Examining whether physical activity location choices were associated with weekly physical activity maintenance across 13 months of the COVID-19 pandemic in emerging adults

Abstract

Background:

Recent studies have shown potentially detrimental effects of the COVID-19 pandemic on physical activity (PA) in emerging adults (ages 18-29 years). However, studies that examined the effects of COVID-19 on PA location choices and maintenance for this age group remain limited. The current study investigated changes in PA location choices across 13-months during the pandemic and their associations with PA maintenance in this population.

Methods:

Emerging adults (N=197) living in the U.S. completed weekly survey on personal smartphones (May 2020 - June 2021) regarding PA location choices and maintenance. Mixed-effects models examined the main effects of PA location choice and its interaction with weeks into the pandemic on participants’ PA maintenance.

Results:

On a given week, participants performing PA on roads/sidewalks or at parks/openspaces were one-and-half and two times as likely to maintain PA levels, respectively. Moreover, after September 2021, weeks when individuals performed PA on roads/sidewalks had a protective effect on PA maintenance.

Conclusions:

Performing PA on roads/sidewalks and at parks/openspaces were associated with PA maintenance during the COVID-19 pandemic. PA promotion and intervention efforts for emerging adults during large-scale disruptions to daily life should focus on providing programmed activities in open spaces to maintain their PA levels.

1. Introduction

Recent studies have shown potentially detrimental effects of the coronavirus disease 2019 (COVID-19) pandemic on adults’ physical activity (PA) levels.^1-7 Emerging adults (ages 18-29 years) often undergo major life-course transitions (e.g., in social support, autonomy and resources) and thus may be more susceptible to declines in PA behavior resulting from disruptions such as a global pandemic.^2,8-10 However, studies also report that changes in PA behavior in adults may vary considerably by their ability to adapt to PA contextual changes due to community-level pandemic prevention strategies, such as stay-at-home orders and limited/restricted access to exercise facilities, parks, and open spaces.^2,3,11-13 Thus, understanding where emerging adults performed their PA is critical in formulating effective PA promotion and intervention strategies targeting this population.

Moreover, previous research has shown that certain domains and locations (e.g., home, parks, sidewalks) were associated with adults’ PA levels.^1,3,14 For instance, a recent study by Dunton et al. found that performing PA at home was associated with more moderate or vigorous types of PA in adults during the COVID-19 pandemic.³ However, the study did not specifically focus on emerging adults and only examined this relationship during the first three months after the U.S. President declared a national emergency concerning the COVID-19 pandemic. Thus, more studies are needed to understand further the long-term effects of physical environments (e.g., location choices) on PA behaviors in emerging adults in this extended pandemic.

To fill the above-mentioned gap, using ecological momentary assessment (EMA) survey data collected from 18–29-year-olds in the U.S. between May 2020 and June 2021, we aim to:

1) describe weekly changes in PA location choices in participants across 13 months of the COVID-19 pandemic.

2) examine time-varying and longitudinal effects of various locations where PA had been performed on odds of reporting same/more weekly PA levels compared to the pre-COVID-19 periods across the 13-month period.

2. Methods

2.1. Data, sample, and survey questions

Data for this analysis comes from the ongoing Temporal Influences on Movement and Exercise (TIME) study. TIME study uses a prospective within-subject case-crossover design and examines real-time micro-temporal influences underlying adoption and maintenance of physical activity, sedentary behavior, and sleep. A total of 206 ethnically diverse, emerging adults were recruited nationally across the U.S. between March 2020 and August 2021. Participants were recruited by emailing individuals enrolled in the Happiness & Health Study, a USC longitudinal cohort study of young adults, obtaining referrals from existing participants (word of mouth), or contacting participants identified using ResearchMatch, a national health volunteer registry that was created by several academic institutions and supported by the U.S. National Insitutes of Health as part of the Clinical Translational Science Award (CTSA).¹⁵ More details on study recruitment are available in Appendix Table 1. The University of Southern California Institutional Review Board approved all study procedures, and participants signed an informed consent form before enrolling in the study.

