Low Discretionary Time as a Barrier to Physical Activity and Intervention Uptake

Abstract

Objective

To determine whether self-reported discretionary time was associated with physical activity and uptake of a physical activity promotion intervention in a multi-ethnic urban sample.

Methods

We examined the association of self-reported discretionary time with hours/week of leisure-time physical activity at baseline and physical activity intervention uptake.

Results

Low levels of discretionary time were significantly (P<0.01) associated with fewer hours/week (β = −0.78, 95%CI = −1.34, −0.22) of physical activity at baseline. Discretionary time was not associated with physical activity intervention uptake.

Conclusion

Lack of discretionary time may serve as barrier to physical activity, but its importance on intervention uptake is less clear.

Regular physical activity is associated with a reduced risk of obesity, numerous chronic diseases (eg, cardiovascular disease, diabetes, and some cancer types), and premature mortality.¹ Despite the clear and well-established benefits of physical activity, physical inactivity is pervasive in the United States.^2,3

Previous work has identified a host of social contextual barriers to physical activity, emerging from a range of individual, social environmental, and physical environmental sources.^4–16 These barriers likely result from a confluence of adverse occupational and domestic social contextual exposures and may restrict opportunities for participation in discretionary pursuits.¹⁷

The literature has suggested that individuals who report having lower levels of discretionary time may be less likely to be physically active; however, research in this area is limited and findings have been mixed.¹³ It can be argued that those who perceive themselves to have limited free time may be less inclined to spend it in physically active pursuits. In light of emerging evidence suggesting that individuals have more discretionary time than they may initially perceive,¹⁸ integrating perceived discretionary time as a target of physical-activity promotion interventions would seem prudent. However, we are aware of no evidence discussing the association between discretionary time and change in physical activity following intervention. Furthermore, the particular impact of this on populations with significant numbers of women and racial/ethnic minorities, especially those in an urban environment, who may face particularly high levels of competing demands and thus, theoretically less discretionary time, is unknown. Lack of discretionary time is frequently given as a barrier to physical activity in qualitative studies,^6,8,12,15 but the effects on actual physical activity levels is largely unknown. Furthermore, although intervention studies may try to impact perceived and actual discretionary time, their ultimate success at increasing physical activity is unknown.

These omissions represent a major gap in the literature. Although there is mounting evidence that a range of social contextual factors might be associated with physical inactivity, we know relatively little about how any of these barriers might affect intervention uptake. Discretionary time may be a particularly salient barrier given the frequency of its mention in the qualitative literature.^6,8,12,15 Therefore, the present study sought to determine whether self-reported amount of discretionary time is positively associated with self-reported baseline physical activity and intervention uptake in a behavioral physical-activity promotion intervention conducted in a multi-ethnic working-class urban sample.

METHODS

Sample

This study includes data collected as part of the Harvard Cancer Prevention Program Project’s Small Business Study, which was designed to target multiple cancer risk behaviors among employees (n = 1740) in small manufacturing businesses in eastern Massachusetts.^19,20 All study procedures were fully approved by the relevant human subjects committees.

Businesses enrolled in the study were identified using Standard Industrial Classification codes (20–39) and were subsequently limited to those located in and around the metropolitan Boston area that employed between 50 and 150 workers. Potentially eligible worksites were those that employed at least 25% of workers who were first- or second-generation immigrants or people of color, had a turnover rate of less than 20% per year, and were in a position to make the decision to participate in the intervention study. One hundred thirty-three worksites met these criteria, and 26 consented to participate in the study. There were no significant differences between worksites that consented to participate and those that refused. ²¹ Individual workers were eligible if they were permanent employees, worked over 20 hours a week, worked on site at the facility, and spoke English, Spanish, Portuguese or Vietnamese. The worksite was the unit of randomization. Of 2069 eligible employees, 1740 completed the survey (84% response rate).²⁰ Workers participated in an interviewer-administered survey delivered in English, Portuguese, Spanish or Vietnamese. There were 974 participants who completed both the baseline and follow-up surveys.

