Examining physical activity and menstrual cycle experiences among insufficiently active women: Project PACE protocol

Cardiovascular disease (CVD) remains the leading cause of morbidity and mortality in the US, imposing substantial health and economic burdens (Kazi et al., 2024 ; Slavin et al., 2021 ). Women of reproductive age (ages 18–50) with CVD risk factors such as obesity and high blood pressure represent a high-priority and vulnerable subgroup, as alarming trends show that these chronic conditions that confer CVD risk are increasingly prevalent (Abramson & Melvin, 2014 ; Giardina et al., 2016 ; Norhammar & Schenck-Gustafsson, 2013 ; Peng et al., 2019 ). CVD risk is further escalated by high (and increasing) rates of anxiety and depression among women in this age range (Faravelli et al., 2013 ; Krittanawong et al., 2023 ). Reproductive health conditions such as Polycystic Ovary Syndrome (PCOS) are also relevant given their strong associations with increased cardiometabolic risk, while Premenstrual Dysphoric Disorder (PMDD) is characterised by cyclical affective and behavioural symptoms that may influence engagement in cardioprotective health behaviours. Thus, the reproductive years represent a critical window for establishing regular engagement in these behaviours. Physical activity (PA; i.e. bodily movement that requires energy expenditure) (World Health Organization, 2024 ) is an essential component in the prevention of CVD and occurs across multiple domains (e.g. leisure time and occupational activity), which may differentially relate to health outcomes (Li & Siegrist, 2012 ). Importantly, the total amount of PA accumulated across domains (reflected in measures such as daily step count) helps to mitigate CVD risk by lowering blood pressure (Kokkinos et al., 2009 ), improving insulin sensitivity (Bird & Hawley, 2017 ), reducing inflammation (Burini et al., 2020 ; Nimmo et al., 2013 ), and enhancing mood, sleep, and metabolic health (Franklin et al., 2022 ). Despite these well-documented benefits, women of reproductive age do not meet recommended levels of PA (Bucciarelli et al., 2023 ). This gap is especially concerning among women with CVD risk factors, for whom the benefits of PA are greatest, yet participation is often lowest. This is due to converging barriers such as the competing demands of work, school, and caregiving responsibilities (Batalha et al., 2025 ) as well as stress and limited time (Ominyi, 2025 ), all of which constrain opportunities for consistent engagement in PA. Among women with CVD risk factors, these barriers may be further compounded by health limitations, increased healthcare demands, and socioeconomic constraints. Hormonal fluctuations across the menstrual cycle may also influence women’s PA engagement (Srinivasa Gopalan et al., 2024 )—specifically, menstrual cycle symptoms may present challenges with adherence to PA programs or individual PA goals (Carmichael et al., 2021 ). To promote PA among women of reproductive age, targeted interventions should consider the biological, social, and behavioural contexts that shape PA patterns, including experiences related to the menstrual cycle. The menstrual cycle is characterised by cyclical fluctuations in sex hormones that are associated with common physical and affective symptoms (e.g. bloating, breast tenderness, fatigue) (Bruinvels et al., 2022 ; Srinivasa Gopalan et al., 2024 ). These symptoms represent proximal, subjective experiences through which hormonal fluctuations may influence daily behaviour (including PA). We propose that hormonal fluctuations may be associated with PA behaviour indirectly through their effects on menstrual experiences (i.e. subjective physical and affective experiences across the cycle), which in turn may shape PA cognitions (i.e. psychological determinants of behaviour such as motivation, intentions, and self-efficacy) and daily PA engagement. Consistent with Social Cognitive Theory (Bandura, 1986 ), behaviour is influenced by dynamic and reciprocal interactions among personal factors (i.e. biological, cognitive, and affective processes), environmental influences, and behaviour itself, suggesting that these pathways may also operate bidirectionally. Much of the existing work in this area—particularly in sport and exercise science—has focused on athlete populations, with broad patterns of PA and performance assessed in laboratory settings (e.g. laboratory-based performance testing), rather than among general populations of women (Bruinvels et al., 2022 ; Carmichael et al., 2021 ; McNulty et al., 2020 ). Traditional experimental approaches typically overlook day-to-day variability in behaviour and lived experiences across the menstrual cycle. For example, menstruating women may be especially sensitive to daily changes in pain and related symptoms, as well as to fluctuations in PA cognitions, sleep quality and quantity, and stress. Emerging evidence suggests that cycle-related hormonal fluctuations are associated with variation in physical and emotional experiences that are relevant to daily health behaviours (Schoep et al., 2019 ), including PA. Women commonly report changes in energy (Meignié et al., 2021 ), mood (Farage et al., 2008 ; Sundstrom Poromaa & Gingnell, 2014 ), sleep (Rugvedh et al., 2023 ), and motivation (Reusch et al., 2025 ) across the cycle, which may affect their ability or desire to engage in PA - particularly when symptoms such as pain, fatigue, or irritability are present. In this way, hormonal changes may be associated with PA behaviour through their impact on symptoms and cognitive determinants such as motivation, intentions, and perceived capability. Critically, these fluctuations may