Preferences and responses for physical activity partnerships among women in midlife with elevated risk for cardiovascular disease: a digital, mixed methods approach

Abstract

Women in midlife (ages 40–65) experience significant risk for cardiovascular disease (CVD), including declining physical activity (PA). Lack of social support is a key barrier, but PA partnerships show promise for increasing PA. Little is known about women’s preferences for PA partnerships or how preferences are linked to women’s broader conceptions of PA. In this study, we used a mixed methods design to evaluate beliefs about PA and preferences for PA partnerships among women in midlife with ≥1 CVD risk factor. Participants (N = 27, M_Age=53.3, M_BMI=32.6 kg/m²) used a proprietary web application for 7 days to view hypothetical PA partners and completed a 1-hour qualitative interview. Thematic analysis indicated that some women in midlife hold specific perceptions about PA that inform a desire to feel validated and understood. Their definitions of PA and the role of caregiving are critical to PA and partner preferences; they prefer to partner with women at similar starting levels of PA, though they are open to a range of partners who can foster motivation and growth. Findings show that PA partnerships may be desirable to women in midlife with elevated CVD risk and point to critical next steps for optimising partner-based PA promotion for women.

Introduction

Women in midlife (ages 40–65) face a unique combination of stressors, including professional obligations, caregiving demands, shifts in social roles, and hormonal changes associated with menopause and biological ageing (Bédard et al., 2005; de Kat et al., 2017; Maffei et al., 2019). These stressors contribute to women’s elevated risk for cardiovascular disease during midlife (CVD; Slavin et al., 2021): by midlife, 50% of women in the US have CVD risk factors such as hypertension or obesity (Hales et al., 2017; Ostchega et al., 2020). Engaging in physical activity (PA) can be particularly helpful for mitigating this risk during midlife (Franklin et al., 2022; Reddigan et al., 2011). However, women tend to decrease their PA engagement during this period (whereas men maintain their previous PA; Finkel et al., 2018), and this trend has only increased since the COVID-19 pandemic—particularly for individuals with medical conditions (Stockwell et al., 2021; Wunsch et al., 2022).

Sustaining cardioprotective levels of PA presents challenges for women in midlife; they often struggle to set specific and attainable goals, identify intentions or plans to achieve these goals, and follow through on their intentions in the midst of competing demands and stressors (Arigo, Romano, et al., 2022; Knapova et al., 2024; Williams & French, 2011). This may be due to a broader difficulty maintaining sufficient motivation during habit formation. Lack of predictability in daily responsibilities is common among women in midlife, and PA plans are often set aside to manage unexpected needs such as child or elder care (Hendry et al., 2010). This lack of follow-through is known to be discouraging (Rhodes et al., 2022), and is likely to undermine women’s efforts to increase or maintain PA.

Women in midlife also cite lack of social support and lack of positive role models for PA as critical barriers to engagement, and these barriers have increased since the onset of the COVID-19 pandemic (Lum & Simpson, 2021). Social support is known to facilitate PA (Kouvonen et al., 2012; Uchino et al., 2018; Yuan et al., 2025), by providing information resources (such as suggestions for specific activities or places to be active), encouragement, accountability, and companionship (Marques et al., 2023). A strong sense of interpersonal connection (e.g. via support) has been linked to overall improvements in physical and emotional health and an increase in health-promoting behaviours such as PA (Wilson et al., 2021). Similarly, social comparison (i.e. self-evaluation relative to others) can offer inspiration, behavioural modelling, and motivation to keep up with or ahead of others (e.g. via friendly competition; Arigo et al., 2023; Bandura, 2013; Merchant et al., 2017).

As a result, many PA interventions are designed specifically for women in midlife, and several purport to activate social processes to promote PA in this group (Arigo, Romano, et al., 2022). As these interventions are only modestly effective (Murray et al., 2017), there is much to learn about women’s PA during midlife that could inform intervention improvements. This may be because the current understanding of women’s beliefs about PA and their preferred methods of activating social support and/or beneficial social comparison is incomplete. Consequently, there is a unique opportunity to address key barriers to PA among women in midlife who experience heightened risk for CVD, by harnessing social influences more effectively.

Physical activity partnerships to support women’s PA

The use of PA partnerships shows promise for meeting women’s social needs related to PA. Specifically, as women in midlife often do not have sufficient support or modelling in their existing social networks, PA partnerships can be created between women who do not already know one another for the purpose of providing support focused on increasing PA for its health benefits (Arigo, Schumacher, et al., 2015). Consistent with the theory of transactive goal dynamics (Fitzsimons et al., 2015), partnering with someone else who is insufficiently active and interested in increasing PA promotes mutual support for goals, creating a reciprocal relationship whereby the dyad facilitates resource-sharing. This may be more effective for women in midlife who are insufficiently active than learning from someone who is already active, as in peer coaching (Rowland et al., 2020).

Preliminary evidence indicates that PA partnerships are of interest to women and that they may facilitate daily PA change (Arigo, 2015; Arigo, Schumacher, et al., 2015; Murray et al., 2023). Importantly, however, evidence shows that health behaviour “support” can also hinder motivation and engagement (Mackert et al., 2011). This may occur if partners compare themselves with one another in ways that trigger negative emotional responses or otherwise reduce motivation for PA (Baga et al., 2024; Buunk & Ybema, 2013). For example, if one partner is already very successful with PA (cf. peer coaching), this could be inspirational for the partner starting with less PA, or could make achieving similar levels of PA seem unlikely (leading to discouragement). The extent of heterogeneity in women’s preferences for partners’ starting level of PA is not yet clear. Moreover, other factors that may be important to women in midlife with elevated CVD risk with respect to a potential PA partner—and how they may be related to broader beliefs or values related to PA—are unknown.

Given the potential for PA partners to meet women’s social needs around PA, a deeper understanding of the types of relationships or social connections that are crucial for this population would inform improvements to existing partner-based approaches. To achieve this understanding, we studied PA partner preferences in two ways. First, we asked women to use a new, proprietary web application to select and view profiles of other women in midlife, and to consider whether they might want this person as a PA partner. Participants responded to these peer profiles in real time, once per day for 7 consecutive days (quantitative method). At the end of this period, participants completed an in-depth qualitative interview to review their perceptions of peer profiles and their overall takeaways about their PA partner preferences in the context of their broader perceptions of PA (qualitative method). Exposure to peer profiles was intended to offer women opportunities to consider and reflect on their partner preferences, to inform their overall feedback during interviews. We then combined findings across data sources to address our research aims as comprehensively as possible (mixed method).

Aims of the present analyses

The overarching goal of the parent study was to inform a range of PA promotion resources for women in midlife with elevated risk for CVD. As in our previous work (Arigo et al., 2015), one resource of interest was an intervention program that would form PA partnerships between program participants; partners would be unfamiliar with one another and would be asked to provide mutual social support, which could be uncomfortable. Thus, decisions about how to pair partners may be critical to the success of PA partnerships, though the decision points for pairing are not yet known. Further, PA may be connected to larger goals, beliefs, and values, which may shape PA partner preferences. To better understand these factors and inform the design of a partner-based intervention, we followed the ORBIT Model of Intervention Development, which is uniquely applicable to behavioural interventions to promote women’s health (Walker & Czajkowski, 2019). The present study was designed to address ORBIT Phase 1a: specifically, guidance to “identify attitudes, norms, and values” of the target population that can affect the feasibility and acceptability of an intervention (Czajkowski et al., 2015; pp. 976). We conducted supplementary analyses of data from our parent study (Arigo et al., 2023), with the following aims: (1) to explore women’s beliefs about and barriers to PA, and (2) to describe women’s preferences for PA partners with respect to starting level of PA and individual difference characteristics. We selected a mixed methods approach to map the heterogeneity and depth of this topic; this approach offered women different ways to consider PA partners and partnerships and the opportunity to describe their preferences in their own words.

