A cross-sectional study design with self-report survey data was used. Data were collected as part of a broader study examining how immigration and integration experiences of recently immigrated first-generation Latinas influence perceptions of and opportunities for PA. Participants were a convenience sample of Latinas recruited from the metropolitan area of Phoenix. Recruitment strategies included flyer advertisements and in-person recruitment at predominately Hispanic-serving community outreach centers, health clinics, religious institutions, and local businesses. The Phoenix metropolitan area includes the cities of Phoenix, Mesa, and Chandler and is predominately comprised of urban and suburban neighborhoods. The region has an estimated 4.8 million residents, with 30% (ie, 1.5 million) identifying as Hispanic or Latinx [27]. Among residents identifying as Hispanic or Latinx, 400,000 are foreign-born and approximately 60,000 immigrated to the United States after 2010 [27]. Women were eligible for participation in this study if they (1) self-identified as Latina, (2) immigrated into the United States within the past 10 years, (3) were aged ≥18 years, and (4) self-reported an ability to read in English or Spanish. Only women immigrating into the United States within the previous 10 years were included because the purpose of the larger study from which data were collected was to explore how contemporary social and contextual factors, including the impact of more restrictive immigration policies enacted during the US presidency of Donald Trump, were related to PA engagement among recently immigrated Latinas.

Women interested in study participation completed an eligibility screening interview either in-person or via telephone (according to the recruitment method). Eligible women were then given the option to complete the study questionnaire packet in English, Spanish, or a combined English/Spanish language using a completion method of their choice. Methods of survey completion included online via Qualtrics or by an oral interview. All participants were provided with an informed consent document prior to completing the survey. Participants completing the survey online provided informed consent by selecting the “continue” button on the webpage displaying the informed consent document that was accompanied by the following statement: “By clicking the Continue button, I acknowledge that I am at least 18 years old and that I consent to conducting the survey online.” Verbal informed consent was obtained from participants completing the survey via the oral interview or pencil-and-paper format. After completing the survey, participants were provided a US $25 gift card for study participation. All study procedures were approved by the Institutional Review Board of Arizona State University.

Descriptive statistics (ie, means, 95% confidence intervals, medians, interquartile ranges, and frequencies) were used to summarize participant demographic, PA, and neighborhood variables. A series of regression models were used to examine associations between neighborhood environmental factors and the PA outcomes of leisure-time PA, transportation PA, and overall moderate-to-vigorous PA. Associations between neighborhood factors and work and household PA were not examined because we lacked a theoretical rationale for why neighborhood factors would influence these PA variables. Ordinal least squares (OLS) regression models, controlling for age and level of education, were used to examine the independent effect of each neighborhood factor on continuous PA outcomes in MET-minutes/week. Logistic regression, controlling for age, was used to examine associations between neighborhood factors and dichotomous PA outcomes of engaging in leisure-time PA (ie, engaging in 0 versus >0 MET-minutes/week of leisure-time PA), transportation PA (ie, engaging in 0 versus >0 MET-minutes/week of transportation PA), and meeting national PA guidelines (ie, engaging in <500 versus ≥500 MET-minutes/week of moderate-to-vigorous PA). Level of education as a control variable was not included in logistic models because of perfect prediction. As a result of the collinearity of neighborhood variables, associations between each neighborhood variable (ie, walking environment, aesthetic quality, safety, violence, social cohesion, and activities with neighbors) and PA outcomes were examined separately. Stata/SE version 16.0 (StataCorp) was used for data analysis.

A total of 50 first-generation Latinas participated in the study. However, preliminary data cleaning revealed that 10 participants reported unreasonably high PA data (ie, the sum of all walking, moderate, and vigorous PA time was greater than 960 minutes or 16 hours/day) and 1 participant had incomplete IPAQ data. These women were excluded from data analysis according to IPAQ scoring guidelines [31], resulting in a final sample size of 39 participants. Sensitivity analyses (ie, chi-square test for categorical variables and t test for age and duration of US residency) were conducted to explore demographic differences between participants excluded from outcome analyses (n=11) and those included. Results showed that the women excluded from the study had a lower education level than those included (P=.003). No other demographic differences were observed between women included in the study and those excluded.

Among the 39 women included in outcome analyses, the mean age was 40.5 (SD 4.3) years and the mean duration of US residence was 4.6 (SD 1.0) years. The majority of the participants were from Mexico (27/39, 69%), with the remaining participants from various Central and South America countries. Approximately half (22/39, 56%) of the women were married and the majority spoke Spanish exclusively at home (32/39, 82%). Based on data provided by the IPAQ, only 7 (18%) of the 39 participants reported performing paid or unpaid work outside of the home. Complete demographic characteristics are presented in Table 1.