have an even greater impact among women who are not sufficiently active (vs. athletes), as physically active women often exhibit reduced hormonal variability across the menstrual cycle (Ahrens et al., 2014 ). For women with CVD risk who may experience systemic or psychosocial barriers to PA, cycle-related barriers may exert more influence on PA engagement (Csontos et al., 2024 ; Salvatore et al., 2025 ; Srinivasa Gopalan et al., 2024 ). The nature and magnitude of these day-to-day associations are poorly understood, highlighting the need for intensive, within-person research to inform improved strategies for promoting cardiovascular health in this population. Advances in behavioural research methods offer an approach that could directly address these weaknesses. Traditional global assessment methods such as validated self-report measures are designed to attenuate short-term variability and capture stable, person-level differences. This approach aggregates over many experiences; with respect to frequent or repeated experiences (or behaviours), can introduce considerable recall bias (Smyth et al., 2017 ), and does not capture the natural variability in experiences that occur daily (e.g. menstrual symptoms, PA). Consequently, little is known about variability in these experiences for the same person at different times or under different circumstances (Dunton, 2017 ; Reichert et al., 2020 ; Smyth & Stone, 2003 ; Smyth et al., 2017 ). In contrast, intensive ambulatory assessment in the natural environment, including device-based assessment of PA, can reduce retrospective recall bias and increase measurement accuracy by assessing experiences as they occur in daily life. These methods can capture individuals’ experiences in real time as well as in their natural environments, providing critical insights informing whom, when, and how to deliver behavioural PA interventions among women of reproductive age. Women in the US may be reluctant to self-disclose menstrual symptoms in global self-reports for reasons such as stigma, internalised gender norms, embarrassment (Åkerman et al., 2024 ), or the broader sociopolitical climate (Salvatore et al., 2024 ) (which may not be supportive of women’s healthcare). These factors can hinder accurate assessment of menstrual experiences across many measurement approaches. In contrast, intensive assessment methods (e.g. repeated, brief assessment of the same participant over time through digital devices) may reduce barriers to disclosure by allowing participants to report symptoms privately, in naturalistic settings, as well as closer to when they occur; this could increase ecological validity (Heron & Smyth, 2010 ; Stephen et al., 2007 ) and minimise the need for disclosure to another person (such as in an interview). Specifically, intensive assessment methods and wearable technology are feasible and acceptable among menstruating women when used independently to assess PA (Heinonen et al., 2012 ; Maijala et al., 2019 ; Shilaih et al., 2017 ) or menstrual health (Epstein et al., 2017 ; Goodale et al., 2019 ). Concurrent use of wearable devices and electronic questionnaires to track menstrual symptoms and PA also shows promise among women in this life stage (Lang et al., 2024 ; Salvatore et al., 2025 ; Salvatore et al., 2025 ). A recent scoping review identified only seven studies that examined PA behaviour among insufficiently active women of reproductive age using intensive ambulatory assessment methods (Salvatore et al., 2025 ), suggesting opportunities to better understand PA behaviours in this group and key opportunities for intervention in the natural environment. For example, this work can identify optimal windows for intervention delivery (such as on days when PA motivation is lower or symptom severity is higher than a person’s typical levels) and how to adapt content and PA support accordingly. First, however, there is a need to gather evidence regarding the temporality of these daily experiences for effective tailoring. This approach will increase our understanding of associations between daily experiences of the menstrual cycle, PA, and other biopsychological processes in daily life (e.g. sleep), including when symptoms are most likely to occur (e.g. irritability, sadness). Assessing these experiences in daily life could identify critical gaps in knowledge of women’s health and enable just-in-time behavioural interventions. To address these limitations, the present study combines device-based ambulatory assessment of PA with intensive self-report methods, providing detailed insight into potential associations between women’s daily experiences. Specifically, we aim to determine the extent of within-person variability in (1) device-assessed PA behaviour (e.g. steps per day), (2) self-reported menstrual experiences and cognitive PA determinants (e.g. motivation), and (3) temporal associations between menstrual symptoms and objectively measured PA at the day level. We hypothesise that participants will show highly variable PA behaviour, menstrual experiences, and cognitive PA determinants (i.e. motivation, intentions, self-efficacy) within-person, reflecting low stability within and across days (H1a-1c). We also hypothesise that on days when participants experience more (vs. less) intense menstrual symptoms relative to their typical levels, they will report lower cognitive PA determinants (i.e. motivation, intentions, self-efficacy; H2). Finally, we hypothesise that lower cognitive PA determinants will be associated with less PA behaviour (e.g. steps per day; H3). These associations will provide preliminary evidence regarding potential pathways linking menstrual experiences and daily PA.