Between May 3^rd, 2020, and June 21^st, 2021, participants completed an ecological momentary assessment (EMA) survey on Sundays. Survey questions included reactions to the COVID-19 pandemic including physical contexts of PA and PA levels compared to the pre-COVID-19 period. EMA data was self-reported using a custom app built for Android smartphones. EMA surveys were prompted two hours before a participant’s anticipated sleep time for that day, assessed by self-report earlier in the day. The complete list of items in the EMA survey is available in Appendix Table 2.

The survey question used to measure the outcomes was: “Over the past week, the COVID-19 outbreak has affected how much PHYSICAL ACTIVITY/EXERCISE I got compared to before the COVID-19 outbreak. I exercised…” Answers (“a lot more,” “more,” “neither more nor less,” “less,” and “a lot less”) to the above question were classified as more/same (including “a lot more,” “more,” “neither more or less”) or less (including “less” and “a lot less”). PA location choices were measured using the question: “Over the past week, where did you do physical activity/exercise (choose all that apply)?” Answers to this question were regrouped into four frequently-used location categories (at/around home, on roads/sidewalks, at parks/open spaces, at exercise facilities) as designated by previous studies examining location-based PA.¹⁴ Then, four binary predictor variables were generated (PA at/around home: yes/no; PA on roads/sidewalks: yes/no; PA at parks/open spaces: yes/no, and PA at exercise facilities: yes/no).

Covariates included weeks into COVID-19 since the U.S. President declared national emergency on March 13^th, 2020, states where participants resided (coded as 0: non-California residents vs. 1: California residents), and using remote or streaming services to facilitate PA (classified as no: 0 day vs. yes: ≥1 day). Sociodemographic characteristics included age, sex (female, male), ethnicity (Hispanic vs. Non-Hispanic), and personal financial status (classified as “live comfortably,” “meet needs with a little left,”, and “just/don’t meet basic expense”).

2.2. Statistical analysis

Logistic mixed-effects models were applied to account for between- and within-person variations in reporting same/more weekly PA, given the hierarchical structure of the data (i.e., EMA weekly survey prompts nested within people). The models estimated the between-subject (BS) and within-subject (WS) effects of performing PA at a particular location and their interaction effects with weeks into COVID-19 (both linear and curvilinear terms) in predicting odds of reporting same/more weekly PA levels during a given week. All models were adjusted for the covariates listed above. The coefficients generated from the model were exponentiated for more straightforward interpretation (i.e., odds of reporting same/more weekly PA compared to pre-COVID-19 periods). All analyses were performed using R (version 4.0.2; R Core Team, 2020) and the lme4 package (version 1.1-27.1).¹⁶

3. Results

3.1. Data availability and participant characteristics

Of the 206 participants enrolled in the study by June 2021, a total of N=197 participants completed the weekly COVID EMA survey between May 3^rd, 2020, and June 21^st, 2021. Of these individuals, we removed 29 cases with missing data on one or more outcome or covariates variables, resulting in a final analytical sample size of 168 with a total of 4,303 weekly COVID EMA survey occasions. We show participant characteristics of the analytical sample in Appendix Table 3. Overall, the mean age was 23.0 years (SD=2.9). A little less than half of the participants were female (47.0%), and about three out of ten participants were Hispanic (30.4%). Close to three-quarters of participants reported financial status of “live comfortably” and “meet needs with a little left.”

3.2. Changes in maintenance of PA levels and PA location choices

We illustrate changes in PA location choices from May 2020 (Week 9 of the COVID-19 pandemic)) through June 2021 (Week 68 of the COVID-19 pandemic) in Figure 1. The top two PA location choices during the earlier pandemic period were on roads/sidewalks (87.5% of participants) and at/around home (85.7% of participants). Between these two locations, proportions of participants performing PA at home declined across the study period while proportions performing PA on roads/sidewalks declined initially (reached lowest point 30.8% at week 47) then flatted out with a gradual increase. Both proportions were at about 50% at week 68, the end of the analysis period. Parks/open spaces were the third most popular location (57.1% of participants at the beginning) and followed a similar curvilinear trajectory as roads/sidewalks across the study period with initial declines followed by a flattening and gradual increase. The proportions of participants who reported performing PA at exercise facilities was stable and low during the entire observation period, with a slight downward trend due to seasonal effects (after about 40 weeks, which corresponds to November-December 2020) and recovery after in the spring of 2021. Notably, the percentage doesn’t seem to increase even after the gradual reopening of indoor exercise facilities across the latter half of 2021 into early 2021. Lastly, the proportions of participants who reported not performing any PA slowly increased, reaching 20-25% in June 2021 (Week 68 since the U.S. President declared national emergency).