Table 2 provides a detailed overview of intensity- and domain-specific PA patterns of our sample. Participants reported a median of 4512 MET-minutes/week of total PA. Intensity-specific PA outcomes showed that the majority of this PA was achieved through moderate-intensity activity (median 2820 MET-minutes/week). Median self-reported MET-minutes/week of walking and vigorous-intensity PA were 594 and 0, respectively. Overall, 85% (33/39) of participants reported PA levels that met the requirements of the PA guidelines (ie, ≥500 MET-minutes/week of moderate-to-vigorous PA).

Examination of domain-specific PA patterns showed that domestic/household PA was the most frequent type of PA reported by participants (34/39, 87%) and, on average, how participants acquired the majority of their activity (ie, median 2160 MET-minutes/week). Leisure-time PA was the second most frequently reported domain of activity (26/39, 67%), followed by transportation (23/39, 59%) and work (7/39, 18%) PA. Additionally, although only 7 participants reported working outside of the home according to the IPAQ, the amount of PA acquired by these participants through work activity was higher than any other domain (4590 MET-minutes/week; see Table 2).

Table 3 shows the descriptive values for scales assessing neighborhood environmental factors. OLS regression results, presented in Table 4, revealed no significant associations between neighborhood environmental factors and continuous PA measures. Table 5 presents logistic regression models examining associations between neighborhood environmental factors and dichotomous PA outcomes of engaging in transportation PA versus not engaging in transportation PA, engaging in leisure-time PA versus not engaging in leisure-time PA, and meeting PA guidelines versus not meeting PA guidelines. Results showed that the neighborhood factors of walking environment, aesthetic quality, and safety were significantly associated with engaging in leisure-time PA (OR 5.95, 95% CI 1.49 to 23.74; OR 2.45, 95% CI 1.01 to 5.93; and OR 3.30, 95% CI 1.26 to 8.67, respectively) and meeting national PA guidelines (OR 4.15, 95% CI 1.13 to 15.18; OR 6.43, 95% CI 1.15 to 28.39; and OR 2.53, 95% CI 1.00 to 6.36, respectively). No other significant associations were found.

This study explored PA patterns among first-generation Latina immigrants residing in the metropolitan area of Phoenix, Arizona, and the influence of perceived neighborhood environmental factors on these PA patterns. This work adds to the limited body of research on domain-specific PA patterns among first-generation Latina immigrants, with the majority of previous PA studies among Latinas having focused exclusively on leisure-time PA [34] and failing to differentiate by generational status. Given that first-generation immigrants likely have different lived experiences and lifestyle activities than Latinas who were born in the United States, work is needed in order to inform the development of PA interventions for this population.

Results showed that the majority of our sample (85%) reported PA levels that met or exceeded national PA guidelines. Domain-specific outcomes indicated that participants accrued the majority of their PA through domestic and household activities, with limited PA performed for leisure or transportation purposes. The perceived neighborhood environmental factors of walking environment, aesthetic quality, and safety were positively associated with engaging in leisure-time PA and meeting national PA guidelines. The neighborhood factors of violence, social cohesion, and activities with neighbors were not related to PA outcomes.

Our finding that domestic and household activities accounted for the majority of PA performed reflects the outcomes of several previous studies examining PA patterns among Latinas [8,20] and supports the notion that Latinas, regardless of generational status, perform extensive caretaking and household activities as part of their daily routine. Likewise, low levels of leisure-time PA and a limited number of participants engaging in paid or unpaid work outside the home have also been previously reported among studies comprised predominantly of first-generation Latinas [19,20,33]. Leisure-time PA, in particular, has profound benefits for reducing cardiometabolic disease conditions [23,35,36]. Moreover, when compared with other PA domains (ie, work and transportation), leisure-time PA may be more amendable to change in the context of a PA intervention because it is within the volitional control of many Latinas. In future work, researchers should explore intervention strategies to increase leisure-time PA among first-generation Latinas. Such work may be key to reducing the disproportionate metabolic disease burden in this high-risk population.

Our examination of overall and intensity-specific PA patterns showed that our sample reported engaging in rather high levels of overall PA (ie, 4512 MET-minutes/week), with the majority of the PA being at a moderate intensity. We speculate that these high PA levels may be a result of overreporting, as this is a commonly reported occurrence with self-report PA measures [37]. Nicaise and colleagues [20] illustrated this issue in a previous study with low-income Latinas using the IPAQ and accelerometers. When assessing PA levels using the IPAQ, the authors found that 73% of Latinas met PA guidelines, with the bulk of the PA being accrued through domestic or household activities [20]—similar to the outcomes of our study. However, when examining PA levels with accelerometers, this percentage was reduced to 20%, suggesting that participants likely overestimated the intensity and the amount of PA accrued through domestic and household activities [20]. We speculate that a similar phenomenon may have impacted our study outcomes. Another key finding of our study was that our sample of Latinas reported limited vigorous-intensity PA, which also mirrors the outcome of several previous studies [18,20]. We hypothesize that this outcome was related to the limited amount of leisure-time PA reported by our sample, as vigorous PA is predominately achieved through purposeful exercise as opposed to daily activities [38].

