to target menstrual-specific measures. Generalised measures of psychosocial functioning and well-being are beyond the scope of this review. One included measure study described the development of a measure of menstrual-related psychosocial outcomes including sub-scales on stress and self-efficacy, but noted that these concepts overlapped in pilot testing and did not report a final measure [ 60 ]. Two trials comparing menstrual products assessed discomfort. Beksinska et al. [ 14 ] asked participants to report comfort levels and adverse events associated with the sanitary pads and menstrual cups, including vaginal irritation and dysuria. Stalder et al. [ 45 ] asked participants to report on discomfort attributes including; feeling hot/sweaty/stuffy, itching, chafing, soreness/tenderness, redness or presence of a rash, wet or sticking feeling, or burning. The Phillips-Howard trial [ 34 ] compared the impact of sanitary pad and menstrual cup provision on sexually transmitted diseases, C . trachomatis , T . vaginalis , and N . gonorrhoea , and reproductive tract infections bacterial vaginosis or C . albicans using vaginal self-swabs. They also tested for adverse events, including the presence of Staphylococcus aureus . A study nested within this trial further explored potential for adverse events by testing the presence of S . aureus vaginal colonization and Escherichia coli grown on samples of menstrual cups provided to girls [ 37 ]. These were compared according to the duration of cup use. One nested study investigated the association of menstrual hygiene practices reported by participants in the trial with self-reported reproductive tract infection symptoms which included: skin irritation or rashes in the pelvic area, itching or burning in the pelvic area, and white or green vaginal discharge since the start of the school year [ 31 ]. Reproductive tract infection and biomarker methods were beyond the scope of our Review B inclusion criteria focused on menstrual experience. In five studies, investigators assessed Water, Sanitation and Hygiene (WASH) conditions. WASH conditions were assessed to describe the comparability of infrastructure between eligible schools and intervention conditions. In one trial, self-reported WASH infrastructure at home was used to screen for eligible participants (those with a municipal water supply). In one nested study, investigators assessed if there were any changes in WASH conditions throughout the duration of the study; infrastructure to support menstruation did not improve, although there was increased soap availability [ 39 ]. One study used a combined measure of ‘menstrual health’ as the primary outcome [ 19 ]. This included items on a range of different features of menstrual experiences, including menstrual practices, impacts of menstruation on daily activities and school attention, dietary choices, iron supplementation and exercise. Two included measure studies also concerned broad conceptualisation of menstrual health. Menstrual Health Instrument . Developed in a South Korean context [ 69 ], the construct was informed by literature review and authors priorities to include both symptoms experienced during the menstrual cycle as well as self-care, attitudes and perceptions about menstruation. Authors validated the measure against self-reported menstrual cycle symptoms and quality of life measures. Menstrual health seeking behaviours questionnaire . Developed in Iran among adolescent girls,[ 58 ] items were based on the theory of planned behaviour (TPB) and included all TPB model components from perceived control and attitudes through to intentions and behaviours. Items were drawn from review, qualitative studies and theoretical constructs.

Through this systematic review and audit, we provide a reference point capturing the current measures used in menstrual health and hygiene research. We found that measures were inconsistent in conceptualisation and operationalisation across studies. Findings suggest that interdisciplinary efforts are needed to better define core constructs such as menstrual and hygiene practices, menstrual knowledge, attitudes, norms and restrictions. In absence of consensus among researchers, authors should clearly define the concepts measured for the purposes of their study and should transparently report measures used. We recommend including full surveys in published supplementary materials where feasible. Research is needed to develop and test reliable and valid measures for core concepts. To support measure development, mid-level theory of menstrual experiences [ 2 ] and intervention effects [ 78 ] should be used to inform priority measures and identify concepts against which such measures can be tested for validity.

