Barriers to physical activity among pregnant women in China: a mixed methods study based on the COM-B model and the theoretical domains framework

Abstract

Background

Although physical activity has numerous benefits during pregnancy, most pregnant women remain physically inactive. This study aimed to explore barriers to physical activity among Chinese pregnant women through an integrated application of the Capability, Opportunity, Motivation, and Behavior model and the Theoretical Domains Framework.

Methods

A sequential explanatory mixed methods study was conducted. First, a cross-sectional survey of 400 pregnant women with convenient sampling (December 2023–January 2024) was conducted using the Chinese version of the Barriers to Physical Activity during Pregnancy Scale. Data entered using EpiData version 3.1 were analyzed with SPSS version 27.0, employing descriptive statistics, univariable analysis, and multivariable regression to identify factors associated with barriers to physical activity among pregnant women. Subsequently, semi-structured interviews were conducted with 13 participants purposively sampled from the survey respondents. The interview data were organized using NVivo version 12.0 and analyzed through thematic analysis. Quantitative and qualitative findings were integrated via side-by-side comparison.

Results

The mean barrier score was (79.58 ± 10.97). Univariable analysis revealed that there were statistically significant differences in the scores of trimesters, desired delivery mode, parity, history of abortion, specific physical activity guidance during pregnancy, and pre-pregnancy habitual physical activity (all P < 0.05). Multivariable regression identified desired delivery mode, parity, specific physical activity guidance during pregnancy, and pre-pregnancy habitual physical activity as significant predictors (F = 6.081, P < 0.05), explaining 23.40% of the total variance. Thematic analysis mapped barriers to the Capability, Opportunity, and Motivation components and nine domains of the Theoretical Domains Framework, including skills, knowledge, memory, attention and decision processes, environmental context and resources, social influences, emotion, goals, belief about capability, and beliefs about consequences. Findings from both quantitative and qualitative studies were mapped to the two theoretical frameworks.

Conclusions

This mixed methods study used the Capability, Opportunity, Motivation, and Behavior model and the Theoretical Domains Framework to identify barriers to physical activity among Chinese pregnant women. Future interventions should combine individualized counseling to enhance skills and motivation with family-level initiatives to provide supportive environments and accessible resources, integrating structured physical activity guidance into routine prenatal care.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-026-26418-0.

Introduction

Physical Activity (PA) is defined as any bodily movement produced by skeletal muscle contractions that result in energy expenditure across various domains, including household/caregiving, occupational, sports/exercise, and transportation activities [1]. The World Health Organization (WHO) recommends that all pregnant women without contraindications participate in at least 150 min of PA per week throughout pregnancy [1]. Growing evidence suggests that regular PA is crucial for optimal pregnancy outcomes. For the mother, the benefits include reducing excessive gestational weight gain [2], enhancing cardiorespiratory fitness [3], increasing the likelihood of vaginal delivery [4], and improving sleep quality [5]. For the fetus, regular PA reduces the incidence of preterm birth [6] and improves neurological development [7]. Conversely, inadequate PA during pregnancy can increase the risk of several pregnancy-related complications, such as obesity, gestational hypertension, and gestational diabetes [8, 9], which can adversely affect both maternal and fetal health.

Although regular PA has several benefits, most pregnant women remain physically inactive during pregnancy and do not meet the WHO recommendation. Studies conducted in several countries revealed that 7% to 21% of pregnant women achieved the goal of 150 min or more [10–12]. A study in China indicated that only 2.8% of 1636 pregnant women met the guideline recommendations for PA throughout their pregnancy [13]. Thus, an urgent challenge is how to effectively promote daily PA among pregnant women. Since inadequate PA is partly due to various barriers [14], it is essential to identify and address these obstacles to increase activity level.

Pregnant women encounter many barriers to PA during pregnancy, which can be categorized into intrapersonal, interpersonal, and environmental factors [15, 16]. Intrapersonal barriers include pregnancy-related issues (e.g., lethargy/lack of energy, pregnancy gastrointestinal problems) and non-pregnancy-related factors (e.g., lack of confidence, pre-pregnancy PA habits) [17]. Interpersonal barriers encompass a lack of PA advice, a lack of support from friends/partner, and family responsibilities [14]. Environmental barriers include unfavorable weather, limited access to suitable spaces, and a lack of facilities [15]. A systematic review found that due to differences in economic development and cultural norms, barriers to PA pregnant women experienced in different regions vary significantly [10]. In China, affected by traditional philosophy, families often advise pregnant women to get plenty of rest to guard against miscarriage, especially during the first trimester [18]. This may result in more pregnant women engaged in inactive PA [13]. Consequently, it is vital to explore PA barriers and factors associated with them among Chinese pregnant women.

Current research on barriers to PA during pregnancy primarily relies on either quantitative or qualitative methods alone [17, 19, 20]. This approach limits the ability to comprehensively understand the causes of PA barriers or to provide an objective assessment of the levels of these barriers. Applying mixed methods research can overcome these limitations [21]. By integrating quantitative and qualitative data, it provides deeper insights into PA barriers during pregnancy, enabling findings to be triangulated and complemented.

This study adopted the Capability, Opportunity, Motivation, and Behavior (COM-B) model and the Theoretical Domain Framework (TDF). These two frameworks were selected based on their complementary strengths in analyzing health behavior barriers, their established validity in prior research, and their direct relevance to the study’s objectives. The COM-B model provides a multidimensional structure that categorizes behavioral determinants into three interdependent components, including capability (physical and psychological), opportunity (social and physical), and motivation (reflective and automatic) [22]. This model enables systematic categorization of PA barriers, thereby overcoming the limitations of single-dimensional analytical frameworks. To operationalize these components into specific and modifiable determinants, this study employed the TDF. The TDF elaborates on and refines the COM-B model into 14 domains and 84 constructs (e.g., Knowledge, Skills, Social Influences, Environmental Context) [23]. Importantly, each TDF domain can be systematically mapped onto the corresponding COM-B component (e.g., “Knowledge” onto psychological capability), providing a granular framework for diagnosing barriers and designing interventions. This integrated approach has been effectively applied in diverse health-related behaviors, such as sleep protection for premature infants [24], lifestyle management after childbirth [25], and rural diabetic older adults’ cognitive function interventions [26], demonstrating its utility and validating its application in complex health contexts. By applying this integrated framework, this study systematically identified modifiable determinants of PA through cross-theoretical mapping, providing a structured foundation for developing targeted, theory-informed interventions.

