An Integrative Review

Abstract

The purpose of this review is to synthesize the research evidence for the role of the work environment—workplace physical activity policies and resources and job strain factors—in explaining physical activity in white-collar workers. White-collar workers are at risk for developing a sedentary lifestyle, which contributes to all-cause mortality. Understanding how work environment can influence worker physical activity is important for the development of effective interventions. We reviewed 15 research articles that describe the relationship between work environment factors and physical activity in predominantly white-collar workers. Relatively consistent evidence was found for the effects of supportive workplace policies and resources. Weak evidence was found for the effects of job strain. Both work environment factors have the potential to influence physical activity but require further exploration to fully understand their contribution to physical activity in white-collar workers. Limitations and implications are discussed.

Keywords

review physical activity/exercise occupational health workers environment

Regular physical activity is protective against numerous chronic diseases and health problems, such as cardiovascular disease, hypertension, diabetes, various cancers, obesity, osteoporosis, and mental illness (U.S. Department of Health and Human Services, 2008; World Health Organization, 2002). However, today’s living and working environments contribute to inadequate physical activity and prolonged sitting, both of which contribute to all-cause mortality (Andersen, Schnohr, Schroll, & Hein, 2000; Hamilton, Hamilton, & Zderic, 2007; Patel et al., 2010). White-collar workers are at greater risk for developing sedentary lifestyles, mainly because of their extended sitting at work (Jans, Proper, & Hildebrandt, 2007; Miller & Brown, 2004). There is also evidence that workers do not compensate for prolonged sitting at work by spending less time in sedentary leisure activities (Jans et al., 2007). Given that employees spend most of their waking hours at work, a better understanding of the possible relationships between work environment factors and physical activity of white-collar workers is needed to inform the development of interventions.

There is an increasing interest in changing the work environment by creating supportive workplace physical activity policies and providing resources that reduce barriers and promote physical activity, such as placing motivational signs to encourage stair use and increasing access to places for physical activity (Gates, Brehm, Hutton, Singler, & Poeppelman, 2006; Pronk, 2009; Pronk & Kottke, 2009; Sallis, Bauman, & Pratt, 1998). Moreover, it is thought that health promotion programs should go beyond educational initiatives to support behavior change in the target population (Engbers, van Poppel, Chin, Paw, & van Mechelen, 2005). Policies and resources that support individual level efforts are more likely than educational initiatives to produce successful physical activity outcomes (Pronk & Kottke, 2009). However, this is an understudied area and it is not clear whether workplace physical activity policies/resources are associated with employee physical activity because most of the research examines workplace programs that target multiple health-related behaviors, making it difficult to determine the isolated effects of workplace policies/resources on physical activity.

In addition, the psychosocial work environment, specifically job strain, as a determinant of physical activity, has received attention in recent years. The leading theoretical model that tries to explain this mechanism is Karasek’s (1979) demand/control model (also termed the job strain model; Figure 1). Job strain, a form of psychological stress, refers to jobs characterized by high psychological work demands combined with low decision latitude. Job demands are the psychological demands of work, including workload, work pace, organizational constraints on completing work, and conflicting demands. Job decision latitude is the primary measure of job control and is referred to as the combination of making autonomous decisions on the job and using a variety of skills on the job. According to the model, low control combined with high demands represents a “high strain job,” which may lead to adverse health outcomes, whereas high control combined with low demands represents a “low strain job.” High control combined with high demands represents an “active job,” which may lead to active behavior, while low control combined with low demands represents a “passive job,” which may lead to a passive lifestyle. Such work characteristics are hypothesized to influence individual behavior and health. In addition, work characteristics are modifiable and may thus provide an opportunity to influence individual behavior, such as physical activity (Karasek, 1979; Karasek et al., 1998). Therefore, a comprehensive review of the existing literature on the relationship between job strain factors and employee physical activity is warranted.

Figure 1.