Figure 1.

Changes in proportions of participants reported performing PA at home, on road/sidewalks, on parks/open spaces, and on exercise facilities, as well as not performing PA as the weeks into the COVID-19 pandemic increased.

3.3. Associations between PA location choices and maintenance of PA levels

We show model results that examine the main effects of various locations that PA have been performed and their interaction effects with weeks into the pandemic in predicting odds of reporting the same/more weekly PA compared to pre-COVID-19 periods in Table 1. Results indicated that maintenance of PA levels changes in a curvilinear fashion over time (see Appendix Figure 1), such that the probabilities of reporting the same/more PA as compared to before the COVID-19 pandemic decreased until about Week 32 (November 2020) and then increased through Week 68 (June 2021). After adjusting for age, ethnicity, personal financial status, whether participant resided in California, using PA remote/streaming services, within a given week, participants who performed PA on roads/sidewalks or at parks/open spaces were one-and-half (WS OR=1.54; 95%CI: 1.12-2.12), and two (WS OR=2.10; 95%CI: 1.52-2.90) times as likely to report same/more weekly PA compared to pre-pandemic periods, respectively. Performing PA at/around home or at exercise facilities; however, was not associated with maintenance of PA levels.

Table 1.

Results of mixed-effect models in testing the hypothesis of main and 2-way interaction effects in predicting odds of reporting of more/same weekly physical activity compared to the pre-COVID-19 periods.

	Level 1 = 168 participants Level 2 = 4,303 occasions
	Main effects only	2-way interaction with weeks into COVID
Predictors	Odds Ratios (95% CI)	Odds Ratios (95% CI)
Performing PA at or around home
BS	0.59 (0.17 – 2.07)	0.62 (0.18 – 2.14)
WS	1.20 (0.90 – 1.58)	1.18 (0.89 – 1.56)
Performing PA on roads/sidewalks
BS	0.50 (0.12 – 2.01)	0.54 (0.13 – 2.18)
WS	1.54 ** (1.12 – 2.12)	0.69 (0.38 – 1.26)
Performing PA at parks/open spaces
BS	0.84 (0.15 – 4.71)	0.84 (0.15 – 4.69)
WS	2.10 *** (1.52 – 2.91)	2.26 * (1.17 – 4.38)
Performing PA at exercise facilities
BS	1.66 (0.23 – 12.20)	1.66 (0.23 – 12.20)
WS	1.24 (0.75 – 2.03)	1.25 (0.76 – 2.06)
Weeks into the COVID-19 pandemic
Linear term	0.42 * (0.18 – 0.99)	0.42 (0.18 – 1.00)
Quadratic term	1.15 ** (1.05 – 1.26)	1.13 * (1.03 – 1.24)
Resided in California (yes)	1.53 (0.33 – 7.08)	1.03 (0.76 – 1.38)
Using remote/streaming to facilitate PA	1.65 ** (1.13 – 2.41)	1.63 * (1.11 – 2.39)
Age in years	1.01 (0.85 – 1.19)	1.01 (0.85 – 1.19)
Sex (female)	1.24 (0.51 – 3.04)	1.22 (0.50 – 2.97)
Hispanic/Latino	1.78 (0.71 – 4.47)	1.73 (0.69 – 4.34)
Personal Financial Status (Meet needs with a little left)	0.87 (0.25 – 3.02)	0.86 (0.25 – 3.01)
Personal Financial Status (Live comfortably)	0.73 (0.29 – 1.81)	0.73 (0.29 – 1.81)
Performing PA at roads/sidewalks x Weeks into COVID-19
WS x quadratic term		1.04 ** (1.01 – 1.07)
Performing PA at parks/open spaces x Weeks into COVID-19
WS x quadratic term		1.00 (0.97 – 1.02)
Marginal R2 / Conditional R2	0.045 / 0.731	0.049 / 0.731
AIC	3166.075	3159.616

Notes. PA = physical activity. BS = between-subject. WS = Within-subject.