Therefore, this study adopted a mixed methods design to explore barriers and factors associated with PA among pregnant women by integrating the COM-B model and the TDF, thereby providing a theoretical foundation for designing interventions to promote PA engagement during pregnancy.

Methods

Study design

The study employed a sequential explanatory mixed methods design, comprising an initial quantitative cross-sectional survey (Phase 1) followed by a qualitative phase involving semi-structured interviews (Phase 2) [27]. This two-phase approach enabled a comprehensive exploration and explanation of both subjective and objective factors associated with barriers to PA among pregnant women.

Quantitative study

Participants

Convenient sampling was utilized to recruit pregnant women receiving routine prenatal care in the hospital obstetrics departments at three teaching hospitals in Yangzhou, China, between December 2023 and January 2024. Participants were included if they met the following eligibility criteria: (1) aged ≥ 20 years; (2) at any trimester with a singleton pregnancy; (3) able to read and write in Chinese and participate voluntarily. Participants were excluded if they had any contraindications to PA [9], such as unexplained persistent vaginal bleeding, pre-eclampsia, intrauterine growth restriction, incompetent cervix, or musculoskeletal disorders.

Data collection

Before the investigation, trained researchers explained the objective and significance of the study to pregnant women. After obtaining consent from eligible pregnant women for participation in the survey, they were instructed face-to-face by investigators using uniform guidelines to independently complete the paper questionnaire in a quiet room. The investigators promptly reviewed the questionnaire after completion to identify any missing data. If the investigators identified any missing data, they promptly reminded the participant and explained the significance of completing the questionnaire fully, thereby encouraging cooperation. According to the rough estimation method proposed by Kendall [28], the minimum sample size is 5–10 times the number of variables in the study. The study aimed to analyze 16 variables in total. Considering a 15% nonresponse rate, a minimum of 95 participants should be included. In this study, 415 questionnaires were distributed, and 15 invalid questionnaires were excluded. Ultimately, 400 valid questionnaires were collected, with an effective recovery rate of 96.39%.

Instruments

Demographic characteristics

The questionnaire collected individual characteristics, including demographic information (such as age, education level, employment status, place of residence, and income) and pregnancy-related information (such as trimester, parity, history of abortion, pre-pregnancy body mass index (BMI), desired delivery mode, whether specific PA guidance was received during pregnancy, and pre-pregnancy habitual PA). Participants were classified according to Chinese criteria for pre-pregnancy BMI (underweight, < 18.5 kg/m²; normal weight, 18.5–23.9 kg/m²; overweight, 24–27.9 kg/m²; and obese, ≥ 28 kg/m²) [29]. Regular pre-pregnancy PA was defined as participating in PA ≥ 5 times/week and lasting ≥ 30 min/time before pregnancy [1, 30].

Barriers to physical activity during pregnancy scale (BPAPS)

Amiri-Farahani et al. developed the BPAPS to assess the barriers to PA during pregnancy [31]. The scale has been translated into Chinese by our research team at an early stage. The Chinese version of the BPAPS was constructed strictly following the Brislin translation-back translation procedure. The final Chinese version of the BPAPS has good reliability and validity, consisting of 28 items with 4 domains (pregnancy-related intrapersonal barriers, non-pregnancy related intrapersonal barriers, interpersonal barriers, and environmental barriers). It was rated on a Likert 5-point scale ranging from strongly disagree to strongly agree. The total score is from 28 to 140, with higher scores associated with more barriers to PA during pregnancy. The Cronbach’s alpha coefficients of the total scale and subscales of pregnancy-related intrapersonal barriers, non-pregnancy related intrapersonal barriers, interpersonal barriers, and environmental barriers were 0.824, 0.815, 0.732, 0.734, and 0.722, respectively. In our study, the Cronbach’s alpha coefficients of the total scale and subscales were 0.942, 0.883, 0.831, 0.882, and 0.895, respectively.

Statistical analysis

Data entry and analysis were performed using EpiData 3.1 and SPSS 27.0, respectively. The normality of data was assessed using skewness and kurtosis. First, descriptive statistics were calculated for the BPAPS total and subscale scores. Given the normal distribution of the data, continuous variables are presented as mean ± standard deviation (M ± SD), and categorical variables as frequency (percentage). Secondly, univariable analyses were conducted to compare PA barrier scores across groups with different characteristics: independent samples t-tests were used for two-group comparisons, and one-way ANOVA was applied for comparisons involving more than two groups. Finally, variables with statistically significant differences (P < 0.05) in the univariable analyses were entered as independent variables into a multiple linear regression model to identify factors associated with PA barriers among pregnant women. We reported the adjusted β coefficient, 95% confidence intervals (CI), and p-values. A p-value of < 0.05 was considered statistically significant (two-tailed).

Qualitative study

Participants

Purposive sampling with maximum variation was used to select pregnant women who had previously participated in the quantitative research and were willing to participate in qualitative interviews. The participants varied in terms of age, education level, parity, place of residence, personal monthly income, and gestational week, thereby ensuring the inclusion of diverse experiences relevant to PA barriers. Data saturation was reached at interview 11, meaning that no new information and themes emerged at the end of the interview process [32]. Two further interviews were conducted to confirm the results. In our study, data saturation was assessed using three criteria: no new themes in two consecutive interviews, sufficient depth and richness to address all research questions, and no new codes in two additional interviews [32]. Interviews were continued until all criteria were met.

Development of the interview guide

Before the interviews, the interview guide was collaboratively developed by the primary author and two experienced nurses holding doctoral degrees specializing in maternal health and qualitative methods. This development was based on the research objectives, the findings of the quantitative research, a review of the literature, and the characteristics of psychosomatic changes experienced by women during pregnancy. The guide was piloted with two pregnant women (not included in the main study) to assess clarity and relevance. The final version was structured with questions designed to explore both general perceptions and specific experiences related to PA during pregnancy. The complete guide is available in Supplementary Material 1.

