1 Introduction

Mexican Americans are documented as being among the least active race/ethnic groups in the United States, most notably during leisure time [1, 2]. Nearly a third of Mexican Americans do not engage in leisure time physical activity (32.1%), compared to 23% of non-Hispanic whites [2]. Some research suggests that despite not engaging in regular leisure-time physical activity, Mexican Americans are more likely to hold non-sedentary occupations and may be more physically active through work activities [3]. Still, physical activity interventions that target this group are designed to improve leisure time activity, not work-based activity. This leads to observed disparities in physical activity engagement and leaves a knowledge gap in evidence-based strategies that may be more effective in improving physical activity engagement [4].

Work-based lifestyle interventions are widely tested and evidence-based to improve physical activity engagement in non-Hispanic groups [5,6,7,8,9,10,11,12,13,14,15,16,17]. Strategies have included yoga, general physical activity, and walking to increase work-time physical activity [7, 18,19,20]. By design, work-based lifestyle interventions are short-term, and documentation of the long-term benefits is mixed [7, 13]. While some evidence suggests that work-based lifestyle interventions lead to sustained physical activity engagement, follow-up is usually within a year [7, 13].

An evidence-based strategy that addresses shortcomings related to study design and follow-up in physical activity is adding 'boosts' during the maintenance phase of the intervention [21, 22]. Boosting generally includes mailed education during the maintenance period, web-based prompting or reminders, and phone-based motivational interview contacts [21,22,23]. Seamos Saludables is one example of a tailored physical activity intervention to improve physical activity among Latinas [24] that incorporated boosting into the study design. During a 6-month maintenance phase, participants received additional mailed education and reminders. Latinas who received the intervention maintained their physical activity engagement at 12 months, suggesting that boosters may facilitate sustained engagement among Latinas [23, 24].

The Border Coalition for Fitness (BCF) is a university-community-based coalition to improve park use and physical activity engagement in El Paso, Texas. El Paso is predominantly Mexican American and highly employed [25]. Through the BCF, a community-wide walking campaign was established to improve physical activity engagement. The BCF has offered three walking challenges per year since 2018. The challenges are targeted toward employers. Since their inception, more than 5000 people from multiple employers, including universities, public schools, and the City of El Paso, have participated in the challenges [26].

While walking challenges are established evidence-based strategies to increase physical activity engagement [10, 13], the BCF community-wide campaigns are unique because they are offered multiple times per year each year [26]. Additionally, employer representation cuts across the city, county, schools, and private employers, providing a variety of work settings to engage participants. Due to the repeated nature of the campaigns offered each year, the walking challenges provided an opportunity to assess multiple challenge participation, as repeated exposures, in place of a boost for sustained increased walking engagement. This design differs from the typical one-year follow-up design, which was customary in other studies. Typically, a booster will rely on cognitive processes to make a participant alter their behavior; however, in this study, the repeated exposure operated behaviorally by providing social reinforcement for new walking challenge participants.

This paper leverages program evaluation data to assess the relationship between multiple walking challenge participation and walking engagement among Mexican Americans in El Paso, Texas. Two research questions guide the analysis for this paper. First, is repeated challenge participation associated with higher average walking engagement at baseline and after challenge completion? Second, are there quantifiable differences in step averages by the number of previous challenges?

2 Methods

2.1 Participants

Study participants were recruited from walking challenges created by the BCF during registration. The three walking challenges created and used by the BCF included the Walk-the-Walk Team Challenge, Step Bootcamp, and 10,000 Steps for 100 Days. The Walk-the-Walk Team Challenge offered a group-based experience that provided social support for physical activity as a cultural element [26]. In this challenge, teams of 10 participated in a month-long competition. The Step Bootcamp taught participants of all skill levels how to achieve their full potential in every day steps over a 9-week challenge [26]. The goal for this challenge was to get participants to walk three times their baseline average by the end of the challenge [26]. The 10,000 Steps for 100 Days challenge, which ran from October 1 to January 9, required individuals to walk 10,000 steps daily for 100 days (more than 3 months) [26].

As part of the online registration form, participants were asked if they would participate in an evaluation study of the walking challenges. If they indicated 'yes,' they received a study invitation by email. Participants who clicked on the link would get directed to a REDCap survey, where they would complete a consent form and a brief survey. Study data were collected and managed using REDCap electronic data capture tools hosted at Texas Tech University Health Sciences Center El Paso [27, 28]. REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies, providing (1) an intuitive interface for validated data capture; (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages; and (4) procedures for data integration and interoperability with external sources. Participants were also asked to provide screenshots of their tracking device summaries for the previous week. Participants were then automatically sent follow-up emails 2 weeks post-challenge completion. At follow-up, participants were asked to upload their step-tracking device screenshots online. Screenshot values would be inputted in separate fields as numeric values by the research staff. A total of 354 participant records were analyzed.

