Physical Activity Intervention Effects on Sedentary Time in Spanish-Speaking Latinas

Abstract

Background:

Latinas have high rates of sedentary behavior and related health disparities, but it is unknown if interventions to increase physical activity will also reduce sedentary time. The current study examined changes in objectively measured sedentary time among Latinas in a randomized controlled trial of a physical activity intervention.

Methods:

Spanish-speaking Latinas (N = 202) were randomized to an exercise or wellness group and wore an accelerometer at baseline, 6 months, and 12 months.

Results:

Participants were sedentary on an average of 8.86 hours per day (SD = 2.60) at baseline. The intervention group had significantly greater increases in sedentary time compared with the control group, with the intervention group engaging in 146 more minutes per week of sedentary time at 6 months and 254 minutes per week of sedentary time at 12 months than the control group (P = .02). The intervention effect on sedentary behavior remained after controlling for moderate to vigorous physical activity. Additionally, time spent in moderate to vigorous physical activity was positively associated with more sedentary time (P = .04).

Conclusion:

An intervention to increase moderate to vigorous physical activity resulted in greater sedentary time, raising concerns regarding compensation and highlighting the need for interventions to address both physical activity and sedentary behavior to improve public health.

Past research has found associations between greater time spent in sedentary behaviors and increased risk for mortality and poor health outcomes.¹ Findings from prospective studies have linked sedentary behavior to increased risk of developing obesity, type 2 diabetes, and hypertension, as well as cardiovascular events.^2-5 While many of these studies relied on self-report data (eg, time spent sitting, watching TV, etc), objectively measured sedentary time (<100 accelerometer counts per minute) has also been positively correlated with waist circumference, 2-hour glucose, and carotid artery thickness.^6-8 Several large epidemiological studies have found a dose–response relationship between daily sitting time and all-cause and cardiovascular disease mortality at long-term (6–12 y) follow-ups, including the Canada Fitness Survey,⁹ the Australian Diabetes, Obesity and Lifestyle Study,¹⁰ and the European Prospective Investigation of Cancer—Norfolk Study.¹¹

Questions remain regarding whether sedentary behaviors are distinct from overall physical activity. In response, some studies examining the association between sedentary behavior and mortality have included physical activity in the analyses as a covariate and found their results largely unchanged.^9-13 Other studies stratified their analyses by physical activity levels or included interaction terms in their statistical models.^9-11 Clear associations between levels of sitting and mortality risk in both the physically active and inactive men and women were found in the Canada Fitness Survey study,⁹ which supports the independence of the effects of such sedentary behavior from overall physical activity levels. However, more recent studies have found that the relationship between sedentary behavior and mortality (or other chronic disease outcomes) is either attenuated or completely eliminated after accounting for physical activity.^14-18

Despite some mixed findings in the literature, sedentary time remains an important intervention target as American adults spend most of their day in this activity category.^19,20 However, intervention research has more commonly focused strictly on promoting moderate–vigorous intensity physical activity. A meta-analysis of randomized trials found that interventions promoting physical activity or even those including both a physical activity and sedentary behavior component generally result in only (inconsistent) small reductions in sedentary time.²¹ As both increased physical activity and decreased sedentary time are likely critical to improving public health,²⁰ more research is needed to determine whether these interventions have the capacity to address the current epidemic of sedentary lifestyle or if perhaps a more concentrated approach to sedentary behavior will be required.

In addition, there has been little examination of sedentary behavior outcomes among underserved populations such as Latinas,²¹ despite the particularly high rates of sedentary behavior and related health disparities in this community. In fact, objectively measured data from a large population-based study of US Latino adults found that 74% of their time was spent in sedentary activities. Furthermore, results indicated an adverse relationship between such sedentary time and cardiometabolic biomarkers (eg, decreased high-density lipoprotein cholesterol, increased triglycerides, 2-h glucose, and fasting insulin), independent of physical activity.²² Similar to concerns in the general population, physical activity interventions have shown to be quite efficacious for increasing activity levels in Latinas ^23-29; however, it is unknown whether targeting physical activity in this group will be sufficient to produce meaningful reductions in sedentary time. Thus, the current study explored secondary outcomes of changes in sedentary time among Spanish-speaking Latinas in a randomized controlled trial of a 12-month physical activity intervention.