Data collection

Potential participants were initially contacted by telephone (using their voluntarily provided phone numbers). During this contact, the researcher reintroduced the study, confirmed their willingness to participate, and offered interview options based on their schedule, location, and preference. These options included in-person interviews (conducted in a quiet and private setting) and online video interviews (arranged for those unable to attend in-person via secure platforms like Tencent Meeting). Before each interview, the pregnant participants were fully informed about the purpose, significance, and main content of the study. Written informed consent was obtained before commencing audio recording and note-taking. During the interviews, the researcher employed active listening techniques, avoided asking leading questions, and captured contemporaneous notes documenting key points as well as participants’ facial expressions and tone of voice. Each interview lasted approximately 30 to 60 min until we reached saturation. All audio recordings were stored encrypted on a password-protected institutional server. Immediately after each interview, all personally identifiable information was removed and replaced with unique codes (e.g., P1). Field notes and transcripts followed the same anonymization protocol. Access was restricted to the principal investigator, and all data will be securely retained for five years post-study as per ethical requirements.

Statistical analysis

All records were promptly transcribed and validated by two researchers within 24 h after each interview. The codes derived by the researcher are organized, categorized, and managed in NVivo 12 Plus software. In this study, the collected interview data were analyzed using thematic analysis, following a six-step process: familiarization, generating codes, searching for themes, reviewing themes, defining and naming themes, and producing a report [33]. Data were read, analyzed, coded, categorized, and refined independently by 2 researchers trained in qualitative research, and when disagreements arose, they were discussed by the research team to reach a consensus and determine the final themes. In this study, credibility was enhanced through prolonged engagement, data triangulation, and member checking; dependability through audit trails, peer debriefing, and systematic data management with NVivo 12 Plus; confirmability through reflexivity, dual coding, and thematic triangulation with the COM-B and the TDF; and transferability by providing thick descriptions and direct quotations.

Integrating quantitative and qualitative findings

Guided by BCW and TDF, we conducted parallel analyses of quantitative and qualitative data, integrating them during interpretation and reporting through a weaving approach and joint displays [34]. The weaving method allowed for narrative synthesis, intertwining quantitative and qualitative findings thematically or conceptually. Meanwhile, joint displays visually presented the integrated data from both sources.

Results

Quantitative study

Demographic characteristics of the pregnant women

The mean age of the participants was 29.72 ± 3.58 years. Most participants had an education level of diploma degree or above (84.8%), and lived in urban areas (79.5%). 67.0% of the women were primiparous, and 76.0% were employed during pregnancy. The characteristics of the participants are shown in Table 1.

Table 1.

Univariable analysis of factors associated with barriers to PA among pregnant women (n = 400)

Variables		n (%)	Mean ± SD	t/F	P
Age (years)	< 35	356 (89.0)	79.30 ± 10.95	-1.479	0.140
	≥ 35	44 (11.0)	81.89 ± 11.01
Education level	High school or less	61 (15.3)	80.67 ± 9.80	0.469	0.626
	Diploma or bachelor’s degree	307 (76.8)	79.48 ± 11.35
	Master’s degree or above	32 (8.0)	78.50 ± 9.31
Current employment status	Employed	304 (76.0)	79.44 ± 11.23	-0.470	0.639
	Unemployed	96 (24.0)	80.04 ± 10.12
Place of residence	Urban	318 (79.5)	79.45 ± 10.94	-0.465	0.642
	Rural	82 (20.5)	80.09 ± 11.12
Personal monthly income (¥)	≤ 3000	16 (4.0)	83.69 ± 14.92	1.072	0.361
	3001–5000	106 (26.5)	78.92 ± 10.43
	5001–8000	172 (43.0)	79.99 ± 10.24
	> 8000	106 (26.5)	78.96 ± 11.92
Pre-pregnancy BMI (kg/m2)	Underweight (< 18.5)	37 (9.3)	80.19 ± 11.96	0.905	0.439
	Normal (18.5–23.9)	251 (62.7)	79.25 ± 10.44
	Overweight (24-27.9)	80 (20.0)	79.18 ± 10.01
	Obese (≥ 28)	32 (8.0)	82.50 ± 15.40
Trimester (week)	First trimester (≤ 13)	16 (4.0)	91.69 ± 16.90	10.838	< 0.001
	Second trimester (14–27)	50 (12.5)	79.94 ± 11.01
	Third trimester (≥ 28)	334 (83.5)	78.95 ± 10.30
Desired delivery mode	Vaginal delivery	198 (49.5)	77.87 ± 10.91	9.313	< 0.001
	Cesarean section	73 (18.3)	84.22 ± 9.98
	Both	129 (32.3)	79.59 ± 10.91
Parity	Primiparous	268 (67.0)	77.49 ± 9.49	-5.647	< 0.001
	Multiparous	132 (33.0)	83.83 ± 12.47
History of abortion	Yes	276 (69.0)	78.61 ± 10.30	-2.680	0.008
	No	124 (31.0)	81.76 ± 12.10
Receiving specific PA guidance during pregnancy	Yes	98 (24.5)	79.29 ± 10.44	-2.574	0.010
	No	302 (75.5)	86.24 ± 18.69
Pre-pregnancy habitual PA	Yes	101 (25.3)	74.70 ± 10.57	-5.347	< 0.001
	No	299 (74.8)	81.23 ± 10.62

Scores of the BPAPS and its domains

The total score of the BPAPS was 79.58 ± 10.97. The scores for the domains of the BPAPS were as follows: pregnancy-related intrapersonal barriers (26.86 ± 4.16), non-pregnancy related intrapersonal barriers (11.01 ± 2.76), interpersonal barriers (21.64 ± 4.33), and environmental barriers (20.15 ± 4.79). Interpersonal barriers were the most critical barriers among our study population, and the lowest scores were related to environmental barriers. The scores of the BPAPS domains and each item are shown in Supplementary Material 2.