Materials and methods

Participants

As part of a larger program of research focused on developing tailored PA interventions for women in midlife with CVD risk factors, participants were recruited from prior study databases for the present phase of evaluation. This allowed us to ensure that women were in the target population for the intervention—specifically, those who met the criteria below and who had previously indicated an interest in increasing their PA. Recruitment focused on maximising available heterogeneity with respect to age, racial/ethnic identification, and socioeconomic status. Women ages 40–65 with one or more conditions that confer risk for CVD (i.e. hypertension, type 2 diabetes, high cholesterol, metabolic syndrome, or current smoking or quit within the last 3 months) were eligible if they regularly used an internet-connected device and were interested in testing a new digital PA tool designed for women in midlife. Evaluating potential partnerships represented only a subset of feedback collected and responses to other components are described elsewhere (Arigo et al., 2022, 2023). Of the 76 women contacted initially, 39 were screened for eligibility, and 27 enrolled in the study (M_Age=53.3, M_BMI=32.6 kg/m²). This sample size of 27 participants with 7 days of web application use each was expected to generate 189 brief partner preference observations and 27 unique narratives, affording considerable range for quantitative descriptive statistics and ample opportunity to describe women’s experiences and perceptions in depth (Braun & Clarke, 2021).

The majority of participants identified as white (59.3%) and a small subset identified as Hispanic/Latina (3.7%). Most had at least a bachelor’s degree (81.5%), endorsed a household annual income greater than $75,000 (52%), and indicated that they were postmenopausal (i.e. had their last menstrual cycle more than 12 months ago; 59.3%). Additionally, 48.6% of participants reported that they worked outside the home, 44.4% reported regularly providing care for children, and 18.5% reported providing care for other loved ones. See Table 1 for additional demographic information.

Table 1.

Participant demographics and descriptive statistics (N= 27).

	M (SD)			M (SD)
Age	53.3 (5.2)		Steps per day	6344 (3854.74)
BMI	32.6 (6.7)		Active minutes per day	36.50 (29.42)
	n (%)		n (%)		n (%)
Racial Identification		Household Income		Menopause Status
White	16 (59.3%)	<$25,000	4 (14.8%)	Pre-menopause	3 (11.1%)
Black	9 (33.3%)	$25,000–550,000	5 (18.5%)	Perimenopause	2 (7.4%)
East Asian	1 (3.7%)	$50,000–575,000	5 (18.5%)	Post-menopause	16 (59.3%)
American Indian	1 (3.7%)	> $75,000	13 (48.2%)	Other	6 (22.2%)
Ethnic Identity		CVD Risk Conditions		Have Children
Hispanic	1 (3.7%)	Hypertension	17 (63%)	Yes	19 (70.4%)
Non-Hispanic	26 (96.3%)	High cholesterol	14 (52%)	No	8 (29.6%)
Marital Status		Type 2 Diabetes	10 (37%)	Provide Childcare
Never Married	8 (29.6%)	Metabolic Syndrome	2 (7%)	Yes	12 (44.4%)
Married	10 (37%)	Current Smoker	1 (4%)	No	15 (55.6%)
Divorced	6 (22.2%)	Other Health Conditions		Provide Other Care
Separated	1 3.7%)	Yes	13 (51.9%)	Yes	5 (18.5%)
Widowed	2 (7.4%)	No	14 (48.1%)	No	22 (81.5%)
Highest Educational Level		BMI Category		Work Status
Associates’ degree, partial college, or technical degree	5 (18.5%)	Healthy Weight	4 (14.8%)	Working	13 (48.6%)
Bachelor’s degree	13 (48.2%)	Overweight	6 (22.2%)	Not working (retired, unemployed, or other)	14 (52.4%)
Graduate degree	9 (33.3%)	Obese	17 (63%)

Procedures

Procedures were approved by the relevant Institutional Review Board and all participants provided written informed consent, including separate documentation of consent to record exit interviews. Women who expressed interest in participating were asked to complete an electronic survey and to schedule an initial orientation session with research staff. Those who requested to use a study-owned pedometer for recording their PA behaviour each day were sent a device via U.S. mail, to arrive prior to their orientation session (n=15; 56%). The orientation session was conducted via Zoom; participants were introduced to study procedures and were invited to ask questions. They also received a written list of instructions and frequently asked questions for reference during the study period.

For the following 7 consecutive days, participants engaged in 3 tasks per day. First, they wore the study-issued pedometer or their own activity monitor during all waking hours to capture their steps and minutes of moderate-to-vigorous-intensity activity (MVPA, or exercise/active minutes). On days of observation (which included web application use at the start of each day), the average participant took 6344 steps per day (SD = 3854.74) and engaged in 36.50 min of MVPA per day (SD=29.42). However, PA behaviour varied considerably, both between and within participants; for example, average steps per day per person ranged from 1438–15116, and daily steps ranged from 8260 steps below to 10450 steps above a person’s individual average. Second, participants used the study-specific PA web application in the morning before beginning their work or other activities for the day. Web application use was intended to set a positive tone for PA each day and provide an opportunity for participants to set PA intentions and receive support for follow-through, as well as to consider their priorities for potential PA partners. At the end of 7 days of data collection, each participant engaged in a semi-structured qualitative interview via Zoom. Transcription occurred automatically in the Zoom platform; both the interviewer and a research assistant not involved in the interviews reviewed and corrected each transcript. Each transcript was then coded by two trained members of the research team, in alignment with the codebook described below.

Measures

Peer profile selections and responses

Women were asked to use a new, proprietary web application each morning for 7 consecutive days prior to beginning their usual daily activities. During each use, women were prompted to indicate the kind of profile of another woman in midlife they would like to view, with the following options: a Very active, Somewhat active, or Not very active peer, or they could indicate No preference—choose for me. If they selected No preference, the web application displayed the category of peer profile that the participant viewed least recently. After selection, a profile of a fictional woman was displayed and presented alongside an image. The peer profile described the women’s average PA behaviour during the past week (in steps per day and minutes of MVPA), employment status, any current caregiving responsibilities, favourite ways to stay active, and biggest challenge to PA (see Figure 1 for an example).

Figure 1.

Example of a peer profile shown on the web app, selected from Very active, Somewhat active, Not very active, and No preference – choose for me options.