Methods

Study Design and Sample

The Pasos Hacia La Salud study was a randomized controlled trial of an Internet-based physical activity intervention for Spanish-speaking Latinas compared with a contact control arm. No content related to sedentary behavior was included in either arm. The intervention was 6 months followed by a maintenance phase from 6 to 12 months (Marcus et al^27,28). Data were collected at the University of California, San Diego between 2011 and 2014, and the study protocol was approved by the institutional review board of the University of California, San Diego.

Eligible participants were women who self-identified as Hispanic or Latina (or of a group defined as Hispanic/Latina by the Census Bureau), were able to read Spanish fluently (defined as scoring above the “inadequate” range on the Short Test of Functional Health Literacy in Adults), and self-reported insufficient physical activity (defined as reporting less than 60 min/wk of moderate to vigorous physical activity [MVPA] on the 7-Day Physical Activity Recall). Additional eligibility included 18 to 65 years of age, body mass index (BMI) < 45 kg/m², and regular access to an Internet-connected computer. Exclusion criteria included medical condition or taking medication that would make unsupervised physical activity unsafe, currently pregnant or planning to be pregnant in the next year, and planning to move from the area within the next year.

Protocol

Study protocol and participant recruitment details have previously been published (Marcus et al²⁸). Briefly, participants were recruited through several methods including online and print advertisements, participant referrals, and mailed and e-mailed study information through primary care doctor offices. Potential participants were first screened over the phone for eligibility, then attended an orientation session, and provided written informed consent. Participants returned for a measurement visit during which height and weight were measured, and an ActiGraph GT3X+ accelerometer (ActiGraph Corp, Pensacola, FL) was distributed with instructions to wear it during waking hours for 7 consecutive days. One week later, participants returned the ActiGraph, completed the remaining baseline measures, and then were randomly assigned to one of 2 groups: Tailored Physical Activity Internet Intervention or Wellness Contact Control Internet Group. Group assignment was determined using a permuted block randomization procedure, with small random sized blocks.

A total of 838 individuals expressed interest in participation. Of these, 258 did not meet inclusion criteria, 333 declined to participate, 25 failed to complete the screener, and 4 were unable to be scheduled for an orientation. A total of 205 eligible women were randomly assigned to the intervention (n = 104) and control (n = 101) groups. Of these, 3 participants did not complete the baseline ActiGraph measure; therefore, 202 participants were included in the current analyses. Full consort diagram is available in Marcus et al.²⁷

Tailored Physical Activity Internet Intervention (Intervention Group)

Participants in the intervention arm set personalized physical activity goals and received access and orientation to the intervention website. Features on the website included (1) self-monitoring of minutes of activity and steps, (2) graphs to compare personalized goal to reported activity, (3) message board to foster social support between participants, (4) “ask the expert” where participants could anonymously ask questions to a PhD-level researcher, and (5) online resources such as maps to create walking routes and free exercise videos. Each month, participants completed questionnaires through the website that generated individually tailored physical activity reports. These reports included information regarding: (1) current stage of motivational readiness for physical activity; (2) self-efficacy; (3) cognitive and behavioral strategies associated with physical activity (processes of change); (4) how the participant compares to individuals who are physically active and meeting national guidelines of 150 minutes per week of MVPA (normative feedback); (5) how the participant compares to her prior responses (progress feedback); and (6) useful facts about physical activity, such as health benefits, stretching, and heart rate monitoring. In addition, they received an online physical activity manual that was matched to their motivational readiness for physical activity behavior change. E-mail prompts to access the intervention website were sent weekly during month 1, biweekly during months 2 and 3, monthly during months 4 to 6, and every other month during months 7 to 12. They also received 2 brief phone calls during the first 6 months and 1 brief call at 9 months to review progress and problem solve around activity goals. No content about sedentary behavior was included in the intervention.