Project PACE (Physical Activity and Menstrual Cycle Experiences) was designed to examine temporal relations between menstrual symptoms, PA, and other daily experiences (e.g. sleep, social comparisons) among women of reproductive age. Eligibility requires that women are 18–50 years old (inclusive), insufficiently active (e.g. <6,000 steps per day or <90 minutes of moderate-to-vigorous PA per week), and menstruating regularly (i.e. having ≥10 menstrual periods per year; not yet entered menopause based on self-report). As the average age of menopause is 51 (Menopause 101: A primer for the perimenopausal, n.d. ), with the transition phase (i.e. perimenopause) lasting between 4–8 years (Menopause 101: A primer for the perimenopausal, n.d. ), we selected this age range to represent the target population of menstruating women. Eligibility criteria also include ≥1 marker of CVD risk (i.e. overweight or obesity [BMI of 27–50], hypertension or prehypertension, type 2 diabetes or prediabetes, hypercholesterolaemia or hyperlipidemia, metabolic syndrome, current smoking or quit within the past 3 months), ability to read and write in English, absence of medical contraindications to PA, and no severe psychopathology that would interfere with participation (e.g. psychosis or substance use disorder requiring immediate clinical treatment) (Arigo et al., 2020 ). Hormonally related mood changes are not exclusionary unless they reflect a level of severity that impacts one’s ability to complete the study. Women with menstrual disorders (e.g. PCOS, endometriosis) are otherwise eligible to ensure a comprehensive understanding of women’s daily experiences. Globally, PCOS affects 6–13% of women of reproductive age (Dubé-Zinatelli et al., 2025 ), endometriosis affects about 10% (Yan et al., 2025 ), and PMDD affects 3–8% (Halbreich et al., 2003 ). Participants using contraceptives are also eligible, as contraceptive use represents a critical component of reproductive healthcare (Daniels & Abma, 2025 ). We aim to recruit a minimum of 60 participants who successfully complete study activities, including 45 days of observation and an exit interview. This sample size was calculated to afford minimally adequate power for primary tests in multilevel models (described below) to detect small- to medium-sized effects, following estimates from simulations by Hox ( 2002 ). Power is more limited for between-person predictors and moderator analyses; see Limitations. Participants are recruited through advertisement from a range of electronic sources, including postings to social media (e.g. X/Twitter, BlueSky), women’s health listservs, daily email announcements from the supporting university, and a database of women who previously expressed interest in studies on women’s health. A brief electronic questionnaire includes demographics, self-reported reproductive health history, contact information, recent psychosocial experiences (e.g. sleep disturbance, feelings of anxiety or depression), and overall social perceptions (e.g. menstrual attitude, social comparison orientation). Additional details about these measures are available in Table 1 . Baseline measures. Participants wear a Fitbit Inspire device continuously during the 45-day data collection period. Outcomes of interest include steps per day and minutes of exercise as calculated by the Fitbit company, to capture total daily PA accumulated across domains (e.g. occupational, leisure-time, and other daily activities). See Table 2 for an additional description of PA outcomes. Intensive ambulatory assessment of physical activity behaviour and sleep experiences. Presence of a menstrual period, presence of menstrual symptoms and severity, and other experiences are assessed with 3 items, twice per day on all study days. First, participants indicate whether they are menstruating today. They then rate the presence and severity of 16 typical symptoms (modified from Steiner et al. ( 2003 ); see Table 3 ). Respondents can also complete an open-ended item to provide additional context or experiences that may not be covered by other items (e.g. less common symptoms), thereby ensuring that diverse menstrual experiences are represented. Intensive assessment via self-report measures. Angry or frustrated Sad or down Happy or excited Content or satisfied Anxious or stressed Confident or hopeful Fatigued or sleep – Not at all-A little bit – Somewhat – Quite a bit – Extremely Not at all-A little bit Somewhat Quite a bit Extremely – Friends – Coworkers – Acquaintances – Family – Strangers – Classmates – Other (please specify) Friends Coworkers Acquaintances Family Strangers Classmates Other (please specify) Comparisons to people who seem to be doing better than I am Comparisons to people who seem to be doing about the same as I am Comparisons to people who seem to be doing worse than I am These experiences are assessed twice daily using 3 items (see Table 3 for items and response scales). These items were adapted from validated scales and were designed for intensive, repeated measurement, allowing detection of within-day fluctuations (Arigo et al., 