Factors associated with barriers to PA among pregnant women

Univariable analysis of factors associated with barriers to PA among pregnant women.

The results of the univariable analysis are shown in Table 1. There were statistically significant differences between the total score of the barriers to PA and trimester (P < 0.001), desired delivery mode (P < 0.001), parity (P < 0.001), history of abortion (P = 0.008), receiving specific PA guidance during pregnancy (P = 0.010), and pre-pregnancy habitual PA (P < 0.001).

The results of the multiple linear regression analysis are shown in Table 2. Those independent variables with P < 0.05 in the univariable analysis were then entered into the multivariable linear regression model. The results revealed that cesarean section, parity, receiving specific PA guidance during pregnancy, and pre-pregnancy habitual PA were significantly related to PA barriers (P < 0.05). Specific assignments for variables are in Supplemental Material 3.

Table 2.

Multivariable linear regression analysis of factors associated with barriers to PA among pregnant women (n = 400)

Independent variables		β	95% CI	P
Desired delivery mode	Cesarean section (vs. Natural labor)	5.022	[2.275, 7.768]	< 0.001
	Both (vs. Natural labor)	2.060	[-0.240, 4.359]	0.079
Parity	Multiparous (vs. Primiparous)	6.235	[3.701, 8.769]	< 0.001
Receiving specific PA guidance during pregnancy	No (vs. Yes)	6.166	[1.271, 11.405]	0.014
Pre-pregnancy habitual PA	No (vs. Yes)	4.040	[1.541, 6.540]	0.002

R² = 0.280, Adjusted R² = 0.234, F = 6.081, P < 0.05

Qualitative study

Characteristics of participants

Following the principles of “adequacy” and “saturation” in qualitative research, a total of 13 interviewees were ultimately included in the study. The participants’ ages ranged from 23 to 40 years. Nine participants lived in urban areas, while four resided in rural regions. Table 3 shows the characteristics of the interviewed participants.

Table 3.

Characteristics of interviewed participants (n = 13)

Serial number	Age (years)	Education level	Place of residence	Personal monthly income (¥)	Parity	Gestational weeks
P1	30	Bachelor	Urban	3001–5000	Primiparous	35+ 2
P2	31	Bachelor	Urban	5001–8000	Primiparous	12+ 3
P3	28	Master	Urban	5001–8000	Primiparous	37+ 5
P4	37	Junior high school	Rural	3001–5000	Multiparous	8+ 2
P5	31	Master	Urban	3001–5000	Primiparous	20+ 3
P6	40	Diploma	Urban	5001–8000	Multiparous	26+ 2
P7	23	Junior high school	Rural	3001–5000	Primiparous	39+ 4
P8	36	Diploma	Rural	3001–5000	Multiparous	38+ 2
P9	38	Primary school	Rural	≤ 3000	Multiparous	23+ 4
P10	28	Master	Urban	> 8000	Primiparous	29+ 2
P11	33	Bachelor	Urban	5001–8000	Primiparous	35+ 3
P12	26	Diploma	Urban	3001–5000	Primiparous	16+ 4
P13	29	Bachelor	Urban	> 8000	Primiparous	39+ 1

Findings from thematic analysis

Thematic analysis of 13 interviews mapped barriers to the COM-B framework (capability, opportunity, motivation) and nine TDF domains, including skills, knowledge, memory, attention and decision processes, environmental context and resources, social influences, emotion, goals, belief about capability, and beliefs about consequences.

Capability-physical capability

Skills

Pregnancy-related physiological changes are common barriers to PA engagement, directly corresponding to the “Skills” domain within the TDF, which involves physical capacity and ability. Abdominal expansion during gestation directly compromises mobility, with progressive enlargement reducing both physical capacity and motivation for PA as pregnancy advances. P1 described how her diminished physical capability influenced her activity: “I wasn’t very active before getting pregnant, and now with this big belly, moving around is such a hassle. So honestly, I’m even less active now.”

Furthermore, the theme extended to include activity restriction due to medical advice, which shaped participants’ perceptions of their own physical capabilities. Gestational risk awareness leads to self-perceived physical inadequacy and subsequent activity restriction among high-risk pregnant women. As P7 explained, her medical condition and the doctor’s counsel directly influenced her assessment of what physical activity she could safely perform: “I have a low-lying placenta - the doctor advised me to cut back on exercise early in pregnancy. That’s why I’m extra careful about moving around now.”

Capability-psychological capability

Knowledge

Knowledge deficits, a key dimension of psychological capability in the COM-B model, emerged as a prominent barrier. Specifically, inadequate understanding of evidence-based activity types and appropriate intensity thresholds for gestation prevented most pregnant women in this study from achieving guideline-recommended PA levels. This lack of knowledge was directly articulated by participants who described receiving insufficient professional guidance. P2 explained: “At my prenatal checkups, the doctor didn’t really discuss PA stuff with me, so I’m not really sure about it.” Similarly, P13 highlighted the gap between basic awareness and actionable, detailed knowledge: “I know the basics about staying active, but not the details - like what types of PA are actually safe, or how intense it should be.”

Pregnant women expressed a clear need for trustworthy and well-defined activity guidance specifically designed for pregnancy. They desired information directly linked to pregnancy outcomes and offering diversified options for personalized adaptation. This unmet demand for credible, tailored information is exemplified by P5’s statement: “Mostly I just go for walks. Beyond that, I don’t really know how to do other physical activities—nobody ever showed me how. But if I could get that info from trustworthy sources, I’d probably try moving more.”

Memory, attention, and decision processes

This TDF domain, which encompasses the cognitive processes guiding behavioral choices, was operationalized through pregnant women’s considerations regarding their planned mode of delivery. These considerations directly influenced their PA engagement. Decisions regarding the planned mode of delivery influence pregnant women’s PA levels. Specifically, the cognitive appraisal of PA’s relevance to their chosen delivery outcome served as a key decision-making factor. Women planning a cesarean section tend to have lower motivation to engage in PA, whereas those planning a vaginal delivery are generally more active. This perception is captured in P4’s rationale for her limited activity: “Since I’m planning to have a C-section, I don’t think it matters that much.” Conversely, women planning a vaginal delivery often held the belief that PA would facilitate a smoother birth, thereby enhancing their motivation to be active. P11’s statement illustrates this motivating belief: “I want to deliver naturally, so I feel like staying active helps with that.”