When examining associations between neighborhood environmental factors and PA outcomes, linear regression models failed to show any significant associations. This might be due to the nonlinear patterns of the associations and a low statistical power due to the small sample size. However, logistic regression models revealed that the neighborhood factors of walking environment, aesthetic quality, and safety were associated with reported engagement in leisure-time PA and meeting national PA guidelines. These findings confirm previous studies showing that these factors are positively associated with PA engagement [19,39-42]. We also found it interesting that these 3 factors are ones that recent immigrants likely have limited control over (ie, these factors are primarily driven by public policies and allocation of local government funds). This outcome highlights the importance of considering all levels of the social ecological model, including public policy and urban design planning, when developing PA promotion interventions.

The lack of significant associations between the 6 neighborhood environmental factors and engagement in some versus no transportation PA in logistic models may be related to transportation PA being a necessity for daily activities rather an option (ie, participants may have had no other type of transportation). However, we did not include survey items asking participants about their primary mode of transportation, which is needed to draw a more precise conclusion on this outcome. Likewise, the lack of associations between social cohesion, violence, and activities with neighbors and PA was reported previously in a study among Latinx residing in Massachusetts [40], suggesting that these 3 factors may not be key determinants of PA engagement among Latinas. Future studies with larger sample sizes are needed before definitive conclusions can be drawn.

Limitations of our study include the use of a relatively small sample from a single metropolitan geographic region. The sample size was determined based on available funds, and the geographic region chosen was based on the population that the subsequent intervention would be developed to target. The small sample limited our ability to conduct additional subgroup analyses to further examine associations among duration of US residency (ie, ≤5 years versus >5 years), neighborhood environmental factors, and PA. Further examination of neighborhood characteristics and PA patterns based on duration of residency status in the United States would be interesting to explore in future work. Likewise, given that our sample comprised women residing in urban and suburban areas of metropolitan Phoenix, the study findings should not be generalized to residents residing in rural areas. Additional research with a larger, more geographically diverse sample is warranted to confirm and expand on our findings. Another limitation was that we relied exclusively on self-report measures to assess PA and neighborhood characteristics. Including an objective PA measure would have provided further context regarding the PA patterns of our sample. However, this was not possible because of the limited resources of the study. Likewise, preliminary data analyses resulted in 11 women being excluded from the study based on IPAQ scoring guidelines for either reporting nonplausible PA data (n=10) or providing incomplete data (n=1). Sensitivity analyses showed that these participants reported lower education levels than participants who provided valid data. This suggests that the IPAQ may not be appropriate for Latinas of lower education levels, further limiting generalization of the study findings. Additionally, perceived neighborhood characteristics may differ from objectively measured characteristics (eg, neighborhood characteristics measured using geographic information systems, local crime/violence data). Future research among first-generation Latinas should consider the use of both subjective and objective PA measures and neighborhood measures, as both types of measures provide important insights into the PA patterns and neighborhood characteristics of Latinas. A final limitation of this study is that it did not include items assessing BMI or health status. Such information would have provided additional insight on the generalizability of the study findings.

Despite its limitations, the current study has several strengths. First, this is one of few studies to examine domain-specific PA patterns among first-generation Latina immigrants, as most PA studies among Latinas have only included a measure of leisure-time PA [34] and failed to differentiate by generational status. Another strength of the study is that our sample was comprised of women from diverse countries of origin (ie, 69% from Mexico and the remainder from various Central and South American countries). The composition of our sample likely reflects the changing Latinx immigrant population in the United States (ie, immigration from Central and South American countries has increased in recent years, with immigration from Mexico slowly declining) [11].

Findings suggest that many first-generation Latinas could benefit from a leisure-time PA intervention. Given that leisure-time PA has pronounced benefits for promoting positive health outcomes, intervention efforts targeting leisure-time PA may be an effective method for researchers and public health professionals to reduce obesity and diabetes health disparities among Latina immigrants. Such interventions should consider the neighborhood environments in which first-generation Latinas reside, as these factors will likely influence the types of PA promoted in an intervention. Results of this study will be used to inform development of a culturally tailored PA intervention for the reduction of metabolic disease risk among first-generation Latina immigrants.