Peer profiles were personalised to depict a woman with similar demographic characteristics to the participant (i.e. age and racial/ethnic background); they also adapted to present the peer’s PA behaviour each day based on the participant’s PA from the previous day, so that the peer’s PA was never too distant from that of the participant (Arigo et al., 2022, 2023). The web application recorded the category of peer profile that the participant selected each day and their perception of the peer’s PA behaviour relative to their own (i.e. How was this person doing with their physical activity, compared to you?). For the latter question, response options were: Seemed to be doing much worse than I am, Seemed to be doing a little worse than I am, Seemed to be doing about the same as I am, Seemed to be doing a little better than I am, and Seemed to be doing much better than I am. They were then asked whether they would want the peer described in the profile as a PA partner (described below as potential partner desirability), and chose from the options Yes, No, and Maybe.

Next, women were asked to indicate what type of PA message they would like to see out of the following options: Encouragement to help me stay on track, Tips for being more active, Holding myself accountable, and No preference - choose for me. These messages were guided by the PA Messaging Framework (Williamson et al., 2021) and categories were selected to represent different types of social support (i.e. emotional, informational, and accountability, respectively). Messages encouraged the adoption of a broad definition of PA that focuses on overall movement as cardioprotective (Ahmadi et al., 2023; Paluch et al., 2021), rather than focusing exclusively on structured exercise (see Figure 2). Finally, women completed their web app use each day by setting PA intentions for the day in an open-response text box: Please describe your plan for reaching your activity goals today. How will you get your steps and/or active minutes today? The present report does not include information about message selections or responses. We include their description here to provide the complete context of women’s experiences with the web application, as these may have influenced their narratives as expressed in qualitative interviews.

Figure 2.

Example of a support message shown on the web app, selected from Encouragement, Tips, Accountability, and No preference – choose for me options.

Exit interview responses

At the end of each participant’s 7 days of web app use, trained researchers conducted semi-structured exit interviews via Zoom to capture women’s perspectives on the web application and potential PA partnerships. Interviewers followed a guide that enabled flexibility based on participants’ accounts of their experiences. Interview questions were initially drafted by the senior (last) author based on a narrative approach to the present research questions, guidance from transactional goal dynamics theory and social comparison theory, and their past experiences conducting research with the population of interest (Arigo, Mogle et al., 2020; Arigo, Brown et al., 2020; Arigo, 2023; Arigo et al., 2014). The fourth author suggested revisions to initial questions to facilitate information gathering; feedback was based on their considerable experience conducting qualitative interviews (Lassiter & Fulton, 2024; Lassiter & Mims, 2022; Denis et al., 2025).

Questions focused on eliciting general preferences for a PA partner (e.g. “What qualities would be most helpful in a partner to support your PA engagement?”) and the perceived importance of specific similarities between themselves and their partner, including age and racial/ethnic/cultural background. Women were also encouraged to identify any additional experiences or characteristics relevant to their partner preferences and the anticipated impact of a partner on their own PA. We used elicitation as part of this process: participants were shown the peer profile they saw on the web application most recently, and were asked to describe their response to this peer profile (e.g. whether they would want her as a PA partner) and the reason for their response. At the end of the interview, they were asked whether they would be interested in joining a PA program that involved being paired with a partner. Assessing the desirability of potential partners, and of a partner-based approach to PA promotion overall, was selected as a straightforward way to assess the feasibility and acceptability of a future intervention among those likely to participate (i.e. those who have some interest in increasing PA; Czajkowski et al., 2015). Interviews averaged 50.41 min in length (SD=7.08) and ranged from 34.01–60.03 minutes.

Positionality statement

Our qualitative coding team included six members with a diverse range of backgrounds, which were considered as potential influences on team members’ perceptions of qualitative data. Specifically, the team included two members who identified as men, two who identified as Black, two who identified as white, one who identified as South Asian, and one who identified as Asian-Australian (who was raised outside of the US). With respect to areas of expertise or focus, three members specialised in women’s health, whereas three did not. In team meetings, we discussed similarities and differences in our interpretations of participants’ statements and the lenses through which these interpretations were filtered (e.g. cultural and gender differences in use of language). Our discussions identified additional depth, which is reflected in the interpretations below, but revealed no meaningful disagreement in use of codes or recognition of patterns. Notably, although the team represented a range of personal identities and backgrounds, all team members were highly educated and belonged to the same professional discipline: two were licenced clinical psychologists working in academia and four were doctoral trainees in clinical psychology. We discussed the context of our disciplinary and training experiences as influences on our coding and interpretation; for example, we acknowledged the possibility of over-interpreting participant statements to reflect avoidance of anxiety or self-presentation bias and discussed alternative explanations (where relevant). We return to these perspectives and their potential influence below.

Analytic approach

We first provide context for women’s narrations of their perspectives on PA partners by describing the peer profiles they selected, saw, and rated as preferred over 7 days. With respect to these quantitative data, we used frequencies to describe the number and proportion of days on which women chose to view profiles of highly active, somewhat active, and not very active peers, as well as the proportion of days on which they elected to let the system choose for them. We also report the proportions of each type of peer profile viewed and the proportions indicated as Yes, No, and Maybe with respect to the desirability of profiled women as PA partners. We describe the desirability ratings of the peers profiled, based on participants’ comparisons of PA to their own PA behaviour, and we report the results of Chi-square (χ²) tests to indicate whether proportions indicated as Yes, No, and Maybe meaningfully differed by type of peer profile selected or viewed, or by perception of the peer’s PA relative to the participant. Finally, we report the number of participants who expressed interest in joining a PA program that involved being paired with a partner during their qualitative interviews.

We then describe the results of reflexive thematic analysis using a semantic approach to generate insights from the qualitative data (Braun & Clarke, 2021). This included collaborative development of a codebook using a combined inductive and deductive process. Specifically, the senior author (principal investigator) developed an initial list of a priori codes, which were aspects of participants’ narratives that were expected to be relevant based on the research questions as well as past experience with this population. The research team added codes following the identification of new meaning units in transcripts and agreement by the research team that these represented distinct information that was relevant to the present aims (i.e. emergent codes). This process occurred over several months and included individual coding with ongoing discussion to support reflexivity and data integrity.

The first 10 transcripts were coded by each team member and coding discrepancies were resolved through discussion. The remaining transcripts were each coded by two members of the research team and discrepancies were resolved by the senior author. Team discussions were also used to reflect on the influence of coder’s identities and professional experiences on their perceptions of the data, including potential biases and these differences among coders, which also informed final coding decisions. Final versions of coded transcripts were reviewed by the senior author to inform the generation of themes that represented deeper patterns of meaning relevant to the initial research questions (i.e. specific to the feasibility and acceptability of the full web application, not reported here).

For the present report, the first and senior author independently identified themes relevant to PA partner preferences. We refined these themes iteratively with input from the fourth author, which required review of candidate quotations and identification of connections among meaning units. Below we describe themes identified through our qualitative analysis of interview transcripts. We also specified endorsement of each theme at the participant level (cf. Braun & Clarke, 2021), though we did not identify additional noteworthy patterns through this process. Exemplar quotations appear in the text and additional illustrative quotations are presented in Table 2, for each theme and its definition. We also provide examples from elicitation components of the qualitative interview (i.e. showing participants the peer profile they saw most recently) and a thematic map, linking our themes to one another. Finally, we summarise findings by integrating across data sources and identifying broader conclusions.

Table 2.

Themes, definitions, and exemplar quotations for each theme from qualitative Interviews (N = 27).

