Background: Ecological models posit that interactions among factors at multiple levels (e.g., individual, psychosocial, and environmental) influence physical activity (PA). However, interactions involving environmental factors are the least understood. Intervention studies also suggest PA behavior change may depend on the environments in which participants are encouraged to be active but such evidence from interventions targeting Latinos is limited.

Methods: Chapter 2 tested interactions between neighborhood environmental and socio-demographic factors on total moderate-to vigorous physical activity (MVPA) among an international adult sample (N=10,258). Chapter 3 examined interactions between neighborhood environmental and psychosocial factors on context-specific PA among an adolescent sample from the Baltimore/Seattle regions (N=910). Chapter 4 assessed the moderating effects of neighborhood environmental factors on the impacts of a physical activity intervention targeting Latinas in San Diego, CA at 12-months post-intervention (N=319).

Results: Chapter 2 found moderating effects by education and gender on the association between safety from crime and meeting high PA levels, with inverse associations found only among the high-education group and men. Education and gender moderated associations of safety from crime and the presence of transit stops with meeting minimum PA guidelines, with positive associations found for safety from crime only among women and the presence of transit stops only among men and the high-education group. Chapter 3 found moderating effects by decisional balance on the association between recreation facility density and neighborhood leisure-time PA among female adolescents, with a negative association found only among those with high decisional balance. Decisional balance also moderated the associations of neighborhood walkability with non-school MVPA among females and non-neighborhood leisure-time PA among males, with positive associations only among adolescents with high decisional balance. Chapter 4 showed higher total MVPA and leisure-time MVPA at 12-months post-intervention among participants with favorable perceived neighborhood aesthetics than those with less favorable evaluations.

Conclusions: This dissertation supports the ecological models’ hypothesized interactions between environmental and individual/psychosocial factors on PA and the moderating role of the neighborhood environment on the impacts of a multilevel PA intervention targeting Latinas. Findings support global efforts targeting multiple levels of influence to promote population PA, health, and environmental sustainability.

Background: Many adult Latinos are not meeting guidelines for healthy eating. Theory posits the important role of family and family members on individual’s health behaviors (e.g., diet). However, few studies examine how children impact their parents’ diet. Furthermore, empirical evidence shows that acculturation is an important predictor of diet among Latino adults. The purpose of this dissertation was to examine the role of children, families, and acculturation on mothers’ dietary intake and related behaviors.

Methods: Chapters 1 and 2 used secondary data from mother-child dyads enrolled in a randomized controlled trial (RCT) and Chapter 3 used primary data collected from mother-child dyads. Chapter 1 used baseline survey data from 314 Latino mother-child dyads to test how children’s acculturation and the mother-child acculturation gap was associated with mothers’ dietary intake and related behaviors. Building on the results from Chapter 1, in Chapter 2, longitudinal data from 162 dyads enrolled in delayed treatment group of the RCT were used to test the temporal relations of the family environment on traditional mothers’ dietary intake and behaviors, and whether these relations differed between mothers of assimilated versus bicultural children. Finally, Chapter 3 used qualitative interview data from mothers and their bicultural (n=11) or assimilated (n=10) children and quantitative data from mothers to explore how children influenced their mothers’ dietary intake and behaviors.

Results: Chapter 1 provided evidence that accounting for mothers’ acculturation, having an assimilated versus a bicultural child was negatively associated with mothers’ vegetable intake and positively associated with mothers’ sugary beverage intake, percent of calories from fat, and frequency of away-from-home eating. The most at-risk dyads for lower quality diet among mothers were traditional mothers of assimilated children. Chapter 2 found that less positive family interactions around food at baseline predicted more frequent away-from-home eating four months later among mothers of assimilated children. Additionally, more family expressiveness at four months predicted more dollars spent on fruits and vegetables at ten months among mothers of bicultural children. In Chapter 3, participants described how mothers’ feeding styles shaped children’s food preferences and in return, children’s food preferences influenced mothers’ dietary intake and behaviors.