^*p<0.05

^**p<0.01

^***p<0.001

The interaction of performing PA at roads/sidewalks and the quadratic terms for weeks into the pandemic was significant (OR=1.04; 95%CI: 1.01-1.07). The predicted trajectories for the probabilities of reporting same/more weekly PA were estimated with stratifications by whether PA had performed on roads/sidewalks. As shown in Appendix Figure 2, up to about Week 30 (October 2020), the probabilities of reporting the same/more PA compared to the pre-pandemic period was lower on weeks when individuals performed PA on roads/sidewalks. However, after Week 30, the probabilities of reporting the same/more PA than the pre-pandemic period were higher on weeks when individuals performed PA on roads/sidewalks. The interaction effect between performing PA at parks/open spaces on any given week and quadratic weeks into the COVID-19; however, was not significant.

4. Discussion

This study examined changes in PA location choices across a 13-month period (May 2020 and June 2021) during the COVID-19 pandemic. It also examined time-varying and longitudinal effects of these locations on PA level maintenance during COVID-19. The higher proportions of participants performing PA at/around home or on roads/sidewalks shown in late May 2020 were consistent with other studies showed PA was confined to home and neighborhood locations during the early COVID-19 pandemic among U.S. and Canadian adults.^1-4 However, the proportion of participants performing PA at parks/open spaces during the early COVID-19 period was higher in our study than in previous research. Watson et al. found a negative association between age and performing PA at a park or public trail during the pandemic, which may help to explain this discrepancy since our study targeted emerging adults.² Moreover, our study also extended these early studies by providing evidence of weekly PA location choice changes across a 13-month period during the pandemic, including both declines in performing PA at/around home, on roads/sidewalks, and at park/open spaces. This finding indicates the potential changes in PA location choices among U.S. emerging adults during this extended pandemic.

The null finding of associations between home locations and emerging adults’ PA in our study differed from previous adults that reported positive associations between performing PA at home and levels of moderate-to-vigorous PA in U.S. and Canadian adults during the pandemic.^3,17 However, both studies only examined early COVID-19 periods (March to May 2020) and focused on all adult age groups. A recent study by Folk et al. found that almost four-fifths of their U.S. emerging adult samples increased time spent in outdoor PA (e.g., running, hiking).¹⁰ Therefore, it is likely that our samples of emerging adults may have stronger preferences to perform PA in outdoor spaces (e.g., parks, trails, beaches), particularly after the lift of public health measures and restrictions, which attenuates the effects of home location on their PA.

Additionally, our discovery of positive associations between performing PA on roads/sidewalks and at parks/open spaces and maintaining PA levels suggests these types of locations may be essential for long-term maintenance of emerging adults’ PA during the pandemic. Our results further corroborate this finding that participants who reported performing PA on roads/sidewalks after October 2020 were more likely to have similar or higher PA levels than before the pandemic.

To our best knowledge, this is the first study that examined changes in PA location choices and their time-varying and longitudinal associations with PA maintenance across 13 months during the COVID-19 pandemic. Our results fill in existing gaps in understanding of the long-term effects of this pandemic on PA behaviors in emerging adults. Yet, our study has several limitations. A major one is using the self-report measures of PA location choices and retrospective measures of PA maintenance, which may be subject to reporting errors and biases. Additionally, study participants resided in 28 states across the U.S. As a result, they were subject to different COVID-19 restrictions and business operating guidelines, which may have influenced the PA location choices and maintenance of participants differently. Moreover, the sample only consisted of emerging adults in the U.S. who were largely middle to high income. Thus, our results may not be generalizable to other adult age groups or emerging adults in other countries/regions. Furthermore, we did not use social media to recruit participants and limited participants to Android-based smartphone users, which may restrict the representation of our samples.