Theme Name	Definition	Example Quotations
(In)flexibility of beliefs about PA	Adoption of either a rigid perspective that physical activity must include a particular type of exercise, intensity level, or structured routine in order for the activity to be considered effective and meaningful, or embracing a flexible approach to activity (e.g., via exposure to a range of peers on the web application)	“She’s a biker so I thought she’s probably a little more serious than me… I found bikers to be pretty serious, physically fit people…they take it seriously. She’s probably a little more dedicated than I can be…so I gave her a maybe.” (age 62, white) “It gives you an idea of an activity you didn’t think about doing before that you might want to try to incorporate.” (age 50, white) “Moving is what’s important and if I think it’s exercise, oh my God, it might scare me. But movement, using my heart, you know hiking, making steps, that’s fun.” (age 54, white) “I appreciated seeing people who put in reasonable amounts of time in a very busy life in lots of different circumstances. And I appreciated the message that something is better than nothing.” (age 62, white) “Even if I just get up from my desk for five minutes, walk around the building for that amount of time and then come back, is better than not doing anything at all, so I guess it takes the pressure of feeling like you have to have a total block of time taken out of your day [for physical activity].” (age 51, white) “[The messages gave] a fun fact like, did you know that just 1000 extra steps today can do this or that? Those are things that kind of stick with you…it’s an encouraging thing and it supports what you’re doing. Or [the message would] say, ‘did you know that just lOminutes of exercise could lower your glucose by this much’…it’s just something that’s reinforcing what you’re doing.” (age 58, white) “[The messages] showed me that any movement is good movement and it doesn’t need to be for an extended period of time, even though an extended period of time is better for you, but you can get in a minute here a minute there, and at the end of the day it all adds up.” (age 52, Black)
Tension of caregiving as a value and a barrier to PA	Commitment to caring for family (children, grandchildren, parents) serves as both a motivator for staying healthy and an unpredictable time commitment that results in difficulty finding time or energy for PA (desire for sharing PA efforts with others who understand/can relate to this, used profile information to contextualize)	“My whole life I’ve been raised - well for the most part, I don’t even like to utter this but, I’ve been like, really struggling with caring for myself. Like always putting other people first. I’ve been a caregiver and a pleaser and all that crap for years. ” (age 54, white) “Yeah I don’t prioritise myself, you know but it’s not because I don’t think I’m important, it’s just you know when you’ve got kids and grandkids you’re not number one in the queue. You can’t be.” (age 62, white) “I did look on each of [the profiles], I looked at everything each of them said, but when they had caregiving on there that one touches me - I get it with the caregiving, how that system is.” (age 58, white) “[Most] women have so many things on their plate and…they take a back seat. I’m married and when my husband used to work out, nothing would get in his way…He would make time and he would stick to it. And then me, it would be like somebody gotta go to soccer, I guess I’m the person.” (age 57, Black) “I have a special needs child. I’m working with this population of women and a couple of support groups who do not put themselves first at all, and you’re kind of left with, okay if you don’t take care of yourself there’s no one left to take care of your kid.” (age 58, Latina)
Desire to feel validated and understood SUB: Preference for similar PA goals and level of commitment SUB: Seeking and reflecting a sense of connection based on experiences unique to women	Preference for similarity in specific personality characteristics, support styles, and life experiences as relevant to PA SUB: Preference for partners who want to reach a similar goal/place/type of activity, bring similar energy to the process (not too little or too much focus on PA change), and has similar time for PA and the partnership (seen as most compatible and most likely to result in mutual support) SUB: Use of “we” statements when describing barriers to PA; a desire to feel less alone in their struggles (often related to how women are socialised, such as difficulties prioritising themselves); desire for sharing PA efforts with other women who could understand their experiences of ageing, menopause, motherhood, and caregiving	“For me it really comes back to the physical activity level and then just enough lifestyle similarities, that we can at least understand.” (age 48, white) “I just felt like [the most recent profile] had the appropriate balance. Like if I wanted to go for a walk with somebody let’s say every day. I wouldn’t want somebody that at least 85% of the time could do it. But I would expect that they’re going to be things that are going to get in the way or you know it’s not going to be like gee if I don’t do this 100% of the time I’m not worthy.” (age 62, white) “It was good to know that even though we know that physical activity needs to be a part of our lives, sometimes it just doesn’t happen because life gets in the way. You know, you’re not the only one that sometimes can’t do what needs to be done because of life.” (age 52, Black) “I only have one child, and I always make her my priority like if I’m supposed to do something, and she wants to do something, I drop whatever I’m doing and I go do it for her … I think that some people might not understand that.” (age 50, white) “I’m thinking about all the pictures they put on advertisements for new gyms. They’re all like little skinny tiny heroes. What person who is, you know, overweight and oh yeah, I want to go hang out with them. Especially women, we beat ourselves up about everything. You want someone like you as inspiration.” (age 58, Latina) “Life happens, stuff happens, we’re not perfect, we’re all trying to make progress. I’d want somebody that understands that. [But not someone] so lenient that every day, you can cheat…Maybe the [partners] would do something like talk out self-talk. Because women need that, women need to help each other. Too many times, people are so nasty and I just don’t like to be with people like that. Women need support and they need to be kind to each other and that’s really important.” (age 54, white) “[With this peer] okay [if] we’re gonna work out together, let’s go jump rope together, you know or something like that then maybe we…can we work out together or something virtually or meetup. [Because] look, we will have to motivate each other.” (age 56, Black) “The closer we are in age, we probably do experience some of the same things… menopause, kids. Our kids are grown now, and we’re dealing with grandkids. So that would be great…the more we have in common, the better the better communication we would have.” (age 57, Black)
Openness to many types of potential partners	Willingness to pair with a range of partners with respect to the partner’s background, lifestyle, commitments; little priority placed on race/ethnicity/culture in favour of factors such as menopause and caregiving (with some heterogeneity)	“Everybody needs exercise, and we all want to keep our hearts well and we all want to have our mental acuity for as long as we possibly can have it.” (age 51, white) “There’s beauty in diversity.” (age 45, white) “I would look more for somebody that could challenge me, you know, and if there’s a seven year old male or female, that would be able to do that white, black, purple, whatever colour they were wouldn’t matter to me in the least.” (age 50, white) “You learn more, you’re exposed to more different things, you see things from a different view.” (age 56, white and Cherokee) “I wouldn’t not have a person that’s younger than me really teach me something…We all are going to experience certain things, it has really nothing to do with our race or anything, women we’re going to have menopause … for me that’s not an important issue, I can, like I said, have friends of every race so that’s not a big issue.” (age 57, Black) “If anything I’d really enjoy somebody with a different background, just because then the conversations are very interesting.” (age 62, white) “A partner who’s different than me would definitely be helpful for me. Because if we’re exactly the same way I’m not doing something different, I guess. You know what I mean? Because I can learn from what motivates somebody else and their conversation or their, you know, their personality.” (age 51, Black) “Having similar background in terms of racial and ethnic identity, it can be helpful for similarities as well, so that sounds like that’s an important piece.” (age 59, Black) “Yes, I would prefer Black women because I think that we share something in terms of frustrations and challenges that I don’t think my white female counterparts could really understand…Even though Black people are not monolithic either, there are some similarities in terms of even diseases. What I feel is that the medical field just lumps us in that if they’re Black you’re gonna have these challenges and we’re kind of ignored.”(age 52, Black)
Motivation from a PA partnership that fosters growth and provides opportunities for behavioural modelling	Desire to connect with partners who inspire self-reflection and progress, including individuals who can model positive physical activity behaviours as well as recognising the motivational impact of being a physical activity role model for others	“I found that people on my level or higher than my level, I would be more encouraged to have as a partner.” (age 55, white) “I would rather aspire up [with someone doing more PA than I am] and surround myself with up to motivate myself.” (age 53, white) “I would need somebody that would have the ability to open my eyes, maybe shock me a little bit, just tell me like it is. If you’re too soft with me I would probably take advantage of that… if you were my partner and you said ’hey you told me you wanted to reach these goals and you’re not going to get there if you keep, you know, making excuses.  Yeah you’re tired. I’m tired, but I got to the gym’ that would work better for me.” (age 50, white) “You wanted to see that you could be a little bit ahead of them, because that would be most motivating to you and you could provide some encouragement to them.” (age 40, white) “Her steps are a little bit low… So for me it’s like Okay, we could partner, and then it would be I’m pushing her, and by pushing her I’d be pushing myself.” (age 51, Black) “[I want] Someone that’s not a complainer, someone that is going to take the position seriously. Somebody that’s going to offer me advice, and someone that is going to have time for me… It was good to know that even though we know that physical activity needs to be a part of our lives, sometimes it just doesn’t happen because life gets in the way. You know, you’re not the only one that sometimes can’t do what needs to be done because of life.” (age 52, Black) “Part of my biggest thing is, you know if I have somebody that’s going to say ’oh let’s skip it today’ - I am going to be that person … I need somebody who’s you know, even if they only get maybe 4000 steps in a day, if they say they walk every day, like once a day, then that’s the person I would say yep I want that person. Because I want to know that no matter what, they’re gonna walk.” (age 50, white) “[One of my friends is] very motivating for me. And when we walk together we don’t take breaks. We hike. And we’re at a good pace, we’re not just doing that kind mall shuffle. So I think somebody that is very active themselves [would be a good partner].” (age 58, Latina) “Somebody that is patient and maybe somebody that wont give up on me even if I try to give up.” (age 54, Black)