Opportunity-physical opportunity

Environmental context and resources

This TDF domain, representing the physical and systemic factors external to the individual, was reflected in several key barriers and facilitators. Work was identified as a structural barrier impeding PA engagement, where time constraints often overrode intentions to be active. P2 exemplifies how occupational demands directly compete with PA opportunities in her daily routine: “Work keeps me too busy to be active normally. By the time I get home, all I want to do is crash on the couch. Working out is the last thing on my mind.”

Furthermore, the immediate physical environment presented challenges. Weather conditions and perceptions of public space safety directly influenced participation. P3 described how adverse weather created a physical and motivational barrier: “If it’s cold or super hot outside, or if it’s raining, it’s a hassle to go out. On days like that, I’m not up for doing much activity.” Conversely, a supportive environment could serve as a facilitator. P9 highlighted this: “I really like being active in the park. The environment there is nice and relaxing.” Safety concerns in public spaces were another environmental constraint, as noted by P5: “I think it’s way too crowded outside with all the people and traffic everywhere – it just doesn’t feel safe. So, I mostly just stay in.”

Participants also pointed to the healthcare system as a critical environmental resource. The availability of specialized PA guidance and services was seen as a key facilitator. Both P11 and P12 articulated a demand for structured, professional support within the healthcare environment to overcome knowledge and access barriers: P11 suggested, “The hospital could offer more classes about PA, like yoga or something. With professional guidance, we’d be more willing to get active. If they had both online and in-person options, we could pick what fits our schedule – that would be way more convenient.” and P12 added, “I think PA clinics sound great, but they’re not common yet. Plus, I’ve never had a pro tell me how to exercise safely. That’s why I usually just do workouts from videos at home.”

Opportunity-social opportunity

Social influences

The “Social Influences” domain in the TDF encompasses the ways in which interactions with others affect behavior. In this study, social influences manifested primarily through family, friends, and spouses, acting as both powerful facilitators and barriers to PA. Family members played a vital role in pregnant women’s PA, with their attitudes directly shaping participation. The facilitative impact of supportive encouragement is captured in P2’s description of her family’s active involvement: “My family’s supportive about me staying active. Usually, they’ll come along with me when I exercise!” Conversely, overly protective or restrictive attitudes from family could become a significant barrier, as noted by P8: “I do want to get some physical activities, but my family worries too much – they won’t really let me do too much.”

Beyond family, companionship from peers served as another crucial social factor. Pregnant women recognized friends’ companionship as a significant form of support that enhanced engagement. P4 articulated how the lack of this companionship currently limits her activity, but also its potential motivating power: “Honestly, none of my friends around me really stay active, so I don’t do much either. But if someone could come with me and remind me to get moving, I’d be way more willing to join in.” The positive effect of this tangible peer support is evident in P7’s contrasting experience: “When I go out with friends, I move more – we end up walking around everywhere together. It’s totally different from when I’m just stuck at home barely doing anything.”

Among all social relationships, spousal accompaniment emerged as the most influential factor. It served as the primary motivator for sustained participation, combining practical and emotional support. P6’s account illustrates how her husband’s involvement provides motivation, security, and practical facilitation: “It’s just me and my husband at home. We’re both crazy busy, but he always tries to make time to be active with me. Having him there makes me feel secure… plus honestly, he wouldn’t feel comfortable with me going alone anyway.”

Motivation-automatic motivation

Emotion

Within the TDF, the “Emotion” domain encompasses affective responses that can automatically drive or hinder behavior. In this study, positive emotional rewards emerged as a significant automatic motivator for PA engagement. The sense of accomplishment gained from completing PA was a key factor associated with pregnant women’s engagement. This is clearly illustrated by P3: “When I’m active during the day, it feels like I’ve accomplished something good. That sense of achievement really motivates me to keep it up.”

Motivation-reflective motivation

Goals

The “Goals” domain within the TDF pertains to conscious aims or desired outcomes that guide behavior. In this study, goal-setting was identified as a reflective motivational strategy for enhancing PA engagement. Setting PA goals served as a method for pregnant women to translate future-oriented health intentions into actionable plans. P5’s statement clearly illustrates how setting specific, outcome-linked goals operates within this reflective cognitive process to guide and sustain her behavior: “I’m staying active now hoping for an easier delivery and a quicker post-baby recovery. That’s why I set activity goals and I’m sticking with them.”

Beliefs about consequences

This TDF domain focuses on the beliefs individuals hold about the likely outcomes of a behavior, which in turn shape their engagement. Pregnant women’s PA participation was influenced by their anticipated consequences, encompassing both positive benefits and perceived risks. When pregnant women perceived clear benefits of PA, they were more likely to participate proactively. P11’s statement exemplifies this, showing how her belief in the positive health outcomes for herself and her baby served as a motivating factor: “I believe staying active regularly makes both me and my baby healthier and more energetic. That’s why I try to be more active day-to-day.”

Conversely, beliefs about potential negative consequences, particularly among primiparous women, acted as significant barriers. Worries over potential risks prompted a cautious approach, often limiting activities to low intensity. This risk-avoidant decision-making, rooted in beliefs about harmful outcomes, is clearly articulated by P12 and P5. P12 explained: “Honestly, I’m mostly worried about the baby. I’ve heard people say that being active during pregnancy might increase the risk of premature birth. So, I usually just stick to things like taking walks and don’t do much else.” P5 similarly stated: “Since it’s my first pregnancy, I don’t know much. I’m scared that being too active might not be good for the baby, which is why I mostly stay put.”