Intervention

Details regarding the design of the intervention are presented more fully elsewhere.²² Briefly, the 18-month intervention was designed to incorporate aspects of the participants’ social context that might influence health behavior change.¹⁷ Intervention components were delivered worksite-wide and focused on multiple behavioral outcomes, including fruit and vegetable consumption, red meat consumption, multivitamin intake and physical activity. The intervention was designed to target multiple levels of influence at the workplace. We worked closely with each worksite to form employee advisory boards, which assisted with intervention development and delivery and helped to assure that the intervention responded to the needs and culture of the individual worksite. The intervention included individual, small group, and worksite-wide activities to promote health behavior change and to create a supportive environment for health behavior change. These programs were designed to appreciate the multi-ethnic composition, literacy levels, and language requirements of each worksite. Intervention and survey materials were created in English, translated by professional translators, and back translated to ensure accuracy.

Protocols for each of the 18 different monthly activities included a description of the activity, the objectives as well as a mock-up of any educational materials. The protocols and mock-ups were brought by the intervention staff to the employee advisory boards (EABs) at each of the intervention worksites to tailor to their worksite. As an example for physical activity, the worksite helped us come up with a listing of physical activities the workers there would likely participate in (eg, dancing, soccer, taking stairs instead of elevators, forming walking clubs in worksites where there was great social cohesion). We also tailored the management-level intervention; for physical activity, we worked with the EABs to come up with company policies that supported physical activity (eg, for workers that wanted to join a worksite-based walking club, the worksite would “donate” a half hour if the workers used a half hour of their lunch). Thus, the intervention encouraged both leisure-time (eg, go on a walk with your family) and lifestyle (eg, walk to the next bus stop) physical activity. The intervention resulted in a significant increase in the proportion of employees who met physical activity recommendations in the intervention sites versus a decrease at the control sites (P=0.02).²²

Measures

Discretionary time

On the baseline survey, participants were asked to report “How much time a week do you have for just for yourself, without any other responsibilities?” Responses were coded in hours/week. Participants who struggled to answer were prompted to make their best guess.

Physical activity

Using a modified version of a previously validated self-report physical activity questionnaire, participants reported how many hours per week in the last 4 weeks they engaged in 8 types of leisure-time activities: walking for exercise (including walking to work or to and from the bus or subway); jogging; running; bicycling; aerobics or aerobic dancing; playing soccer, rugby, basketball or lacrosse; playing baseball, football, bowling, or lifting weights; and other activities not mentioned that “get you at least a little out of breath.”^23,24 Prompts (ie, Have you spent any time at this activity in the last 4 weeks? If yes, have you spent more than 5 minutes per week at this?) were employed to help participants identify the time, if any, spent in each activity. The instrument’s example items were modified based on pilot testing in this population. Our analyses focused on hours/week of leisure-time physical activity. In a validation study in a subsample of 84 participants, the Spearman correlation between total hours of physical activity (as recorded on accelerometers) and self-report of leisure-time physical activity was 0.53.

Socio-demographic Characteristics

Educational attainment was reported as the highest level of schooling attained in 8 categories ranging from elementary to postgraduate degree. This variable was collapsed into high school education or less, some college or associate’s degree, or bachelor’s degree or more. Racial/ethnic data were collected by self-report as white, black, Hispanic, Asian, or Native American. Respondents were permitted to report more than one group. Due to small numbers in some categories, these responses were categorized as: non-Hispanic white, non-Hispanic black, Hispanic, Asian, and other (which included all other non-Hispanic participants including those reporting more than one racial or ethnic group). Participant’s gender was self-reported, and age was calculated based on the reported date of birth.

Data Analysis

Baseline analyses include all participants; intervention uptake analyses include only those participants randomized to the intervention arm following the baseline survey. Of those subjects enrolled at baseline, 122 were missing baseline physical-activity data, an additional 13 were missing data on discretionary time, an additional 17 were missing data on age, and 2 were missing data on gender, leaving 1586 participants in the baseline analysis. Of those in the baseline analysis, 706 received the intervention and 387 of those reported follow-up physical activity values. Thus, the analyses of intervention uptake included 387 individuals. Participants at control sites are not included in the intervention uptake analyses.