Results

Quantitative: peer profile selections and desirability of peers as PA partners

Of the peer profile categories available, participants selected No preference—choose for me most often (35.8% of selections across participants and days), followed by Somewhat active peer profiles (28.4% of selections; see Figure 3, panel A). Not very active peer profiles were chosen least often (13.6% of selections). These selections resulted in participants viewing each type of peer profile (highly active, somewhat active, not very active) on approximately one third of days overall (see Figure 3, panel A), indicating that the web application worked as intended to distribute No preference selections to the category seen least recently. Participants perceived peers’ PA as better (more) than their own for the majority of peer profiles viewed (50.8%), whereas they perceived only a small minority of peers to have worse (less) PA than they did (10.3%).

Figure 3.

Peer profile categories selected/viewed (panel a) and interest in each type of peer as a physical activity (Pa) partner by profile category viewed and perception of peer’s Pa relative to the participants’ own Pa (panel B).

With respect to the desirability of the peers described in profiles as potential PA partners, across selections and days, participants most often indicated Maybe (48.0%) or Yes (29.7%) to the peer profiles they viewed. Preferences did not meaningfully differ by the category of peer profile selected (χ²[6] = 11.91) or of the peer profile viewed on the web application (χ²[4] = 2.59, ps > 0.06). However, the desirability of a potential partner depended on their perception of how well the peer was doing with PA, relative to the participants’ own PA (χ²[10] = 22.18, p = 0.01). As shown in Figure 3 (panel B), participants preferred to partner with peers at their own PA level or those they perceived as doing a little better than they were with PA. In contrast, they never indicated Yes to partnering with a peer who they perceived as doing a little worse than they were with PA. They occasionally indicated Maybe for these peers and indicated Yes for peers who they perceived as doing much worse than they were, as well as much better. Finally, the majority of participants expressed interest in joining a PA program that involved being paired with a partner (26/27; 96.30%).

Qualitative: physical activity beliefs and preferences for partners

We generated five themes, which reflected two overarching sets of beliefs or experiences that informed a particular social desire (i.e. to be validated and understood). This desire linked broader perspectives with those specific to PA partner preferences. The latter encompassed two specific sets of perspectives (i.e. openness to many types of partners, desire for motivation and growth from a partnership). Each (sub)theme, its definition, and illustrative quotations appear in Table 2, and Figure 4 presents a thematic map linking themes and subthemes.

Figure 4.

Thematic map of influences on preferences for PA partnerships among women in midlife with elevated risk for CVD.

The first theme, (in)flexibility of beliefs about PA, captures participants’ definitions of PA, their assumptions about what does and does not “count” as healthy (i.e. cardioprotective) PA, and the degree of openness they expressed towards different types of PA. Several participants expressed a rigid perspective, viewing PA as requiring structured exercise of at least moderate intensity, and a sense of relief to learn that this is not the only way to stay healthy (i.e. via peer profiles and supportive messages on the web application). They noted that the web application encouraged thinking flexibly about PA as movement and how it could fit into their hectic and unpredictable lives. For example, participants shared:

• “Moving is what’s important and if I think it’s exercise, oh my God, it might scare me. But movement, using my heart, you know hiking, making steps, that’s fun.” (age 54, white)

• “[The messages] showed me that any movement is good movement and it doesn’t need to be for an extended period of time, even though an extended period of time is better for you, but you can get in a minute here a minute there, and at the end of the day it all adds up.” (age 52, Black)

• “I appreciated seeing people who put in reasonable amounts of time in a very busy life in lots of different circumstances. And I appreciated the message that something is better than nothing.” (age 62, white)

• “[The messages gave] a fun fact like, did you know that just 1000 extra steps today can do this or that? Those are things that kind of stick with you…it’s an encouraging thing and it supports what you’re doing. Or [the message would] say, ‘did you know that just 10 minutes of exercise could lower your glucose by this much’…it’s just something that’s reinforcing what you’re doing.” (age 58, white)

Several participants referenced that their beliefs about PA influenced preferences for potential partners, with respect to specific peer profiles on the web application. For example, in response to viewing the peer profile she saw last on the web application, one participant shared her focus on the activity this person enjoyed and how that informed her decision to choose this individual as a partner:

• “She’s a biker so I thought she’s probably a little more serious than me… I found bikers to be pretty serious, physically fit people…they take it seriously. She’s probably a little more dedicated than I can be…so I gave her a maybe.” (age 62, white)

Another participant expressed appreciation for the opportunity to see multiple approaches to PA illustrated in the peer profiles:

• “It gives you an idea of an activity you didn’t think about doing before that you might want to try to incorporate.” (age 50, white)

A second theme relevant to general perceptions of PA was tension of caregiving as a value and a barrier to PA (see Table 2). Participants frequently referenced caregiving for family members (i.e. children, grandchildren, parents) as both a motivator for committing to PA and as a barrier that impedes their ability to maintain routines or muster the energy for PA. One participant described these challenges:

• “[Most] women have so many things on their plate and…they take a back seat. I’m married and when my husband used to work out, nothing would get in his way…He would make time and he would stick to it. And then me, it would be like somebody gotta go to soccer, I guess I’m the person.” (age 57, Black)