Beliefs about capabilities

This TDF domain, closely aligned with the concept of self-efficacy, refers to individuals’ confidence in their ability to perform a specific behavior. Such beliefs were fundamental to sustaining PA during pregnancy. Pregnant women with strong confidence in their PA capabilities demonstrated resilience in managing activity-related challenges. P3’s statement powerfully illustrates how her high self-efficacy influenced her behavioral persistence: “Every time I do PA, I get really tired, but I believe I can push through it – and I always end up doing it.” Conversely, a lack of such confidence could undermine participation. This positive belief in one’s own capacity, even amidst physical exhaustion, is further exemplified by P10: “Working out always leaves me exhausted, but I know I have it in me to keep going. And in the end, I do.”

Integration of quantitative and qualitative data

Findings from both quantitative and qualitative studies were mapped to the three core components of the COM-B and the nine theoretical domains of the TDF. Table 4 showed the joint displays of quantitative and qualitative results.

Table 4.

Joint display of quantitative and qualitative results

BCW	TDF	Quantitative findings	Qualitative findings	Integration of quantitative and qualitative findings
Capability (Psychological Capability)	Knowledge	The score of the item 20 “I do not do PA because I do not have access to complete information about PA during pregnancy” was (3.37 ± 0.91).	• Inadequate professional guidance on PA during prenatal visits. • Knowledge was limited to basic concepts; specifics on safe PA types, intensity, and duration were lacking. • Expressed a need for credible, tailored information from healthcare providers.	Pregnant women exhibit limited understanding of PA during pregnancy, with an inadequate knowledge base on the subject.
	Memory, Attention, and Decision Processes	The desired delivery mode was the factor associated with the barriers to PA among pregnant women (P < 0.001).	• Women planning vaginal delivery believed PA would facilitate a smoother birth, increasing motivation. • Those opting for cesarean section perceived PA as having minimal impact on delivery outcomes, showing less initiative.	The cognitive appraisal of PA’s relevance to the planned delivery outcome directly influences motivation and engagement, creating a clear behavioral divide.
Capability (Physical Capability)	Skills	The score of pregnancy-related intrapersonal barriers domains was (26.86 ± 4.16).	• Physiological changes (e.g., abdominal enlargement, fatigue) reduced willingness to engage in PA. • Activity restrictions due to medical advice shaped self-perceived capability.	Discomfort stemming from pregnancy-related physiological changes may constrain PA engagement among pregnant women.
Opportunity (Physical Opportunity)	Environmental Context and Resources	The scores for items regarding unfavorable weather and unsafe parks were (3.00 ± 0.95) and (2.54 ± 0.84).	• Weather conditions (cold/heat/rain) and safety concerns in public spaces were frequently cited barriers. • Work commitments limited the time for PA. • Expressed demand for structured, professional PA guidance and services within healthcare settings.	Environmental and structural factors, including weather, safety, time constraints, and lack of accessible professional support, consistently emerged as key external barriers to PA.
Opportunity (Social Opportunity)	Social influences	Item 23 and item 24 on lack of professional advice were (3.13 ± 1.02) and (3.39 ± 0.98). Receiving specific PA guidance during pregnancy was a significant predictor (P = 0.014).	• Spousal accompaniment and encouragement were the most influential motivators. • Family support facilitated PA, while overprotective attitudes could hinder it. • Peer companionship enhanced adherence through mutual monitoring.	Social support, particularly from spouses, is a critical facilitator, whereas lack of professional guidance and overprotective family norms constitute significant interpersonal barriers.
Motivation (Automatic Motivation)	Emotion		Participating in PA can trigger positive psychological feedback among pregnant women, producing a significant sense of accomplishment.	Immediate positive affective responses to PA, such as feeling a sense of achievement, function as intrinsic automatic motivators that help sustain the behavior.
Motivation (Reflective Motivation)	Goals	Pre-pregnancy habitual PA was the factor associated with the barriers to PA among pregnant women (P = 0.002).	• Goal-setting was used to translate health intentions into actionable plans. • Recurrent obstacles during goal pursuit could diminish motivation.	Clear PA goals help establish routines, but their feasibility is critical for sustained motivation, especially for women without pre-existing active habits.
	Beliefs about Consequences	Parity was the factor associated with the barriers to PA among pregnant women (P < 0.001). The score of the item 15 “I am concerned by possible pregnancy complications such as miscarriages and premature labor” was (3.00 ± 0.82).	• Belief in health benefits for the mother and baby motivated PA. • Primiparous women more frequently reported fears of risks (e.g., harming the baby).	Perceived benefits of PA motivate proactive engagement among pregnant women. Conversely, perceived risks associated with PA significantly shape engagement decisions among primiparous women.
	Beliefs about Capabilities	The scores of items on PA being “too hard work” and “lack of confidence” were (2.55 ± 0.86) and (2.62 ± 0.92).	• High self-efficacy was linked to persistence despite fatigue. • Confidence in one’s ability to perform PA was seen as foundational for sustained engagement.	Self-efficacy beliefs are central to overcoming physical and mental challenges, influencing whether women persist with PA or disengage.

Discussion

To our knowledge, this study was the first attempt to use a mixed methods study to explore the barriers to PA and associated factors among Chinese pregnant women based on the COM-B model and TDF. In this research, the overall scores of the BPAPS were 79.58 ± 10.97, which was similar to a study in Nigeria (85.35 ± 22.82) [16]. Specifically, the highest score was related to interpersonal barriers, and the lowest score was related to environmental barriers.

Capability-physical capability

Skills

In this study, the score of pregnancy-related intrapersonal barriers domains was (26.86 ± 4.16), with lethargy, pain, and drowsiness identified as the most prevalent specific barriers. Furthermore, qualitative findings complemented abdominal changes as a notable barrier. These integrated findings aligned with a previous study [35], showing that confirming that such physiological symptoms adversely affect women’s willingness to engage in PA, thereby reducing their overall PA levels during pregnancy. Approximately 63% of pregnant women experienced pregnancy-related low back pain, a common physical discomfort during pregnancy [36]. As many of these women fear that PA could exacerbate pain, they often avoid it [37]. However, these findings contrast with studies reporting that structured and supervised prenatal PA does not worsen pain and may, in fact, alleviate symptoms such as low back pain and improve overall well-being [38, 39]. Evidence from 13 randomized controlled trials demonstrated that prenatal PA reduced pain intensity and improved function among pregnant women with low back pain compared with control groups [40]. This contradiction suggests that pregnant women’s fear of PA may be related to a lack of proper guidance, rather than PA itself. Therefore, to address these specific intrapersonal barriers, it is worth considering the provision of coping skills for managing physiological stimulation during pregnancy to alleviate physical discomfort [41].