Karasek job strain model.

This work is grounded in the ecological model for health promotion (McLeroy, Bibeau, Steckler, & Glanz, 1988), suggesting that important factors include individual and organizational. For purposes of this article, we have limited our review to organizational factors. There are a wide variety of organizational factors that could influence physical activity of employees. So far, the only ones that have been studied include work environment factors for policies and resources and job strain factors. Work environment factors refer to (a) psychosocial work environment, specifically job strain variables including the workload, pace of work, employee involvement in decision making, and management styles and (b) workplace physical activity policies and resources including incentives, supervisor and management support, corporate culture, changes in policies, and environmental modifications to support employee physical activity (McLeroy et al., 1988). Although it seems likely that a combination of job strain variables and workplace policies/resources may contribute to a better understanding of the relationship between work environment factors and physical activity, it remains underexplored.

Previous reviews suggest that the effectiveness of workplace physical activity interventions is not consistent (Conn, Hafdahl, Cooper, Brown, & Lusk, 2009; Dishman, Oldenburg, O’Neal, & Shephard, 1998; Proper et al., 2003). Most of the previous research used multicomponent interventions at the individual level (e.g., information, education, and counseling sessions). While Dishman et al. (1998) and Proper et al. (2003) reported limited or inconclusive evidence due to a small amount of high-quality research, Conn et al. (2009) conducted a comprehensive meta-analysis of the literature and found that workplace physical activity interventions have significant positive effects on physical activity (0.21) and health and work-related outcomes (0.08-0.57). Two reviews were conducted to determine the effectiveness of workplace policy and/or environmental interventions on physical activity and dietary intake (Engbers et al., 2005; Matson-Koffman, Brownstein, Neiner, & Greaney, 2005). Matson-Koffman et al. (2005) concluded that policy and environmental interventions may promote physical activity, whereas Engbers et al. (2005) found inconclusive evidence regarding environmental changes to promote physical activity because of the small number of studies available and methodological quality concerns.

A review of work environment factors associated with physical activity targeting white-collar workers is lacking. Therefore, the purpose of this review is to synthesize the research evidence for the role of the work environment—workplace physical activity policies and resources and job strain factors—in explaining physical activity in white-collar workers. Because most reported job strain factors vary between men and women, our data analysis and presentation were stratified by gender.

Method

Design

An integrative review was conducted to synthesize the empirical literature to more fully understand how work environment factors influence physical activity among white-collar workers. The integrative review method allows the simultaneous inclusion of diverse methodologies (experimental and nonexperimental study) and the incorporation of many purposes, such as defining concepts, reviewing theories and/or evidence, and analyzing methodological issues of a phenomenon of interest (Whittemore & Knafl, 2005). We followed the process of the integrative review outlined by Whittemore and Knafl (2005), including problem identification, literature search, data evaluation, data analysis, and presentation.

Search Methods

The following sources were searched to identify potential research articles: online databases (CINAHL, MEDLINE, PsycINFO, and PubMed), reference lists of included papers, and hand search of journals. Articles were retrieved using a combination of the search terms physical activity or exercise, and workers or employees, and work environment or workplace environment or job strain. Papers were included if they were published in English from January 1990 to November 2011 and described the relationship between work environment factors, specifically workplace physical activity policies and resources and job strain factors as independent variables and physical activity or exercise behavior as the dependent variable in predominantly white-collar workers. As stated earlier, workplace physical activity policies and resources refer to any workplace activities that support or promote employee physical activity. Job strain factors include job/work demands, job control/decision latitude, high-strain jobs, low-strain jobs, active jobs, and passive jobs. White-collar workers in this review refer to those who work in clerical, administrative, and professional nonmanual occupations. Intervention studies were not included in this review because there are very few studies of work environment interventions that solely target physical activity among white-collar workers. We limited our review to studies published in 1990 or later to capture the literature that focused on current white-collar workers, such as information technology workers.