5. Conclusion

In summary, our study results indicate that PA location choices of the U.S. emerging adults may have changed across 13 months during the COVID-19 pandemic. PA promotion and intervention efforts for emerging adults during large-scale disruptions to daily life such as the COVID-19 pandemic should focus on providing amenities and programmed activities in public open spaces such as parks and sidewalks to stimulate their interest in visiting these spaces and maintaining their PA levels.

Appendix

Appendix Table 1.

TIME study recruitment and end-of-day (EOD) EMA data collection details.

Recruitment	Rolling recruitment began in March 2020 and ended in August 2021.To recruit participants, we implemented following strategies: (1) sending emails to individuals enrolled in the Happiness & Health Study, a USC longitudinal cohort study of young adults, (2) referrals from existing participants (word of mouth), and (3) contacting participants identified using ResearchMatch, a national health volunteer registry that was created by several academic institutions and supported by the U.S. National Institutes of Health as part of the Clinical Translational Science Award (CTSA) program.
Inclusion criteria	(1) 18-29 years old living in the United States, (2) intend within the next 12 months to engage in (or already engage in) recommended levels of MVPA (≥150 min/week moderate or ≥75 min/week vigorous intensity), (3) use an Android- based smartphone as their only primary personal mobile device with no intention to switch to a non-Android phone, (4) able to speak and read English and (5) plan to reside in an area with Wi-Fi connectivity during the study period.
Exclusion criteria	(1) physical or cognitive disabilities that prevent participation, (2) health issues that limit physical activity, (3) any diagnosed sleep disorders, (4) unable to wear a smartwatch or answer EMA surveys at home, work, school, or other location where a substantial amount of time is spent (e.g., participant would not be able to answer prompts more than 20% of the time), (5) spends more than 3 hours/day on a typical weekday or weekend day driving, (6) owns an Android phone version 6.0 (or older), or if the app will not function on the phone due to other technical issues, (7) currently owns and wears a smartwatch, (8) uses pay as-you-go data plan or data plan with less than 2 GB of data per month, or (9) currently pregnant.
EOD EMA Data collection	Across the 12 months, an EMA question set is prompted on the smartphone at the end of each day asking participants to summarize experiences occurring that day and their plans for the next day. EOD EMA prompts are delivered two hours before a participant’s anticipated sleep time.

Appendix Table 2.

The list of end-of-day COVID-19 EMA survey items in TIME study (Sunday only 100%).