The associations between discretionary time and physical activity outcomes were assessed using linear random effects models with normal errors to account for the nonindependent nature of observations from individuals clustered within a worksite. Discretionary time was categorized into tertiles. In baseline analyses, the regression coefficient represents the difference between tertiles in number of hours/week of physical activity. In follow-up analyses, the regression coefficient represents intervention uptake, operationalized as the difference between tertiles in hours/week of physical activity from baseline to follow-up. Data were analyzed with SAS, Version 8.02 (SAS Institute Inc, Cary, NC) using the mixed model analysis procedure “PROC MIXED.” Race/ethnicity, education, age and gender were controlled for as covariates. Intervention uptake analyses are also adjusted for baseline physical activity. We used interaction terms to consider the possibility that the association between physical activity and discretionary time could vary by gender or race/ethnicity. Interactions were considered significant at the P=0.10 level.

RESULTS

Study participants ranged in age from 18 to 82, with a mean age of 43.6 years (SD =11.6) (Table 1). Sixty-seven percent of participants were male and 32% were racial/ethnic minorities. Nearly half participants had a high school education or less. Participants reported a mean (SD) of 13.8 (12.5) hours/week of discretionary time, with a range of 0 to 90 hours/week. Self-reported physical activity at baseline ranged from 0 to 24 hours/wk with a mean of 5.6 hours/week. Physical activity uptake (change in physical activity from baseline to follow-up) ranged from a decrease of 16.7 hours/week to an increase of 15.3 hours/week.

Table 1.

Participant Demographics at Baseline, Harvard Cancer Prevention Program Project Small Business (n=1586)

	Na	%
Race/Ethnicity
White	1082	68.2
Black	81	5.1
Hispanic	201	12.7
Asian	104	6.6
Other	109	6.9
Education
High school or less	729	46.0
Some college/associate degree	524	33.0
Bachelor’s degree or more	320	20.2
Gender
Male	1067	67.3
Female	519	32.7
Discretionary Time
Low (<7 hrs/wk)	502	31.7
Moderate (7–15.9 hrs/wk)	571	36.0
High (≥16 hrs/wk)	513	32.4
	Mean Std Dev
Age	43.6	11.6
Leisure-Time physical activity at baseline (hours/week)	5.6	4.7
Leisure-Time physical activity at follow-up (hours/week)	5.5	4.5
Physical activity uptake (hours/week)	−0.9	4.5

^aNote. Numbers may not sum to total because of missing data

Baseline Physical Activity

In a multivariable model adjusting for age, sex, race/ethnicity, and education, discretionary time was positively associated with baseline physical activity (F=8.95, P<0.01) (Table 2). Participants in the low discretionary time group reported over 45 fewer minutes/week of physical activity than did those in the high discretionary time group (β=−0.78 hrs/wk, 95%CI −1.34, −0.22). There was no significant difference in physical activity between those in the moderate and those in the high discretionary time groups (β=0.37 hrs/wk, 95% CI −0.17, 0.91). The association between discretionary time and physical activity did not differ by gender (P for interaction = 0.18) or race/ethnicity (P for interaction = 0.21).

Table 2.

Leisure-Time Physical Activity and Discretionary Time in the Harvard Cancer Prevention Program Project – Small Business

	Baseline physical activity: Age & sex adjusted hrs/ wk (95% CI)	Baseline physical activity: Multivariable hrs/wka (95% CI)	Change in physical activity: Age & sex adjusted hrs/ wk (95% CI)	Change in physical activity: Multivariable hrs/wka (95% CI)
Discretionary Time
Low	−0.75 (−1.30, −0.19)	−0.78 (−1.34, −0.22)	−0.10 (−1.07, 0.87)	−0.14 (−1.14, 0.85)
Moderate	0.49 (−0.04, 1.03)	0.37 (−0.17, 0.91)	0.27 (−0.67, 1.20)	0.06 (−0.90, 1.01)
High	reference	reference	reference	reference