3 Variable measurement

3.1 Primary outcome variable

Steps per day Participants reported steps tracked using a wrist-based pedometer or smartphone. During the study, the number of steps was determined using tracking device screenshots and collected at baseline and 2-weeks. We averaged the 7 days prior to baseline and 7 days prior at 2 weeks. The accuracy of both wrist-based pedometers and step-recording smartphones has been validated previously [6, 7, 14, 21]. For this study, we did not require the participants to record the tracking technology used to track their steps.

Previous challenge participation Previous challenge participation was measured as yes/no to whether participants participated in any prior walking challenges sponsored by this program. Participants were able to check off as many as applied to them. Since BCF maintains records for all their challenges and challenge participants, only walking challenges provided by BCF were included in the list of previous walking challenges.

PA self-assessed engagement Non-exercise test of physical activity was measured using the National Aeronautics and Space Administration/Johnson Space Center Self-reported Physical Activity Scale (NASA SR-PA) [29]. The NASA SR-PA has undergone cross-validation studies and received high praise as a measure of cardiorespiratory fitness (CRF). The tool provides five options to describe their current physical activity engagement. Level 1 is the least engaged and is described as being "inactive." Level 5 is described as "participates in aerobic exercises over 3 h per week."

Demographic Covariates Sociodemographic information was collected at baseline. Variables included for this study include; age, married (yes/no), high school vs. college graduate, and employed (yes/no).

3.2 Analysis

Step totals for each time point (baseline, 2 weeks) were used to calculate means and percent change throughout the follow-up period. Frequencies, cross-tabulations, and means were first conducted to assess for regularity in the data, outliers, and other inconsistencies that may affect the analysis. Categories of some variables were collapsed into smaller categories in cases where there were small numbers or made analytical sense to collapse (i.e., professional vs. non-professional). Regression analysis (OLS) was then conducted to predict average steps at baseline and 2 weeks post-intervention by previous challenge participation. Models were adjusted for demographic characteristics and the current walking challenge. An additional model was run using 2-week step outcomes, adjusting for the baseline step average. Multi-logit regression was conducted to predict previous challenge participation. The first model was adjusted for the current challenge and demographic characteristics. Model 2 added self-assessed physical activity level. All the statistical analyses were done by using STATA 16 SE statistical software. P < 0.05 was considered statistically significant.

4 Results

Table 1 presents the baseline characteristics of the study sample. On average, participants were about 40 (41.5 years), married (60.4%), and college-educated (65%). An equal number of professionals were employed in professional jobs compared to non-professionals; the majority were enrolled in the 10,000 steps for 100 days challenge (39.1%) and had never participated in a previous challenge (77.5%). Non-professionals tended to walk 500 steps per day more at baseline than professionals (7601 vs. 8169 steps), college-educated more than high school or less (8074 vs. 7535 steps), and Step Boot Camp participants the least (6482 steps). The average number of steps taken by previous challenge participants was higher than that of first-time participants, regardless of how many challenges they had previously participated in (7569 steps vs. 8959 (1 to 3 challenges). Three or more challenge participants had the greatest number of average steps (9446). Self-assessed PA levels were close to evenly distributed from inactive to engaged in aerobic activity three or more hours per week. The inactive group had the lowest baseline average steps (5767 steps), and the aerobic 3 h or more a week with the highest (10,114 steps), nearly double that of the inactive.

Table 1 Participants' baseline characteristics and average steps per week
Full size table

We conducted a regression analysis to determine predictors of step averages by demographic characteristics, previous challenge participation, and current challenge (see Table 2). Previous challenge participants had, on average, 1326 (p = 0.01) more steps at baseline than new participants after adjusting for demographic characteristics. At 2 weeks post-challenge, this advantage was slightly attenuated by 103.7 steps (1223, p = 0.09) to near significance, Suggesting in part that first-time challenge participants narrow their disadvantage through challenge participation. Any significant advantage of prior participation at 2 weeks was reduced to insignificance when controlling for average baseline steps. This suggested that previous participants' advantage may be due to the benefit of multiple challenge participation and/or the selection of more athletic participants into numerous challenges.

Table 2 Results from un-standardized OLS regression predicting average steps by previous challenge participation, demographics and current challenge (beta (p-value))
Full size table

Multilogit regressions predicting previous challenge participation indicated two trends (see Table 3). First, being registered for the 10,000 steps for 100 days challenge significantly predicts the number of previous challenge participation. In model 1, 10,000 Step were 7% more likely to have participated in one previous challenge (p = 0.007) than none and about twice as likely to have participated in two (OR = 2.28 (p = 0.001)) or three (1.78 (p = 0.038)) challenges previously. This relationship is only slightly attenuated in Model 2, after adjusting for self-assessed PA level, whereas 10,000 step participants were 3% more likely to have participated in one previous challenge than none. These participants were still about twice as likely to have participated in two (OR = 2.29 (p = 0.001)) or three (1.80 (p = 0.036)) challenges previously.