Wellness Contact Control Internet Group (Control Group)

The Wellness Contact Control Internet Group received access to a Spanish language website with information on health topics other than physical activity. The web-based content focused on diet and other factors associated with cardiovascular disease risk and included information from a series on heart health developed for Latinos by the National Heart, Lung, and Blood Institute. Sedentary behavior was not included as a topic. Control group participants also completed online questionnaires on wellness topics (other than physical activity) on the same schedule as the intervention arm.

Measures

Sedentary time and physical activity were measured with a hip-worn accelerometer (ActiGraph GT3X+) at baseline, 6 months, and 12 months. ActiLife software (version 6.3.4; ActiGraph Corp) was used to screen for sufficient wear time using the Choi et al³⁰ guidelines with sufficient wear time defined as 5 days with ≥600 minutes of wear time or 3000 minutes across 4 days. Nonwear time was defined as intervals of at least 60 consecutive minutes of zero counts, with allowance for up to 2 minutes of observations of <100 counts per minute within the nonwear interval. All complete and valid data were processed in ActiLife using the low-frequency extension and aggregated to 60-second epochs, so activity and sedentary cut points could be applied.³¹ We relied on established cut points to classify sedentary behaviors from accelerometer data. As such, a threshold of 100 counts per minute on the x-axis defined sedentary activities.¹⁹ Time spent per day in sedentary activities was calculated by summing the minutes in a day where the counts were below 100 counts per minute. We averaged day-level totals across measurement days for each participant to yield the average daily time spent sedentary. To determine the time spent in MVPA (activity at 3 metabolic equivalents or higher, eg, brisk walk or faster), established cut points for accelerometer data were used by summing every minute in a day where the x-axis counts were 1952 or above.³¹ Day-level averages for time spent in MVPA were then computed for each participant.

Questionnaires at baseline assessed demographics and acculturation using the Brief Acculturation Scale. The Brief Acculturation Scale is a 4-item measure that asks about language use across different life contexts (ie, at home, with friends) with higher scores indicating a higher level of acculturation.³²

Statistical Analyses

Baseline demographics, physical activity, and sedentary behavior were summarized using descriptive statistics and compared between groups using t tests (for continuous variables), chi-squared tests (for categorical variables), and nonparametric tests as appropriate.

Using a longitudinal mixed-effects model with a random intercept and slope, we simultaneously examined intervention effects on mean sedentary time at follow-up (6 and 12 mo), controlling for baseline and wear time. Models included intervention effects, time effects, and the interaction between time and intervention. Linear and quadratic functions of time were examined, and the final model was chosen to maximize model fit. Models included subject-specific intercepts in order to adjust for repeated measures within participants. A subsequent model was examined which also adjusted for the time-varying physical activity variable (minutes per week of objectively measured MVPA). Specifically, we regressed mean sedentary time at follow-ups on baseline sedentary time, wear time, intervention group, as well as a time-indexed variable of minute per week of MVPA. Models of 6-month sedentary time were adjusted for MVPA at 6 months, and models of sedentary time at 12 months were adjusted for MVPA at 12 months.

As a subsequent step, we considered sedentary bouts as a secondary outcome and tested associations between intervention group and longitudinal objectively measured bouts, adjusting for wear time and baseline activity. Longitudinal mixed-effects models included subject-specific intercepts to adjust for the highly correlated data within participant over time.

Moderators of the intervention effects on total sedentary time were tested using a similar analytic approach to that described previously. Models included the main effect of the posited moderator as well as the interaction between intervention and the moderator. A variable was considered to be a moderator if the interaction was statistically different than zero. Posited moderators were identified a priori and included age, employment, income, baseline minutes of physical activity, self-efficacy, acculturation, and BMI.

Analysis was conducted on the sample of participants with valid ActiGraph data at baseline (N = 202). Models used a likelihood-based approach to estimation and thus made use of available data without directly imputing missing outcomes. All analyses were carried out in SAS (version 9.3; SAS Institute Inc, Cary, NC), and alpha was set at .05 a priori.