Some participants found a sense of camaraderie when viewing potential partner profiles on the web app that mentioned similar responsibilities. For example:

• “I did look on each of [the peer profiles], I looked at everything each of them said, but when they had caregiving on there that one touches me – I get it with the caregiving, how that system is.” (age 58, White)

Others viewed caregiving as a strong value and motivating factor for engaging in PA, recognising the potential importance and benefits of increasing PA while providing for others. Further, although participants often viewed making time for PA as at odds with what caregiving requires (i.e. prioritising others over the self), a subset recognised the importance of PA and how it can benefit their caregiving. For example, one participant offered:

• “I have a special needs child. I’m working with this population of women and a couple of support groups who do not put themselves first at all, and you’re kind of left with, okay if you don’t take care of yourself there’s no one left to take care of your kid.” (age 58, white)

As noted, we identified a third theme that encompassed both women’s broad experiences and their specific PA partner preferences: participants’ desire to feel validated and understood (see Table 2). Several participants expressed a preference for partners with similar characteristics, support systems, or life experiences that made them feel seen. For instance, participants explained,

• “For me it really comes back to the physical activity level and then just enough lifestyle similarities, that we can at least understand.” (age 48, white)

• “I only have one child, and I always make her my priority – like if I’m supposed to do something, and she wants to do something, I drop whatever I’m doing and I go do it for her … I think that some people might not understand that.” (age 50, white)

Under the theme of validation, we identified two subthemes: (1) seeking and reflecting a sense of connection based on experiences unique to women, and (2) preferences for partners with similar PA goals and level of commitment.

For the first subtheme, some participants expressed a desire to feel understood based on shared life experiences related to being a woman. Specific points of potential connection were how the ageing process affects women, the experiences of menopause or motherhood, and caregiving as a gendered responsibility. These experiences demonstrated a collective desire to feel less alone in their struggles. This was noted in statements such as:

• “The closer we are in age, we probably do experience some of the same things…menopause, kids. Our kids are grown now, and we’re dealing with grandkids. So that would be great…the more we have in common, the better the better communication we would have.” (age 57, Black)

• “I’m thinking about all the pictures they put on advertisements for new gyms. They’re all like little skinny tiny heroes. What person who is, you know, overweight and oh yeah, I want to go hang out with them. Especially women, we beat ourselves up about everything. You want someone like you as inspiration.” (age 58, white)

For the second subtheme, a subset of participants highlighted a desire for partners who were working towards goals that aligned with their own. Participants valued partners who could invest similar amounts of effort and energy into PA progress, and who have the same amount of time available to focus on improving their PA and investing in the PA partnership. This mutual commitment was seen as crucial not only for the partnership, but to how participants perceived themselves. As one participant described in response to seeing the profile of the last peer she viewed on the web app:

• “[With this peer profile] okay [if] we’re gonna work out together, let’s go jump rope together, you know or something like that then maybe we…can we work out together or something virtually or meetup. [Because] look, we will have to motivate each other.” (age 51, Black)

With respect to the particular desire for validation through the partnership and what came through in the peer profiles, another participant offered:

• “[I want] someone that’s not a complainer, someone that is going to take the position seriously. Somebody that’s going to offer me advice, and someone that is going to have time for me… It was good to know that even though we know that physical activity needs to be a part of our lives, sometimes it just doesn’t happen because life gets in the way. You know, you’re not the only one that sometimes can’t do what needs to be done because of life.” (age 52, Black)

The theme of openness to many types of potential partners was evident in participants’ discussion of their willingness to be paired with partners of different diverse backgrounds and lifestyles. One participant noted what partnering with someone different could offer:

• “You learn more, you’re exposed to more different things, you see things from a different view.” (age 40, biracial – Cherokee and white)

Another expressed a similar statement, noting:

• “Everybody needs exercise, and we all want to keep our hearts well and we all want to have our mental acuity for as long as we possibly can have it.” (age 51, White)

Overall, shared experiences such as menopause and caregiving held higher importance than demographic factors of age, race, and ethnicity among this sample of participants. However, some individuals highlighted preferred to partner with someone who shared more specific characteristics:

• “Having similar background in terms of racial and ethnic identity, it can be helpful for similarities as well, so that sounds like that’s an important piece.” (age 59, Black)

One participant noted that her preference for a PA partner stemmed from her CVD risk, and she desired a partner who understood how race and ethnicity played a role in the medical care she received:

• “Yes, I would prefer Black women because I think that we share something in terms of frustrations and challenges that I don’t think my white female counterparts could really understand…Even though Black people are not monolithic either, there are some similarities in terms of even diseases. What I feel is that the medical field just lumps us in that if they’re Black you’re gonna have these challenges and we’re kind of ignored.” (age 52, Black)

Our final theme captured the expected motivation to be gained from partnerships and the partnership factors that participants believed would positively influence their PA: Motivation from a PA partnership that fosters growth and provides opportunities for behavioural modelling. Participants indicated a desire for connections that inspire self-reflection and progress, as well as a partner who can model positive behaviours. Participants also highlighted the motivating impact of being a role model for others, nothing that being in this position may not only encourage their partner’s engagement in activity but also motivate their own:

• “I would rather aspire up [with someone doing more PA than I am]… and surround myself with up to motivate myself.” (age 53, white)

• “Her steps are a little bit low… so for me it’s like Okay, we could partner, and then it would be I’m pushing her, and by pushing her I’d be pushing myself.” (age 51, Black)

• “I would need somebody that would have the ability to open my eyes, maybe shock me a little bit, just tell me like it is. If you’re too soft with me, I would probably take advantage of that… if you were my partner and you said, ‘hey, you told me you wanted to reach these goals, and you’re not going to get there if you keep, you know, making excuses. Yeah, you’re tired. I’m tired, but I got to the gym’…that would work better for me.” (age 50, white)

Summary and integration of findings

As noted, Figure 4 displays a thematic map linking the themes and subthemes we identified. These reflected two overarching experiences related to PA (i.e. [in]flexibility of beliefs about PA and tension of caregiving) that we saw as informing the complex theme of a desire to feel validated and understood. This desire was a lens through which women saw their preferences for PA partners: although they were open to many types of partners (with some exceptions), they saw partnerships as opportunities for motivation and growth and preferred partners who they perceived as able to provide these opportunities. In tandem with quantitative findings, the partners perceived as most able in this respect were peers who were at the same level or doing a little better with PA than they were. Peers doing a little worse than they were with PA were least likely to be preferred, though some women were open to testing a partnership with someone in this position (as motivating their partner could also motivate them).

Discussion

Social connection is key to facilitating PA among women in midlife who are at elevated risk for CVD (Lum & Simpson, 2021), though many women lack this connection in their existing relationships (Im et al., 2011). Available intervention programs for women in midlife often purport to increase social support and related processes as mechanisms for promoting PA, though the efficacy of existing approaches is questionable. Most programs seem to assume that support and processes such as positive behavioural modelling or comparison will happen if the program is delivered in groups, and there is little emphasis on intentional and effective activation of these processes (Arigo, Romano, et al., 2022). Many women “fall through the cracks” if they do not feel a sense of connection and accountability to others in the group, which can undermine behaviour change efforts—particularly if connection was a motivator for joining the program (cf. Dutton et al., 2014; Izumi et al., 2015).