Capability-psychological capability

Knowledge

The quantitative findings of this study revealed insufficient knowledge regarding PA during pregnancy (3.37 ± 0.91). These findings were substantiated qualitatively, with participants consistently identifying deficient PA knowledge and inadequate professional support as common barriers. This aligns with a study of 1,283 pregnant women [42]. A previous qualitative study indicated that limited knowledge regarding PA contributed to a decline in motivation among pregnant women [43]. The primary reason for reliance on inadequate information appears to be a lack of professional guidance. In China, obstetricians often prioritize routine medical tasks during prenatal check-ups, leaving insufficient time for PA counseling [42]. Consequently, pregnant women most commonly turn to the Internet for information, which may be inaccurate, confusing, and overwhelming [44]. A study from Australia highlighted that integrated PA education in prenatal care and standardized clinical guidelines have improved PA knowledge among Australian pregnant women, with most receiving tailored guidance from obstetricians or exercise physiologists [45]. This contradicts the knowledge gaps observed in the Chinese context. Countries like Australia have incorporated evidence-based PA guidelines into routine antenatal care, such integration remains limited in China, where clinical focus remains on medical monitoring rather than lifestyle counseling [46]. To address this gap, a collaborative approach where nurses and midwives work with obstetricians to integrate structured PA education into routine antenatal care is recommended.

Memory, attention, and decision processes

The integrated results indicate that the desired delivery mode significantly affects pregnant women’s PA barriers (P < 0.001). Specifically, Participants intending vaginal delivery reported fewer barriers and were more likely to increase PA, whereas those planning cesarean section demonstrated lower activity engagement. This association is supported by Shang’s study, indicating that delivery mode beliefs influence women’s attitudes and behaviors toward PA [17]. Evidence has demonstrated a positive correlation between vaginal delivery beliefs and PA participation, as women perceive that PA can enhance physical stamina, shorten labor duration, and increase the likelihood of vaginal delivery [47]. Therefore, during antenatal education, it is essential to emphasize the advantages of vaginal delivery, such as quicker maternal recovery, to encourage a preference for natural childbirth when medically appropriate [48]. Conversely, women planning cesarean delivery may perceive that PA has minimal impact on the labor process, thus lacking motivation for PA. However, it is crucial to address this misconception. Evidence has shown that PA offers significant health benefits for women undergoing cesarean section. It can reduce the risk of several pregnancy complications associated with a higher likelihood of cesarean section [49]. Healthcare providers should proactively communicate these benefits to motivate this subgroup. In a word, PA is important regardless of the chosen mode of delivery, and healthcare providers need to ensure pregnant women understand this. Pregnant women ought to select their delivery mode based on their individual circumstances and clinical recommendations after being informed of the advantages and disadvantages of vaginal delivery and cesarean section. At the same time, they should be supported and educated to maintain physical activity throughout pregnancy, recognizing its value for their health and well-being. For those planning a cesarean section, counseling should specifically highlight the unique benefits of PA for optimizing pre-surgical health, reducing perioperative complication risks (e.g., infection, thrombosis), and enhancing postoperative recovery, moving beyond its perceived impact on labor.

Opportunity-social opportunity

Social influences

Both quantitative and qualitative findings pointed out that social support serves as a significant facilitator of PA among pregnant women. This aligns with Grenier’s study showing that higher PA levels are associated with those with strong support networks [50]. Conversely, insufficient social support was significantly associated with increased PA barriers [15]. This may stem from emotional companionship from partners and family, which builds confidence in PA capabilities, thereby significantly increasing PA participation [20]. Critically, as pregnant women identify family support (particularly spousal support) as a key PA motivator [51], future initiatives should prioritize family health education targeting partners and other family members specifically. Such programs must enhance their understanding of prenatal PA benefits to foster supportive actions. In addition, this quantitative study found participants lacking specific PA guidance during pregnancy exhibited higher PA barrier scores compared to those receiving specific PA guidance (P = 0.014), aligning with previous research [19]. A randomized controlled trial indicated that providing only written PA guidelines without personalized counseling did not significantly reduce PA barriers among pregnant women [52]. This finding underscores the imperative for providing specialized PA guidance during pregnancy.

Opportunity-physical opportunity

Environmental context and resources

Quantitative results identified unfavorable weather (3.00 ± 0.95) and unsafe PA environment (2.54 ± 0.84) as primary environmental barriers to PA during pregnancy. In-depth interviews substantiated these findings, with participants emphasizing weather constraints and space safety apprehensions. These findings resonate with both national and international research. A study in China similarly reported that extreme temperatures (both hot and cold) and air pollution were primary environmental barriers, closely mirroring our “unfavorable weather” factor [17]. Internationally, a study by Mottola et al. identified “lack of safe and pleasant places to walk” as a key barrier [9]. This consistency across diverse contexts underscores the universality of these environmental challenges for pregnant populations. Crucially, weather-related barriers exhibit significant seasonal and geographical variations [14], necessitating seasonal adaptations: pregnant women should be encouraged toward indoor PA during inclement conditions. To enable this, public sports facilities require retrofitting with all-weather protections (e.g., rain shelters, heating systems). Regarding environmental safety, evidence confirms that perceived risks in public spaces [53, 54], including fear of unleashed dogs [14], deter PA engagement. These findings underscore the need for evidence-based safety interventions targeting community environments and public spaces to facilitate PA engagement during pregnancy. Beyond community-level interventions, individual-level strategies based on behavior change techniques can also be effective [55]. This involves restructuring the physical environment at home to facilitate PA performance. For instance, pregnant women can be encouraged to place PA reminders in prominent locations (e.g., on the refrigerator, TV) to serve as prompts. They can also be guided to use readily available household items for PA, such as using water bottles for resistance exercises.