Table 3 Multiple logistic regression predicting the number of challenges by demographics, challenge and self-assessed PA
Full size table

Second (see Table 3), self-assessed PA level is only significant in predicting one previous challenge participation. Participants who engage in a low-level activity (OR = 1.59, (p = 0.007) or no more than 1 h per week of aerobic activity (OR = 1.69, (p = 0.008)) were the significantly more likely to have participated in a previous challenge compared to non-active participants at baseline. This finding suggests that participants may benefit from multiple challenge participation, leading to higher overall PA engagement.

5 Discussion

In this study, we set out to determine if multiple walking challenge participation, as a form of repeated exposure, in a community-wide campaign, is associated with greater average walking engagement. Findings from our assessment indicate that repeated challenge participants were observed to have higher baseline step averages and higher 2-week post-challenge average steps. Repeated participants were also more likely to participate in the Walk-the-Walk Team Challenge and the 10,000 Steps for 100 Days Challenge. While there was a benefit of having previous challenge participation, benefits were not incremental. Multiple challenge participation did not significantly affect step achievement more than one previous challenge. Finally, baseline self-assessed physical activity levels only extended to participating in one previous challenge. This finding suggests that greater baseline step averages will likely result from challenge participation rather than selection due to more general physical activity engagement.

We employed a program evaluation using survey and fitness tracker data to understand repeated walking challenge engagement as a repeated exposure to increase the likelihood of sustained walking engagement. Repeated challenge participants differed from first-time challenge participants at baseline in several ways. First, on average, repeated challenge participants had higher step averages at baseline than first-time participants. Second, repeated challenge participants also appeared to assess their physical activity engagement higher than first-time participants. One theory behind this behavior is that walking challenges effectively increase physical activity because they provide an environment that warrants opportunities for social comparison and self-evaluation relative to others [30]. For instance, some [31, 32] studies identified that participants from their walking challenge study were motivated to walk more due to social comparison feedback between groups. Self-tracking and the ability to assess self-performance were also possible reasons why repeated challengers may have performed better than first-time participants. Previous studies [32, 33] also noted that there is a substantial beneficial impact on reported physical health and psychological well-being (such as positive feelings and a sense of success) when fitness trackers are used with mobile applications.

For this study, repeated participation was most likely to occur among participants who participated in the Walk-the-Walk Team Challenge, an incentivized group challenge, and the 10,000 steps for 100 Days challenge, an individual-based competition. The 10,000 Steps for 100 Days Challenge participants were likelier to have participated in multiple challenges than the others. This challenge is individually goal based, intended to promote long-term walking engagement equivalent to the recommended guidelines for physical activity. Given the individual focus, the intrinsic nature of the challenge (i.e., individualized goal achievement) may be a key design feature important to sustained walking after challenge completion. Personal feelings of enjoyment and satisfaction are likely to be a pivotal component to participating in the 10,000 for 100 days challenge. This finding may be essential in sustained walking engagement over the more extrinsically-based team challenge design of the Walk-the-Walk Team Challenge. However, in the team challenge, participant motivation was based on external rewards such as team member approval and financial assessment [26]. Personal motivation has previously been observed as a stronger predictor of physical activity than extrinsic factors [34] and could be a vital ingredient to long-term sustainability [35]. Although repeated participation was observed among Walk the Walk Team Challenge participants, this may not translate into sustained walking engagement due to the short-term extrinsic reward of financial incentive that is short-lived.

Other noteworthy findings were that most first-time participants were enrolled in our Step Boot Camp challenge. This challenge targeted first-time participants as a stepping-stone to the more intensive challenges. While the challenge successfully got participants, it does not appear to perform as intended in encouraging participants to continue to other challenges in our series. This failure most likely resulted from the design of the Bootcamp, which encouraged participants to increase the number of steps they conducted weekly until they produced three times their initial step count. We have received feedback of this nature and have made adjustments to facilitate completion. Other boot camp-type challenges have observed similar struggles with encouraging participation and maintaining motivation among physical activity intervention participants [10, 13].

In sum, multiple walking challenge participation in this study was associated with higher average walking engagement. Moreover, the type of walking challenge may enhance long-term physical activity engagement. These findings are promising and provide implications for implementation programming. However, several limitations to this study would need to be addressed in future work before this approach could be considered evidence-based. First, the sample was recruited through a community-wide campaign and not randomly selected. Second, participant selection occurred on a rolling basis and not at the same challenge. Additionally, participants were not followed through time with collected data during each challenge in which they participated. Moreover, the differences in walking challenge length (100 days, 63 days, and 30. days) are also a limitation that could explain the differences in step counts between challenges. Another major limitation is that this study was conducted in a predominantly Mexican American community and did not represent a more diverse population. We have already begun to address these factors in planning future evaluations.

6 Conclusions

This exploratory assessment of walking challenge participation directs the next steps for improving walking and other physical activity engagement. Offering community-level programming at regular intervals may be a sustainable approach to increasing the engagement of at-risk populations due to sedentary lifestyles. Citywide walking challenge campaigns can potentially improve population disparities related to physical inactivity in Mexican American communities.