Facilitating PA partnerships between women who sign up for tailored PA promotion programs is a promising approach to increase the effectiveness of efforts to activate social support and other processes in PA promotion programs (Arigo, Schumacher, et al., 2015; Murray et al., 2023). In partner relationships, each participant is directly accountable to and responsible for communicating with another participant, and partners can be available for support and problem-solving between intervention sessions (Arigo, 2015; Arigo et al., in press). As a result, women’s social needs are more likely to be met. For intervention leaders and health coaches, addressing communication or support challenges in partner dyads may be more straightforward than balancing the needs of multiple women in groups. Yet, there is little evidence available to inform optimal matching of partners—particularly women’s perspectives and preferences for PA partnerships, which could affect their efficacy. In the present study, we used a proprietary web application to offer women in midlife with elevated CVD risk the opportunity to select and read profiles of potential PA partners each day for 7 consecutive days. At the end of this period, they engaged in an in-depth qualitative interview about their experiences and overall preferences for PA partners.

Findings from the present study indicate that PA partnerships with other women may be highly desirable to women in midlife with elevated risk for CVD. Women who shared their perspectives in this study viewed these partnerships as opportunities for mutual support to overcome common barriers to PA, such as lack of motivation and difficulty managing competing responsibilities (cf. Im et al., 2011). Quantitative and qualitative findings showed that women in this study were open to a range of potential PA partners and specific preferences for these partners. For example, women most often selected No preference—choose for me with respect to the type of peer profile they wanted to see each day (35.8% of daily selections), and indicated maybe or yes they would want to partner with the woman profiled for the vast majority of potential partners viewed (77.7% of peer profiles). This may have been due to uncertainty or hesitation to reject opportunities, as we asked about a hypothetical situation only. However, women generally described being open-minded about the extent to which their partner shared various aspects of identity, if they endorsed similar PA- and health-related challenges (e.g. menopause, ageing) specific to women in midlife. These attitudes were informed by perceptions that women with different life experiences may provide interesting perspectives or complementary skills and attributes that promote connection or learning (Friedman et al., 2016; Gomez & Bernet, 2019).

Yet not all potential partners were considered equally desirable, as women indicated the peers with whom they could not imagine a beneficial relationship. In interviews, participants described the most suitable partners as those who could inspire and motivate them to increase PA; these were most often peers who were similar to or slightly above their own level of PA. These peers were seen as having useful insight, as they were at similar starting points with PA, as well as the ability to provide accountability. As noted, these perceptions are consistent with multiple frameworks for understanding health behaviour change, including social comparison theory and the theory of transactive goal dynamics. Social comparison theory (Festinger, 1954) proposes that upward comparison targets (i.e. people who are perceived as doing better than the self in a relevant domain) can provide inspiration and guidance for how to achieve a similar outcome (Bandura, 2013; Derlega et al., 2008), as well as confidence in one’s own ability to achieve that outcome (Buunk & Ybema, 1997), making them desirable PA partners. Lateral comparison targets (i.e. people who are perceived as at the same level as the self in a relevant domain) are also desirable because people are motivated to “keep up” with relevant others (Merchant et al., 2017)—in this case, to keep up with their partner’s PA. Similarity in starting level of PA between partners facilitates the identification of shared goals and the provision of mutual support for goal achievement, creating a transactive (and therefore, effective) process of behaviour change (Fitzsimons et al., 2015; Miller & Stiver, 2015).

Importantly, however, a few features of potential PA partnerships offered caveats to these preferences: the distance between one’s own PA and that of their potential partner, and whether their partner could understand their challenges. Women perceived that peers who already achieve much more PA than they do may have fewer competing priorities, and thus, may be less equipped to empathise with their own PA barriers and provide effective support. These more extreme upward targets often prompted self-criticism and defensiveness and were perceived as alienating. Thus, whereas peers doing a little bit better led to upward identification (i.e. seeing oneself as similar to someone doing well and able to achieve a similar outcome), peers doing a lot better led to upward contrast (i.e. seeing oneself as different from someone doing well and highlighting one’s own inferiority; Buunk & Ybema, 2013). Although either could be motivating depending on the person and context, upward identification (vs. contrast) has shown stronger positive associations with motivation for healthy behaviour among adults with conditions such as type 2 diabetes (Arigo, Smyth, et al., 2015).

There were two additional exceptions to openness with respect to potential PA partners, which were evident from the combined quantitative and qualitative findings. One was for peers who they perceived as doing a little bit worse than they were with PA; these peer profiles were never strongly preferred (i.e. Yes response) and received the largest proportion of not preferred (i.e. No response). Women expressed concern that those not yet at their level of PA (i.e. downward comparison targets) may empathise without supporting an increase in PA, reflecting insight into the complexity of social support: although empathy might be desirable and reduce distress, it needs to be paired with accountability, modelling, or another process to promote change (Ginis et al., 2013; Petosa & Smith, 2014). This observation is particularly noteworthy given that peers who were perceived as doing much worse with PA received similar proportions of yes responses to those perceived as at the same level or doing a little better with PA (see Figure 2, panel B). Yet, even with peer profiles perceived as doing a little bit worse than they were, some women indicated openness to partnerships (i.e. Maybe response), as they could feel motivated by motivating their partner to engage in PA.

Second, a subset of participants shared that racial, ethnic, or cultural background is important for their partnership, indicating that similarities in these domains may improve connection through shared experiences. This was particularly salient among women who identifyied as Black, aligning with evidence that women of colour in the U.S. may suffer disproportionately from inadequate medical care and may benefit from relevant social support (Ho et al., 2022; Johnson et al., 2014). Black women in the U.S. may also share crucial, health-related perspectives related to their intersectional experiences; for example, potentially experiencing higher levels of stress combined with pressures to appear emotionally and physically strong (Woods-Giscombé, 2010). Thus, PA partnerships may provide needed opportunities for these women to receive support for specific health-related problems (e.g. validation or problem-solving related to micro-aggressions or more overt discrimination; Pullen et al., 2014; Watson-Singleton, 2017). Partnering with women who do not have these shared experiences may limit opportunities for benefit or contribute to frustrations if women feel invalidated or misunderstood. Further study is required to confirm our findings and determine how PA partnerships can be tailored for the needs of different subgroups of women.

Unique experiences of women in midlife

In interviews, women in this study also expressed interest in working with someone who shares similar experiences related to being a woman in midlife who could understand the challenges of ageing and menopause as they relate to PA motivation and behaviour (cf. Im et al., 2011). This interest was often expressed as a preference for someone who would commit to the PA partner relationship and make adequate time for them, but who would not take PA so seriously that they wouldn’t have empathy for life getting in the way. In particular, women sought a PA partner who could understand the unique difficulty of making time for PA in the context of caregiving. In the U.S., women still bear the brunt of caregiving burdens across the lifespan, and elder care responsibilities increase sharply during midlife (Ice, 2023; Infurna et al., 2020; Pope et al., 2012). Women who provide caregiving continue to participate in work activities at high rates during this time (Ansari-Thomas, 2024). Consequently, even when women in midlife set concrete intentions to engage in PA, caregiving is primary in a set of responsibilities that can interfere with follow-through (Arigo, Hevel, et al., 2022; Hendry et al., 2010).