Over the past week, how much were you lonely? Not at all, A little, Moderately, Quite a bit, Extremely Over the past week, to what extent has the COVID-19 situation affected your physical health? Not at all, A little, Moderately, Quite a bit, Extremely Over the past week, which of the following are true about your physical health? (Viral symptoms typically include fever, body aches, cough, fatigue, chills, digestive issues) I am not experiencing viral symptoms I am currently experiencing mild-moderate viral symptoms I am currently experiencing severe viral symptoms Over the past week, which of the following behaviors have you been performing to prevent the spread of COVID-19? (check all that apply) Avoiding indoor social gatherings with anyone who does not live with me Avoiding outdoor social gatherings with anyone who does not live with me Avoiding any places with 10 or more people Staying at leasl 6 feet away from other people who do not live with me Avoiding physical contact with other people who do not live with me None of the above Over the past week, which of the following did you do because of tahe COVID-19 (coronavirus) outbreak? (check all that apply) Avoided outdoor spaces (e.g., parks, beaches, forests) Avoided public spaces (e.g., museums, playgrounds, theatres) Avoided restaurants, coffee shops, bars Avoided gyms, fitness centers, yoga studios Avoided participating in team sports None of the above Over the past week, what else of the following did you do because of the COVID-19 (coronavirus) outbreak? (check all that apply) Avoided going for walks or other outdoor exercise activity Avoided grocery stores, pharmacies Avoided public transportation Avoided indoor exercise classes or recreational sports Avoided in-person events in the community Avoided in-person religious services None of the above Over the past week, which of the following behaviors have you been performing? (check all that apply)? Wearing masks or face covering when around those who live with me Wearing masks or face covering when around those who do not live with me Washing hands with soap and water whenever I return home Used hand sanitizer Coughed/sneezed into your elbow Avoiding non-ventilated spaces None of the above Over the past week, where did you do physical activity/exercise (choose all that apply)? Inside my home or garage In my yard or driveway On the sidewalks and roads IN my neighborhood On the sidewalks and roads OUTSIDE my neighborhood At a gym or fitness center At a park, trail, or beach At an outdoor sports facility Other places None of the above Over the past week, with whom did you do physical activity/exercise in the same physical space (choose all that apply)? Spouse/partner Child (ren) Other family members (e.g., cousins, siblings, aunts, uncles) Friends Coworkers Other people None of the above Over the past week, did you use any remote/streaming services to facilitate your physical activity/exercise? (check all that apply) Yes – with friends or family Yes –from my gym/fitness facility Yes –I streamed online workout classes Yes – I played a game online that involves physical activity No – I did not use any remote/streaming services Over the past week, on how many days did you use any remote/streaming services to facilitate your physical activity/exercise? 0 days 1 day 2 days 3 days 4 days 5 or more days Over the past week, how many days did you spend walking a dog? 0 days 1 day 2 days 3 days 4 days 5 or more days Over the past week, the COVID-19 outbreak has affected how much PHYSICAL ACTIVITY/EXERCISE I got compared to before the COVID-19 outbreak. I exercised… A lot more, more, neither more nor less, less, a lot less Over the past week, the COVID-19 outbreak affected how much SITTING I did compared to before the COVID-19 outbreak. I sat… A lot more, more, neither more nor less, less, a lot less Over the past week, the COVID-19 outbreak has affected how much SLEEP I get compared to before the COVID-19 outbreak. I slept… A lot more, more, neither more nor less, less, a lot less Over the past week, what has been the overall level of impact to your daily life due to the COVID-19 outbreak? Not at all, Very little, Moderate, Quite a bit, Extremely Over the past week, what has been your overall level of stress related to the COVID-19 outbreak? Not at all, Very little, Moderate, Quite a bit, Extremely

Notes. EMA = ecological momentary assessment. TIME = Temporal Influences of Movement and Exercise.

Appendix Table 3.

Participant characteristics (Level-2 = 168 emerging adults, Level-1 = 4,303 occasions^a).

Level-1 Variables
Achieved same/more weekly PA compared to the pre-COVID-19 periods, n(%)
No	1176 (27.3%)
Yes	3127 (72.7%)
Performed PA at/around home in the past week, n(%)
No	1911 (44.4%)
Yes	2392 (55.6%)
Performed PA at roads/sidewalks in the past week, n(%)
No	2500 (58.1%)
Yes	1803 (41.9%)
Performed PA at parks/open spaces in the past week, n(%)
No	3332 (77.4%)
Yes	971 (22.6%)
Performed PA at exercise facilitiesb in the past week, n(%)
No	3882 (90.2%)
Yes	421 (9.8%)
Used remote/streaming services to facilitate PA, n(%)
No	3470 (80.6%)
Yes	833 (19.4%)
Occasions in winter season, n(%)
No	2647 (61.5%)
Yes	1656 (38.5%)
Level-2 Variables
Age in years, mean(SD)	23.0 (2.9)
Sex, n(%)
Male	89 (53.0%)
Female	79 (47.0%)
Ethnicity, n(%)
Not Hispanic/Latino	117 (69.6%)
Hispanic/Latino	51 (30.4%)
Personal financial status, n(%)
Just/Don’t meet basic expenses	24 (14.3%)
Meet needs with a little left	65 (38.7%)
Live comfortably	79 (47.0%)

Note. PA = physical activity. LA = Los Angeles.

^aWeekly end-of-day EMA survey (Sunday only 100%).

^bExercise facilities include gyms, fitness centers, and outdoor sports facilities.

Appendix Figure 1.

Effects of weeks into the COVID-19 pandemic on predicting the probabilities of reporting same/more weekly PA compared to pre-pandemic periods in a given week.

Notes. PA = physical activity.

Appendix Figure 2.

Effects of weeks into the COVID-19 pandemic on predicting the probabilities of reporting same/more weekly PA compared to pre-pandemic periods in a given week by whether PA was performed on roads/sidewalks (yes vs no) on the same week.

Notes. PA = physical activity.