Discussion: This dissertation found evidence for the important role of children and families on mothers’ dietary intake and related behaviors. The findings presented here support the theoretical notion that individuals’ health behaviors are directly and indirectly influenced by family-level factors. Further, this dissertation extends previous research on the important role of mothers on their children’s dietary outcomes by building evidence for the reciprocating influence of children on their mothers’ dietary intake and behaviors.

Background: This dissertation aimed to add to the limited body of research on the subjects of PA maintenance, lapses in PA, and relapse into inactivity among Latina participants in PA interventions.

Methods: Study 1 used survival analysis methods to determine the probability of lapses in PA and to identify predictors of time until lapse, among Latina participants in two PA interventions. Study 2 used qualitative methods (i.e., semi-structured individual interviews with 21 Latina women) to (1) identify facilitators and barriers to PA maintenance; (2) examine similarities and differences between those who maintained PA and those who did not; and (3) create a conceptual model to understand determinants of PA maintenance. Study 3 was a pilot randomized controlled trial (N=27) that used a nested analysis of quantitative and qualitative data to (1) assess maintenance of PA among Latina women who completed a PA intervention; (2) understand whether participants were satisfied or dissatisfied with their maintenance; (3) evaluate the feasibility and preliminary efficacy of smartphone apps for PA maintenance; and (4), explore the role that smartphone apps played in PA maintenance (if any).

Results: Study 1 revealed predictors of earlier lapse, including 2-month scores of Self-Efficacy and various Processes of Change. With one exception, environmental and individual factors were not predictive of earlier lapse. Study 2 identified potential facilitators of PA maintenance: habit, ability to overcome obstacles, specific sources of motivation, having knowledge and skills to maintain PA, and a supportive social and built environment. In study 3, mean minutes/week of self-reported moderate to vigorous PA (MVPA) was 16.85 (SD=24.73) before the intervention, 230.50 (SD=199.29) immediately after the intervention, and 163.10 (SD=125.49) after the maintenance period. There were significant effects of time, but not of group assignment, on self-reported PA. Nine (64%) enhanced maintenance participants reported using apps for PA maintenance. Reasons for not using apps included not finding them appealing or necessary.

Conclusions: Lapse and relapse into inactivity were not pre-determined by individual or environmental factors; skills needed to prevent such events and achieve long-term maintenance of PA outcomes may be taught and need to be incorporated into future PA interventions.

The rapid expansion of direct-to-consumer wearable fitness products (eg, Flex 2, Fitbit) and research-grade sensors (eg, SenseCam, Microsoft Research; activPAL, PAL Technologies) coincides with new opportunities for biomedical and behavioral researchers. Underserved communities report among the highest rates of chronic disease and could benefit from mobile technologies designed to facilitate awareness of health behaviors. However, new and nuanced ethical issues are introduced with new technologies, which are challenging both institutional review boards (IRBs) and researchers alike. Given the potential benefits of such technologies, ethical and regulatory concerns must be carefully considered.

Our aim was to understand potential barriers to using wearable sensors among members of Latino, Somali and Native Hawaiian Pacific Islander (NHPI) communities. These ethnic groups report high rates of disparate health conditions and could benefit from wearable technologies that translate the connection between physical activity and desired health outcomes. Moreover, these groups are traditionally under-represented in biomedical research.

We independently conducted formative research with individuals from southern California, who identified as Latino, Somali, or Native Hawaiian Pacific Islander (NHPI). Data collection methods included survey (NHPI), interview (Latino), and focus group (Somali) with analysis focusing on cross-cutting themes.

The results pointed to gaps in informed consent, challenges to data management (ie, participant privacy, data confidentiality, and data sharing conventions), social implications (ie, unwanted attention), and legal risks (ie, potential deportation).

Results shed light on concerns that may escalate the digital divide. Recommendations include suggestions for researchers and IRBs to collaborate with a goal of developing meaningful and ethical practices that are responsive to diverse research participants who can benefit from technology-enabled research methods.

ClinicalTrials.gov NCT02505165; https://clinicaltrials.gov/ct2/show/NCT02505165 (Archived by WebCite at http://www.Webcitation.org/6r9ZSUgoT)

Background

Purpose

Methods

Results

Conclusion

Background

Methods

Results

Conclusions

Impact