Indeed, of the 27 women who participated in the present study, 12 provided regular care for children or grandchildren (44.4%; see Table 1), 5 provided unpaid care for other adults (e.g. ageing parents; 18.5%), and 2 provided care for both children and adults (7.4%); all but one of these women were also working in paid positions. Across the full sample, women’s narratives pointed to a tension around caregiving, as both a highly valued activity and a primary barrier to focusing on one’s own health. Some women seemed to describe the view that caregiving taking away from other areas of life is natural and entirely unavoidable. Others recognised that consistently deprioritizing the self is problematic, in part because it can negatively impact caregiving ability (Lu & Wykle, 2007; Sabo & Chin, 2021).

This tension may be strongly related to the inflexibility in beliefs about PA that we observed in the present study. Specifically, women expressed or alluded to assumptions that only moderate-to-vigorous-intensity exercise in discrete episodes is healthy (e.g. hiking, visiting the gym), which is consistent with previous work in this population (Tudor-Locke et al., 2003). As noted, exposure to upward comparison targets with respect to PA has benefits, though it may reinforce this belief if targets or other peers (appear to) engage in only this type of activity (Luszczynska et al., 2004). Critically, although moderate-to-vigorous-intensity exercise has important benefits for cardiovascular health, particularly for women with elevated CVD risk (Smith et al., 2016), considerable evidence shows that light-intensity PA and overall movement are also cardioprotective (Füzéki et al., 2017; Kraus et al., 2019). Lower-intensity activities are less effortful, as they do not require setting aside large blocks of time or finances (e.g. to maintain a gym membership), and thus, may be much easier for women in midlife to integrate into their daily routines while conferring meaningful cardiovascular benefits (Cheval & Boisgontier, 2021). The rigid beliefs we observed demonstrate the persistence of perceptions that are thus at odds with current science about how much and what kind of PA is necessary for health and are potentially maladaptive, as they can promote all-or-nothing thinking and behaviour (e.g. “I don’t have time to get to the gym, so I won’t be active today;” Segar, 2024). Many women in midlife with CVD risk conditions have regular healthcare check-ins (e.g. for medication management) that involve discussion of regular PA as part of protecting their heart health; correcting misperceptions about what “counts” as PA represents a key area of opportunity for improvement in healthcare and allied settings (Arena et al., 2018).

Strengths, limitations, and future directions

This study has several strengths, including a sample size of 27 women who used our web application for 7 days and completed a qualitative interview. As intended, this sample size and mixed methods approach afforded a deep and rich understanding of women’s preferences regarding PA partnerships as well as identification of heterogeneity with respect to multiple themes. Our team and reflexive methods also represented perspectives from researchers with diverse backgrounds and varying levels of expertise in both quantitative and qualitative analyses. As a result, we were able to represent the complexity of qualitative data, as well as to align qualitative and quantitative findings. We also used a purposeful sampling approach to prioritise diversity, leading to enrolment of several individuals from underserved and marginalised backgrounds. However, although we captured a wide range of experiences among women in midlife, certain subgroups (e.g. caregivers without a partner, those with high school education only) were underrepresented or not clearly identified. Further, as our analysis team was homogenous with respect to disciplinary background (i.e. clinical psychology) and education level, we may have under-represented or missed certain perspectives. This is a particular concern given the similar high education level of several participants, and additional work is needed to accurately reflect the perceptions of women in midlife with a wider range of educational experiences.

Several additional, critical points remain unclear and warrant important next steps. By definition, participants saw peers who were more, less, and similarly physically active than they were the previous day (i.e. each peer’s steps and minutes of MVPA were percentages of their own from the day before). This was intentional, to ensure that a peer’s activity was not so much higher or lower than the participant’s own that the information would be perceived as irrelevant (Wood, 1996). However, women’s assessments of a peer’s activity relative to their own is likely based on what they perceive as their own typical or average level of activity (Sallis & Saelens, 2000), rather than only what they did the previous day. As we did not include assessments of participants’ activity behaviour prior to their use of our web application, the extent to which women’s comparisons were accurate is not clear. Notably, we focused on comparison preferences (selections) and perceptions due to considerable evidence that are more powerful for influencing emotions and behaviour than relative assessments based on “objective” information (Alvero, Bucklin, & Austin, 2001).

Whether women’s preferred partners are more effective for supporting their PA than peers who do not meet these criteria is also unknown, and formal testing is needed. The logistics of a process for identifying and optimally matching partners based on their preferences are complex and require careful consideration, particularly if they are to be implemented on a large scale. For example, it is possible that two women who are similar in all preferred respects do not have overlapping availability for intervention sessions or between-session communication. In addition, even if partners are optimally matched at the start of a PA adoption effort, their needs may change as they progress through PA habit formation and maintenance (Rothman et al., 2011), and there is a need to identify the frequency, method(s), and type(s) of communication that are most effective in PA partnerships.

Further, many women in the present study expressed preference for partners doing a bit better than they are with PA. If partnerships are formed between women who enroll in a PA program, it would not be possible to pair every woman in a PA program with someone doing better, as some would have to be doing better than their partner. As many women also expressed an openness to partners who were doing worse, however, it might be possible to take individual preferences into account when pairing (e.g. matching someone who wants a partner doing better with someone who is doing better and is open to a partner who is doing worse). Critically, as the overall goal is to modify women’s PA behaviour, and as women’s PA varies across days and weeks (Brady, Brown, & Mielke, 2023; Kishida & Elavsky, 2016), it is likely that many partner dyads would experience shifts in who is “doing better” during an intervention. Preparation for this situation and coaching in effective support across levels of achievement may help to avoid any competition or negative social comparisons between partners (cf. Arigo et al., 2015; Miller et al., 2023).

Despite the potential logistical challenges of a partner-based approach to PA promotion for women in midlife, the preferences shared in this study represent a strong beginning point for understanding priorities for these women and testing methods for integrating them in partner-based PA promotion programs. Additional work is needed to determine the frequency and qualities of communication that can lead to success, whether and how these change over time, and how best to integrate flexible partner support with more structured professional support (i.e. from health coaches or intervention facilitators). It will be critical to conduct this work with diverse samples and to explore perspectives from different groups of women, including those who have not yet participated in PA-related research studies. Finally, consistent messaging about the benefits of lower-intensity PA and overall movement is needed across the wide range of healthcare experiences for this population, to effectively support small changes in PA that are sustainable in the context of women’s busy and unpredictable lives.

Conclusions

PA partnerships between women who do not have a pre-existing relationship show unique promise for addressing key barriers to PA among women in midlife with CVD risk conditions, as these women often do not have sufficient PA support in their existing social networks. The present study offers important insights into a subset of women’s preferences for PA partners that are informed by their life experiences and their exposures to potential partners via a personalised, adaptive web application. PA partnerships may be desirable to women in this at-risk population and they are likely to prioritise potential partners who could understand their challenges—particularly around ageing, menopause, and caregiving—and provide positive PA role models. Women tend to acknowledge that a range of peers could meet these criteria and respond positively to the idea that PA is more than structured exercise. Additional investigation of methods to more formally assess the feasibility and acceptability of this approach, optimally pair PA partners, facilitate beneficial communication between partners, and support them through PA adoption and maintenance efforts is needed.