Results

Participants were 202 Latinas who were 39.18 years old on average (SD = 10.33). At least 60.7% had some college-level education and about 65.8% reported less than $30,000 annual household income. Participants predominantly identified as having a Mexican background (84.2%), being first generation in the United States (82.1%), and speaking only Spanish or more Spanish than English in the home (68.8%). BMI was on average in the overweight category (M = 28.90 kg/m²) and ranged from 19.55 to 43.12 kg/m² (see Table 1). Of the 202 participants, 184 (84%) had ActiGraph data at 6 months and 155 (76.7%) had ActiGraph data at 12 months. Those who did not provide ActiGraph data at 6 or 12 months were not significantly different in any baseline characteristics from those who did. Participants wore the ActiGraph on average 6.73 days at baseline (SD = 1.45), 6.26 days at 6 months (SD = 1.21), and 6.20 days (SD = 1.22) at 12 months. As previously published, compared with the control group, participants in the intervention arm had greater ActiGraph-measured MVPA in 10-minute bouts at 6 and 12 months (P = .01).^24,27 The intervention arm engaged in 35.77 minutes per week of MVPA (SD = 69.65) at baseline, 75.82 minutes per week of MVPA (SD = 91.00) at 6 months, and 70.38 minutes per week of MVPA (SD = 86.40) at 12 months. The control arm engaged in 28.67 minutes per week of MVPA (SD = 48.22) at baseline, 43.02 minutes per week of MVPA (SD = 60.90) at 6 months, and 55.51 minutes per week of MVPA (SD = 74.60) at 12 months.^24,27

Table 1.

Baseline Characteristics by Study Arm (N = 202)

	Intervention group (n = 101)	Control group (n = 101)
Age, mean (SD), y	38.8 (10.3)	39.6 (10.4)
Latina	100%	100%
White	45%	58%
Mexican	82%	86%
First generation in United States	87%	77%
Household income <$30,000	68%	63%
Some college-level education or more	55%	66%
Only Spanish spoken in home	39%	35%
Body mass index, mean (SD), kg/m2	29.2 (5.8)	28.6 (4.5)
Hours spent sedentary per week, mean (SD)	64.2 (16.5)	59.8 (19.6)

At baseline, participants were sedentary on an average of 62.03 hours per week (SD = 18.18) or 8.86 hours per day (SD = 2.60). Results showed significant intervention effects such that those randomized to the intervention arm had significantly higher mean minutes of objectively measured sedentary time at 6 and 12 months, P < .05. Compared with the control group, participants in the intervention arm were sedentary for 146 more minutes per week at 6 months and 254 minutes per week at 12 months (P = .02). As a subsequent step, we added MVPA into the model as the intervention group had significantly greater increases in minutes per week of MVPA than the control group at 6 months (mean differences = 31.0, SE = 10.7, P < .01) and at 12 months (mean differences = 11.47, SE = 3.19, P = .01). The intervention group spent significantly more minutes being sedentary than the control group even after controlling for time in MVPA. Furthermore, among the aggregated sample of participants, higher MVPA at 6 months was associated with higher sedentary time at 6 months, controlling for baseline (P = .04). A similar pattern of findings was seen at 12 months. Adjusted models are presented in Table 2, and unadjusted means of weekly hours in sedentary time are presented in Figure 1.

Table 2.

Adjusted Effects of Intervention and Moderate to Vigorous Physical Activity on Total Sedentary Time and Sedentary Time Accumulated in Bouts

	b (SE)	P
Total sedentary time
Intervention
6 mo	146.78 (83.88)	.02
12 mo	254.35 (89.34)	.02
Moderate to vigorous physical activity
6 mo	0.25 (0.09)	.04
12 mo	1.98 (0.86)	.04
Sedentary time accumulated in 10-min bouts Intervention
6 mo	47.26 (18.21)	.04
12 mo	115.78 (92.40)	.04
Moderate to vigorous physical activity
6 mo	0.29 (0.07)	.049
12 mo	0.17 (0.06)	.046

Note: b is unstandardized effect. Models adjust for time-varying indicator of wear time and include subject-specific intercepts.

Figure 1 —

Unadjusted means of ActiGraph measured sedentary hours per week (N = 202).