Search Outcome

The initial computerized search yielded 624 citations (including 100 duplicates). After removing duplicates and screening titles and abstracts, 16 potentially relevant studies were identified for evaluation. Four additional articles were identified through other sources. After full-text readings of these 20 articles, 5 were excluded. The reasons for exclusion were not focusing on the relationship between workplace physical activity policies and resources and physical activity (n = 2), assessing physical activity as the independent variable (n = 1), the outcome measure combining diet and exercise scores (n = 1), and unable to distinguish between blue-collar and white-collar workers (n = 1). The final sample included 15 articles.

Data Evaluation

For each study, the level of evidence was assessed using a 7-level scale (Melnyk & Fineout-Overholt, 2011), ranging from Level 1 (the strongest evidence: systematic review or meta-analysis of all relevant randomized controlled trials [RCTs], or evidence-based clinical practice guidelines based on systematic reviews of RCTs) to Level 7 (the weakest evidence: opinion of authorities and/or reports of expert committees). The level of evidence rating for each of the included studies was generally low. Fourteen studies were graded Level 6 (weaker evidence from a single descriptive study), and one study (Gimeno et al., 2009) was graded Level 4 (moderate evidence from a well-designed cohort study). All 15 studies were included in this review because of the limited research in this area.

Data Analysis

The constant comparison method described by Whittemore and Knafl (2005), consisting of data reduction, data display, data comparison, conclusion drawing, and verification, was used for data analysis. We extracted data from the included studies, including the authors, year of publication, sample characteristics (number of all participants, gender, age, work setting, and location of study), study designs, the work environment measures, the physical activity outcome measures, and main findings.

To demonstrate the level of empirical support for the work environment factors associated with physical activity, we used the following rules (Campbell & Matthews, 2010): (a) relatively consistent support: Numerous studies examining the factor found a significant relationship with physical activity; (b) weak or no support: Numerous studies examining the factor found a nonsignificant relationship with physical activity; and (c) equivocal: The relationship between a factor and physical activity was about evenly split between studies with significant and nonsignificant findings. For each job strain factor, we counted the number of studies that found a significant relationship with physical activity and the number of studies that found no significant relationship. There were a small number of studies in the area of workplace physical activity policies and resources, so we assessed the relationship between overall supportive policies/resources or cumulative policies/resources and physical activity, instead of assessing workplace physical activity policies and resources separately, for each study.

Results

Study Characteristics

Fifteen articles were identified as meeting the criteria. Five studies were categorized under the workplace policies and resources (Table 1), and 10 were categorized under the job strain factors (Table 2). Fourteen studies were cross-sectional and 1 (Gimeno et al., 2009) was a prospective longitudinal study. All but 1 study (Landsbergis et al., 1998) involved male and female participants, and sample size ranged from 202 (Landsbergis et al., 1998) to 46,573 (Kouvonen et al., 2005) participants.

Table 1.

Characteristics and Main Findings of Studies That Examine the Relationships Between Workplace Physical Activity Policies/Resources and Physical Activity Among White-collar Workers.