^aNote. Multivariable model adjusted for age, sex, race/ethnicity and education

Intervention Uptake

Discretionary time was not associated with physical activity uptake (F=0.09, P = 0.92) in multivariable analysis. There was no difference in physical activity between those in either the low (β=−0.14 hrs/wk, 95%CI −1.14, 0.85) or the moderate (β=0.06 hrs/wk, 95%CI −0.90, 1.01) groups, compared to those reporting high discretionary time. The association did not differ by gender (P for interaction = 0.14) or race/ethnicity (P for interaction = 0.34).

DISCUSSION

Interestingly and inconsistent with previous research,^{6,8,10,12,15,25} low levels of self-reported discretionary time were significantly associated with baseline leisure-time physical activity. Participants reporting low levels of discretionary time reported significantly less leisure-time physical activity than did those reporting high levels of discretionary time. However, discretionary time showed no association with intervention uptake. These findings extend the intervention’s primary outcomes;²² this one of the few studies that has investigated whether such an oft-cited barrier is associated with physical-activity intervention uptake.

More generally, these findings expand upon qualitative research^6,8,12,15 that has suggested lack of free time may be a barrier to physical activity among socio-demographically diverse populations. One previous study found that a lack of time for activity was only predictive of meeting physical activity recommendations among individuals in higher socioeconomic strata.¹³ In addition, King et al found that lack of time was not associated with likelihood of being active, though they examined lack of time for physical activity specifically, in contrast to our measure of time for oneself in general.¹⁰ We found that less discretionary time was significantly associated with over 45 fewer minutes/week of physical activity at baseline. It is encouraging that discretionary time was not found to impact intervention uptake; one interpretation of this finding is that the intervention was successful in minimizing the impact of this perceived barrier on participant’s physical activity.

Although there is evidence that other social contextual barriers (eg, perception of and access to places to be physically active, perceptions of crime, neighborhood safety, self-consciousness about appearance) to physical activity may vary by race/ethnicity,^8–10 socioeconomic position⁵ and gender,⁵ we failed to find such racial/ethnic or gender variation for discretionary time in our sample. This may suggest that features of the social environment of lower income populations are more proximate modifiers than race/ethnicity and gender. It is also possible that discretionary time does not show such variation or that we lacked the statistical power to detect such variation.

There are several factors that may affect the interpretation of our findings. Discretionary time was assessed via self-report and, as such, reflects participants’ perceptions of their circumstances. We were unable to objectively evaluate time that participants might otherwise have used to be physically active. However, subjective perceptions may shape the decision to participate in physical activity, making objective assessments of discretionary time solely insufficient. For example, a recent analysis found that there was significant discordance between objectively determined free time and perceptions of time available for physical activity among African American and Hispanic women in Texas.¹⁸ In addition, it is possible that those participants who completed both baseline and follow-up examinations were different from the larger population of small business employees, limiting the generalizability of the findings. Finally, we relied on a self-report measure of physical activity. Despite the fact that the instrument was validated in this and other samples, the physical activity levels reported by our population are higher than those typically estimated for the general population. Emerging research suggests that some population segments (ie minorities and those with low socioeconomic position) may experience higher levels of non-leisure-time physical activity, such as transportation-related physical activity (ie, walking to work). It is possible that our study subjects included nonleisure domains of physical activity, particularly when estimating walking, in their reports.

Given that perceptions of discretionary time are significantly associated with baseline leisure-time physical activity in the urban working class, future prevention efforts should address perceptions of discretionary time and should include strategies to create time and opportunities for physically active pursuits. Particular attention should be paid to cultural and gender norms around notions of free time. Finally future research should consider how to best address participants’ perceptions of their discretionary time in an effort to improve intervention efficacy in increasing physical activity among socio-demographically diverse populations.