Motivation-automatic motivation

Emotion

Qualitative data highlighted the significance of emotional factors in enhancing PA engagement among pregnant women. Positive psychological experiences associated with PA, like enjoyment, happiness, or empowerment functioned as intrinsic motivators that foster constructive PA attitudes and sustain behavioral maintenance [20, 47]. Successful PA experiences produce immediate rewards, enhancing PA participation more than delayed rewards like long-term health benefits [56]. Conversely, negative experiences such as pain, injury, or boredom can decrease motivation to participate [56]. A qualitative study by Hegaard et al. in Sweden similarly identified that the primary predictors of PA maintenance during pregnancy were the affective responses to the activity itself—specifically, finding pleasure and satisfaction in it [57]. This aligns with our emphasis on immediate positive experiences as key drivers. Therefore, monitoring emotional responses during PA is critical, enabling researchers to tailor subsequent tasks accordingly [58]. For those experiencing emotional difficulties, informing them about the mood-boosting benefits of PA and providing emotional support strategies, such as stress management techniques, can enhance PA participation.

Motivation-reflective motivation

Goals

This study found that women who did not engage in pre-pregnancy habitual PA had higher total PA barrier scores than regularly active women (P = 0.002). This finding aligns with studies conducted in the United States and Iran [19, 59]. Furthermore, it was supported by previous research indicating that women physically active before pregnancy encountered fewer barriers to PA during pregnancy and found it easier to maintain their PA levels [60]. Collectively, these results emphasized the importance of promoting PA among reproductive women. Implementing step-by-step goal-setting strategies can facilitate this population in progressively establishing sustainable PA habits.

Beliefs about consequences

Regarding obstetric factors, multiparous women had greater PA barriers than primiparous women (P < 0.001), contrasting with existing literature [42, 61]. We attribute this disparity to multiparas typically assuming greater childcare responsibilities, which impose practical consequences (e.g., limited time and energy) that override their positive perceptions of PA benefits [17]. Our qualitative findings revealed that compared to multiparas, primiparous women overestimate the negative consequences of PA (e.g., risk to maternal-fetal health) due to limited experience, leading to avoidance of activity despite having more time and energy availability. Pregnant women of different parities have different perceptions of PA. Therefore, providing parity-specific PA guidance is necessary. For primiparous women, it is crucial to deliver clear information on the health effects of PA through multiple modalities, such as images, audio recordings, videos, and real-life case studies. For multiparous women, practical time-management training can be offered to facilitate the incorporation of PA into their daily schedules. Future interventions should consider providing childcare services or developing family-inclusive PA programs to address this practical barrier.

Beliefs about capabilities

The integrated results identified PA self-efficacy as a significant factor associated with PA engagement among pregnant women, which was consistent with the previous studies showing PA self-efficacy was directly associated with PA [62, 63]. This finding was supported by Bandura’s self-efficacy theory [64], which proposed that individuals with higher self-efficacy levels had greater capacity to tackle challenging tasks, expend increased effort, and persist more effectively when confronting adversity. PA self-efficacy is influenced by complex factors, including the difficulty of physical tasks, encouragement, communication, as well as the individual’s physiological and psychological conditions [65]. Enhancement of PA self-efficacy can be facilitated through structured interventions involving gradually increasing task difficulty tailored to the individual’s condition and giving timely feedback and recognition for progress at each pregnancy stage.

Limitation

There are several limitations in our study. First, participants in this study were recruited from three teaching hospitals in China, limiting our findings’ generalizability. Future multi-center studies with large samples are warranted to improve the generalizability of this study’s results. Second, self-reported questionnaires were used for data collection, resulting in potential recall bias and measurement inaccuracies. Future studies should incorporate objective measures with subjective tools to measure study variables. Third, longitudinal studies are required to track barriers to PA across the prenatal, antenatal, and postnatal periods, given the cross-sectional design of this study. Fourth, although our qualitative sample included multiparous women (n = 4 out of 13), their relatively small proportion may have limited our ability to fully capture and understand the specific barriers related to childcare responsibilities. Future qualitative studies should purposively recruit a larger and more diverse sample of multiparous women to explore this significant barrier in greater depth.

Conclusions

This study systematically identified a spectrum of barriers to PA among Chinese pregnant women using the COM-B model and the TDF. The barriers spanned all three COM-B components: Capability (e.g., knowledge gaps, physical discomfort), Opportunity (e.g., lack of support, environmental constraints), and Motivation (e.g., low self-efficacy, safety concerns). To effectively address these interconnected barriers, future interventions must be multifaceted. Individualized counseling is needed to build skills and motivation, while family and community-level initiatives are crucial to foster supportive environments and accessible resources. Integrating structured PA guidance into routine prenatal care is essential for sustainable impact.

Abbreviations

PA

Physical Activity

COM-B

Capability, Opportunity, Motivation, and Behavior

TDF

Theoretical Domains Framework

BPAPS

Barriers to Physical Activity during Pregnancy Scale

WHO

World Health Organization

BMI

Body Mass Index

Authors’ contributions

**LY: ** Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft. **FW: ** Methodology, Investigation, writing - original draft. **LL: ** Investigation, Writing – original draft. **YL: ** Funding acquisition, Investigation, Methodology, Writing – review & editing. **XS: ** Conceptualization, Formal analysis, Funding acquisition, Methodology, Writing – review & editing. All authors reviewed the manuscript.

Funding

This study was supported by grants from the Key Project of the Nursing Research Special Program of Northern Jiangsu People’s Hospital (SBHL25004), the Key R&D Projects of Yangzhou (YZ2021067), and the ‘Huxin Fund’ project of Jiangsu Key Laboratory of Zoonotic Diseases (HX2414). The funding body played no role in the design of the study, the collection, analysis, and interpretation of data, and in writing the manuscript.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study followed the guidelines of the Declaration of Helsinki and was approved by the School of Nursing Ethics Committee, Yangzhou University (Approval No. YZUHL20230016). All respondents were informed about the study procedures and provided both verbal and written consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.