Study and level of evidence^a	Sample, setting, and location	Physical activity outcome	Workplace policies and resources	Association with physical activity	Main findings
Crespo et al. (2011) Level 6	N = 1,313 (mostly high-income, sedentary workers; 56% men) M = 45 yr Seattle-King County, Washington and Baltimore, Maryland in the US	Total sedentary and MVPA: objectively assessed via 7-day accelerometry Total job-related PA and recreational PA: assessed by the IPAQ long form Recreational PA at the worksite (1 item): total number of days in the past month employees engaged in any MVPA at their worksite . . .	Worksite Physical Activity Promotion Checklist: The presence or absence of 9 worksite policy and resource strategies to promote PA (e.g., showers, lockers, PA programs)	β = 3.97 (sedentary behavior) β = 1.04 (MVPA) β = 1.11 (recreational PA away from work) OR = 1.39 (recreational PA at work) β = 0.90** (job-related PA)	A higher worksite PA promotion score was related positively to higher total sedentary behavior, MVPA, recreational PA both at work and away from work and negatively to job-related PA.
Dodson et al. (2008) Level 6	N = 977 (mostly sitting/standing while at work; 25% men) M = 43 yr Midwestern communities (Missouri, Tennessee, and Arkansas) in the US	4 PA outcomes: meeting PA guidelines via • moderate PA • vigorous PA • MVPA • walking 6 PA items (focusing on MVPA) derived from the 2003 BRFSS (PA section; in a usual week)	Structured activity policies/resources: • Facilities for PA • Equipment for PA • Subsidized memberships • Group services • Sponsored sports teams Nonstructured activity policies/resources: • Accessible stairways • Personal services • Time for PA • Resource materials • Reduced health insurance • Other monetary incentives	Met PA guidelines via walking OR_adj = 1.7 (1.1-2.8) OR_adj = 2.0 (1.2-3.5) ns ns ns Met moderate/vigorous PA guidelines OR_adj = 1.4 (1.1-1.9) OR_adj = 1.9 (1.2-3.1) ns ns ns ns	Employees with a workplace environment that had more supportive policies and resources for PA were more likely to meet PA recommendations.
Kaczynski et al. (2010) Level 6	N = 375 (mostly white, well-educated workers; 39% men) M = 40 yr Manhattan, Kansas in the US	Active commuting (4 items): “Thinking about the past month: On average, how many days per week do you • walk to work? • walk from work? • bicycle to work? • bicycle from work?”	Cultural supports: • Employer encourages active commuting • Perceived coworkers who actively commute Physical supports: • Bike parking available • Bike storage policies • Showers/lockers	For full sample ns β = 3.98 (1.91-8.28) β = 2.70 (1.40-5.21) β = 5.92 (3.03-11.58) ns	In the cumulative effects, workers reporting more workplace physical and cultural supports for active commuting were more likely to walk or bicycle to work ≥1 time/wk.
Lucove et al. (2007) Level 6	N = 987 (mostly white, well-educated, relatively young workers; 54% men) Age ≥ 18 yr 6 North Carolina counties in the United States	3 LTPA outcomes: • Any LTPA and recommended PA: 10 PA items derived from the 1998 BRFSS (exercise section; during the past month) • Work-break PA (1 item): ‘In a usual week, do you use your lunch or other regular work breaks to do PA or exercise . . . for at least 10 min at a time?’	Individual worksite policies/resources: • Subsidy for health club • Paid time for non-work-related PA • Onsite fitness facility • Safe place to walk outside work • Flexible work schedule Cumulative worksite policies/resources: • 1 • 2 • 3-5	/ LTPA/WPA ns/* LTPA/WPA /ns LTPA/WPA ns/ LTPA/WPA ns/ns LTPA/WPA ns/* LTPA/WPA ns/* LTPA/WPA ns/* LTPA/WPA	Associations between perceived worksite policies/resources and LTPA were strongest for having paid time for non-work-related PA, subsidies for health club use, and onsite fitness facilities at work.
Prodaniuk et al. (2004) Level 6	N = 897 (26% men) M = 42 yr 3 organizations in Canada: a health authority, an educational institution, and an urban center	• Workplace PA (1 item): ‘How much do you incorporate PA into your workday . . . ?’ • LTPA (the modified Godin Leisure-Time Exercise Questionnaire): How many times per week subjects engaged in strenuous, moderate, and mild PA for ≥10 min/time	Perceived Workplace Environment Scale (PWES; 6-item): Each item measures perceptions concerning one of the six environmental dimensions of the ecological workplace PA model, including individual, social, organizational, community, policy, and physical environment	r = .13 (LTPA) r = .23 (WPA)	In regression analyses, a weak association was observed between workplace PA and PWES (R²_adj = .04), and self-efficacy partially mediated the relationship between the two.