2020 ; Bandura, 2006 ; Salvatore et al., 2025 ). Device-based assessment of sleep duration and efficiency is continuously measured over the 45-day data collection period using a wrist-worn Fitbit Inspire device. These values are estimated based on movement and heart rate data, providing an objective measure of sleep experience. See Table 2 for an additional description of sleep outcomes. Additionally, subjective sleep quality is assessed once daily, and fatigue and sleepiness are assessed twice daily. Additional experiences including affect, body satisfaction, pain, social experiences, and social comparison are assessed twice per day. Selecting ‘yes’ to instances of social experiences and social comparisons prompt follow up questions about the frequency, mode, and direction (Arigo et al., 2020 , 2024 ; Salvatore et al., 2026 ). See Table 3 . Enrolment begins with an online screening questionnaire via Qualtrics (Provo, UT), followed by a 15-minute phone call to confirm eligibility. Those who are eligible and elect to move forward are sent a Fitbit device via US mail and asked to complete a baseline electronic survey to assess person-level experiences (e.g. typical level of premenstrual symptoms). They then complete a 60-minute orientation session via Zoom, where they engage in informed consent discussions and are introduced to study procedures (including all survey items). Starting the day after orientation, participants complete 45 days of intensive ambulatory assessment: they report on physical symptoms (i.e. menstrual, body pain), PA cognitions (i.e. motivation, self-efficacy), and other psychosocial experiences (i.e. sleep, mood) twice daily via electronic surveys. They receive prompts to complete these surveys at pre-specified times via text message or email (based on their preference). PA and sleep data are collected continuously on the same 45 days using a wrist-worn Fitbit Inspire device. Participants receive a check-in from the research team on a weekly basis via email (or phone if they prefer) to problem-solve any issues with the technology or survey completion; these check-ins also afford opportunities to maintain engagement and promote retention through the end of the study period. Following 45 days of observation, participants attend a 30-minute exit interview via Zoom to provide feedback on their experience in the study. Participants can earn $75 via an electronic gift card over the course of this study; $60 for completing the study and a $15 bonus for ≥90% compliance with daily survey completion. We will use multilevel modelling to address the nested structure of our data: assessments (Level 1) within days (Level 2) within individuals (Level 3). Our primary research questions focus on within-person associations between menstrual experiences and daily PA, using two categories of predictors: (1) menstrual symptom presence and intensity (Levels 1-2), and (2) secondary exploratory person-level moderators (i.e. Level 3) such as age, BMI, identity indicators, hormonal contraceptive use, self-reported reproductive health conditions, and menopause status. These person-level moderators were selected based on their potential relatedness to menstrual experiences, symptom burden, and PA engagement and will be examined to determine whether within-person associations vary across individuals. Missing data patterns will be evaluated and predictors of missingness (e.g. weekday vs. weekend day) will be included as covariates in subsequent models; sensitivity analyses will be used to determine their effects on results and conclusions. To examine temporal sequencing of these associations, models will include lagged predictors such that symptoms from the prior assessment (e.g. previous survey or day) are used to predict PA outcomes. All models will employ maximum likelihood estimation techniques, which allow for the inclusion of all available cases and are robust to missing data (i.e. missed surveys, missed items within surveys). Initial empty models will evaluate the proportion of variance accounted for at each level, for each PA parameter (e.g. steps per day, active minutes), using the intraclass correlation coefficient (ICC). Subsequent models will use person-mean centreing to differentiate stable versus time-varying associations between predictors and outcomes (Curran & Bauer, 2011 ; Hoffman, 2015 ). Analyses will test components of the overall conceptual model, to offer preliminary evidence to inform future tests of the proposed pathways. Specifically, we will examine components of the proposed mediation framework by examining within-person associations between: (1) menstrual experiences (predictor) and daily PA outcomes, (2) menstrual experiences (predictor) and cognitive PA determinants (i.e. potential mediators: motivation, intentions, self-efficacy), and (3) cognitive PA determinants (predictors ) are associated with daily PA outcomes. Interpretation will emphasise estimates of effect size, confidence intervals, and model fit indicators rather than statistical significance alone, to inform formal evaluation using multilevel mediation models.