Results were similar for time spent in sedentary bouts of at least 10 minutes in duration (see Table 2). There was a significant intervention effect such that those randomized to the intervention arm had significantly higher mean minutes of sedentary bouts (P = .04). The intervention group spent significantly more minutes in sedentary bouts than the control group after controlling for time in MVPA. More time spent in MVPA was positively associated with more sedentary time in sedentary bouts (P = .05).

None of the examined baseline variables, age, employment, income, baseline minutes of physical activity, self-efficacy, acculturation, and BMI, were significant moderators of the intervention effect on sedentary time (all Ps > .05).

Discussion

This is the first trial with Spanish-speaking Latinas to examine how a physical activity intervention impacted sedentary behavior. Although participants were highly sedentary, engaging in almost 9 hours of sedentary behavior per day at baseline, the intervention arm still significantly increased their total sedentary time and sedentary bouts compared with the control arm at 6 and 12 months. The amount of time intervention participants increased their sedentary time was much larger than the amount of time that they increased their MVPA. In addition, greater MVPA was associated with greater sedentary time. This suggests that as Latina women became more physically active, they engaged in compensatory resting behaviors. One possible explanation for such findings is the “ActivityStat” hypothesis, which posits that people maintain an overall physical activity set point by balancing increased physical activity in one part of the day with decreased physical activity in another part of the day.³³ Qualitative methods were helpful in identifying several potential causes for physical activity compensation in past studies in older adults (fatigue, muscle soreness, fear of overexertion, lack of motivation, and time)³⁴ and similar investigations are needed in Latinas.

Findings from the current study highlight the importance of physical activity interventions also addressing sedentary behavior in this at risk target population; however, results are inconsistent with much of the previous literature that has found physical activity interventions either have little impact on sedentary time or result in small decreases in sedentary time. For example, a meta-analysis of 9 physical activity intervention studies did not find that the interventions resulted in significant change in sedentary time.³⁵ The meta-analysis also looked at sex differences and found significant intervention effects for reduced sedentary time among men-only studies, but not women-only studies. One reason for the lack of consistency with previous research may be that the current study only included Spanish-speaking Latinas. Martin et al’s³⁵ meta-analysis showed that there were sex differences but did not look at race/ethnicity differences. In addition, much of the previous research has been conducted in predominantly non-Hispanic white populations and was of short duration, with only one intervention longer than 6 months. A systematic review on overall energy expenditure found energy compensation in response to exercise, and that compensation was greater for longer interventions.³⁶ Interventions of longer duration that also consider overall energy expenditure, including dietary factors, may be important for understanding the compensatory process more fully. The current study advances the existing literatures by demonstrating that Latinas may engage in more compensatory behaviors than other populations when engaging in physical activity and highlights the importance of including diverse and underrepresented populations in research. With the increasing knowledge of the negative impact of extensive sedentary behavior, it is important to understand who is at risk for engaging in unhealthy compensatory behaviors as they attempt to increase their physical activity.

While a strength of this study is its focus on Spanish-speaking Latinas, a limitation is that we cannot examine whether the increase in sedentary time is related to the intervention itself or if it is related to demographic characteristics of the participants. Future studies should investigate race/ethnicity and sex differences to know who is at risk for increasing sedentary time when increasing MVPA. A focus on overall energy expenditure, including diet, may help to advance our knowledge on factors that lead to compensatory behaviors. Another strength is the use of objective measures of sedentary behavior and MVPA, which are less prone to recall and response biases than traditional self-report approaches. However, the ActiGraph cut points used to define sedentary behavior do not differentiate between standing still and seated postures.³⁷ Therefore, standing still may have been classified as a sedentary behavior. Future studies using devices with inclinometers may be able to more accurately distinguish between seated and standing postures.³⁸

Conclusion

In summary, the results of the present study demonstrate that among Spanish-speaking Latinas, an intervention to increase MVPA may inadvertently result in increasing sedentary time. To our knowledge, this was the first study to examine changes to objectively measured sedentary time among an understudied population enrolled in a physical activity intervention. This study shows that both physical activity and sedentary behavior should be targeted in interventions to improve public health.