The review protocol is registered on PROSPERO: [CRD42018089884] and is reported according to PRISMA guidance ( S1 Table ). Two systematic searches were undertaken in English, reported in Box 1 . Trial searches were conducted in the following 12 databases: Cochrane Central Register of Controlled Trials (CENTRAL), Cumulative Index of Nursing and Allied Health Literature (CINAHL), ProQuest Dissertation and theses, Embase, Global Health, Medline, Open Grey, Popline, PsycINFO, Social Sciences Full Text, Socoiological Abstracts, WHO Global Health Library. In addition, we screened the first 20 pages of Google Scholar results, and trial registries: Clinical Trials Registry, Pan African Trials Registry, Trials Register of Promoting Health Interventions (TRoPHI). To identify nested studies, we undertook vertical searching of reference lists and citations of included trials. Searches for measure development and validation studies were undertaken in the same databases but excluded CENTRAL and trial registries. We searched the reference lists and citations of two past systematic reviews [ 3 , 4 ] as well as a report reviewing the state of menstrual health research [ 10 ]. To identify grey literature, we searched online databases specific to menstrual health, hygiene and sanitation: Menstrual Health Hub, Menstrual Hygiene Day ‘Resources’, and Sustainable Sanitation Alliance (SuSanA library). Further we searched the websites of key organisations undertaking work in menstrual health: Oxfam, PATH, Plan International, Save the Children, UNICEF, UNFPA, UN Women, WaterAid, WASHUnited, WSSCC. Initial searches were undertaken in English in March 2018 and updated in July 2019. Titles and abstracts were independently screened by two reviewers using EPPI-reviewer 4, with full text screening undertaken by the first author. Inclusion criteria for trial design were consistent with those applied in a past systematic review of menstrual health and hygiene intervention studies [ 3 ]. Randomized and non-randomized trials which included a control group (including controlled before-after studies) were eligible for inclusion [ 11 ]. Trials were eligible if they evaluated the effectiveness of interventions designed to improve the menstrual experiences of women or girls. Interventions could include interventions such as puberty education or social programs designed to improve social support or reduce menstrual stigma. Interventions providing supportive resources and environments were also eligible, such as the provision of menstrual products (e.g., sanitary pads) or improvements to water, sanitation and hygiene (WASH) infrastructure. We included studies that compared the acceptability, comfort or experience of using different menstrual products when they met study design criteria. Nested studies were eligible if they included quantitative data on the menstrual experiences or environments of participants. Nested qualitative studies were not eligible. Studies of menstrual experiences nested in trials which did not meet inclusion criteria, or trials not yet reported (e.g., only baseline data available) were not included. Studies including women and girls of reproductive or pre-reproductive age in low- and middle-income countries were eligible [ 12 ]. We were unable to include studies that were not available in English or Spanish. Studies were eligible if they reported on the development and validation, or tested the performance of, measures of menstrual experiences. The aim of this review was to inform future measurement in the study of menstrual health and hygiene in low-resource settings and we sought to identify any measures of menstrual management behaviours, attitudes, knowledge or the impact of menstruation on quality of life. To align our review of measures with menstrual health and hygiene research and interventions focused on experiences of non-disordered menstruation, we excluded measures designed to diagnose menstrual disorders such as endometriosis or heavy menstrual bleeding. Measures that focused on symptoms associated with hormonal cycling, that is, physical symptoms associated with menstruation, were also excluded. Similarly, we excluded measures of the experiences of disorders of the menstrual cycle including: Polycystic Ovary Syndrome (PCOS), dysmenorrhea, or endometriosis. We excluded measures of the acceptability of menstrual suppression or experiences of menopause. Studies from any country were eligible. Review A and Review B study data were extracted using piloted forms by one reviewer and checked by a second. For Review A we extracted concepts measured as described by each study’s authors and iteratively grouped the concepts measured. Within each grouping we summarized the ways different concepts had been described and measured. Where described we extracted measure specifics, such as piloting or validation, question wording or recall periods, to further inform comparisons. As the aim of this review was to audit the measures used in trial, we did not assess risk of bias in trial designs. Through Review B we collated the measures developed to date and the contexts in which these had been assessed. Included measure studies ranged from those assessing the accuracy of single self-report questions, to those developing scales relating to latent constructs. Where latent constructs had been described, we extracted authors’ definitions and any subscales along with example items from the measures. We grouped Review B measures according to the iterative conceptual groupings developed in Review A. This enabled us to describe the availability of measures for the concepts of focus in intervention research. We did not undertake formal quality appraisal of measure development studies, as the concepts measured were poorly aligned with trial outcomes, requiring more attention to concept analysis. We supplemented this with brief summary of the performance of the developed measures and the concepts studies used to establish convergent, discriminant or predictive validity as this is likely to be most useful for future measure development efforts.

(PDF) Click here for additional data file. (PDF) Click here for additional data file. (PDF) Click here for additional data file.