This observational study received institutional review board approval in February 2025, and we enrolled the first participant that same month. As of May 2026, we enrolled and followed 64 participants; 4 withdrew from the study due to the length of data collection or technology issues (e.g. ongoing difficulty with Fitbit syncing/pairing). Figure 1 shows a CONSORT flowchart for enrollment and study completion. Data collection is complete and we expect to finish data analysis and report study findings by December 2026. CONSORT flow diagram.

Project PACE has several notable strengths. This study employs intensive real-time data collection, allowing for accurate capture of menstrual symptoms, PA, and related experiences, while limiting recall bias. This study is one of the first to examine temporal associations between menstrual cycle experience and PA and will provide novel insights into these experiences as they occur. Additionally, this work is guided by Social Cognitive Theory (Bandura, 1986 ) and the ORBIT Model of Intervention Development (Czajkowski et al., 2015 ). As an ORBIT Phase 0 study, Project PACE is designed to address a clinically significant question and will clarify key aspects of future PA promotion efforts. In addition, we prioritise broad representation by enrolling both naturally cycling participants and those on hormonal contraceptives, as well as including participants across different (early) stages of menopause. This will allow us to account for variations in daily experiences, helping to identify for whom and when to intervene. The following limitations are noteworthy. First, the aim of this work is to understand women’s daily menstrual cycle experiences and PA from a behavioural and psychosocial perspective, rather than to investigate underlying biological processes. Thus, we rely on women’s self-reported menstrual experiences; although assessing daily urinary hormone levels would lend biological verification, this is not essential for our current objectives and would increase cost and participant burden. The inclusion of hormonal biomarkers should be considered in future work. Further, we do not collect detailed information about menstrual flow (e.g. spotting, light bleeding, heavy bleeding), which may be relevant to understanding patterns in menstrual symptoms. Instead, the presence of a menstrual period is assessed using a binary (yes/no) response to whether participants are currently menstruating. While this simplifies daily reporting and minimises participant burden, it may limit the granularity in our understanding of menstrual experiences. Participants are provided with an open-ended prompt to allow for the reporting of additional menstrual-related information that may not be captured through survey items. Additionally, although this study is designed to examine within-person associations between menstrual experiences, cognitive PA determinants, and daily PA, it was not specifically powered to test a full multilevel mediation model. Analysis of potential mechanisms should be interpreted as preliminary and are intended to inform future hypothesis testing. Use of Fitbit also carries limitations. There is potential for underestimation of non-step based activities (Feehan et al., 2018 ) (e.g. resistance training) and accurate estimates rely on consistent participant device wear and syncing. Additionally, wearable-based measurement may not fully capture all activity contexts or short-duration movements if the device is not worn consistently. As participants are insufficiently active, we will likely capture most daily movement and have built-in weekly check-ins for the purpose of data verification and syncing. However, technological issues may interrupt syncing, and we track these for descriptive purposes. Finally, the generalisability of findings may be limited by the sample size and limited diversity. Project PACE is 100% remote, but as we are based in the northeastern US, this region may be overrepresented in our sample. Participants must also have access to the internet and a smartphone device to receive daily surveys; our sample may skew toward women of higher socioeconomic status and education level due to the length and technological requirements. Despite these limitations, this study will offer information about temporal associations that are difficult to observe using traditional retrospective methods. Examining menstrual symptoms as they occur will enhance our understanding of associations between menstrual cycle experiences and PA behaviour in women’s daily lives, which can identify targets for tailored, individualised interventions.

Women’s experiences of menstruation have been historically understudied and underrepresented in clinical research (Åkerman et al., 2024 ). Although existing evidence suggests that menstrual cycle experiences may influence daily PA behaviours among insufficiently active women of reproductive age (Csontos et al., 2024 ), the traditional approach to studying the menstrual cycle and PA has been to use cross-sectional, longitudinal, or experimental methods to describe (stable) differences between people. These approaches have considerable strengths, but they cannot map experiences as they occur in the natural environment. Project PACE will fill gaps in the existing evidence base using research methods that will capture perceptions and behaviours as they occur in daily life (i.e. intensive ambulatory assessment).