Note: yr = year; M = mean; MVPA = moderate-to-vigorous physical activity; PA = physical activity; IPAQ = International Physical Activity Questionnaire; β = parameter estimate; BRFSS = Behavioral Risk Factor Surveillance System; OR_adj = adjusted odds ratio; ns = not statistically significant; LTPA = leisure-time physical activity; WPA = work-break/workplace physical activity; R²_adj = adjusted R².

Rated according to Melnyk and Fineout-Overholt (2011).

p < .05. **p < .01.

Table 2.

Characteristics and Main Findings of Studies That Examine the Relationships Between Job Strain Factors and Physical Activity Among White-Collar Workers.

Study and level of evidence^a	Sample, setting, and location	Physical activity outcome	Job strain factors	Association with physical activity	Main findings
Brisson et al. (2000) Level 6	N = 6,995 white-collar workers (50% men) M = 41 yr (men); 39 yr (women) 21 organizations in Canada: 20 public and semipublic organizations and 1 private insurance company	Sedentary behavior: exercising <1 time/wk over the last 6 mo One exercise session = a rigorous PA during leisure-time ≥20 min	Job demands Lowest 2^nd lowest 2^nd highest Highest Job control Lowest 2^nd lowest 2^nd highest Highest Job strain Low strain Active job Passive job High strain	Adj OR Reference group 1.0/0.9 F/M 1.3^b/0.8 F/M 1.1/1.0 F/M 1.0/1.3^† F/M 1.1/1.1 F/M 1.0/1.1 F/M Reference group Reference group 1.1/1.1 F/M 1.0/1.3^† F/M 1.1/1.2^† F/M	Prevalence of sedentary behavior was not linearly associated with quartiles of job demands, job control, or high job strain in either men or women. However, in men, this prevalence was increased in the lowest quartile of job control and in the passive and high-strain groups. In women, this prevalence was significantly increased in the third quartile of job demands.
Choi et al. (2010) Level 6	N = 2,019 (mostly white, well-educated workers; 50% men) Age = 32-69 yr Data came from the National Survey of Midlife Development in the United States (MIDUS) II study	LTPA (2 items): “During your leisure time or free time, how often do you engage in • moderate PA that . . . • vigorous PA that . . . ” The items were specified for the summer and the winter to consider a possible seasonal effect of LTPA	Job demands Lowest 2^nd lowest 2^nd highest Highest Job control Lowest 2^nd lowest 2^nd highest Highest Job strain Passive job High strain Middle Low strain Active job	Adj OR [95% CI] Reference group 0.84 [0.63, 1.13] 0.87 [0.70, 1.12] 0.92 [0.68, 1.24] Reference group 1.09 [0.83, 1.44] 1.18 [0.88, 1.57] 1.60 [1.18, 2.16] Reference group 1.07 [0.77, 1.50] 1.27 [0.96, 1.68] 1.60 [1.14, 2.25] 1.46 [1.06, 2.03]	After controlling for covariates, high job control, active jobs, and low-strain jobs were associated with increased odds of active LTPA. However, the associations varied by sex and education level.
Gimeno et al. (2009) Level 4	N = 6,085 (mostly white, office-based civil servants; 71% men) Age = 35-55 yr 20 civil service departments in London, UK	Low LTPA: <1.5 h/wk of at least moderately energetic LTPA LTPA (2 items): average number of hours per week spent in • moderately energetic PA • vigorous PA	Working in a passive job (cumulated over 3 phases) None 1 phase 2 phases 3 phases	Prevalence ratio (PR) Reference group 0.93/0.99 F/M 1.07/1.14 F/M 1.09/1.34^c F/M	Cumulative effect of passive jobs on low levels of LTPA applied only to men. After controlling for all covariates, the PR for low LTPA in men was 1.16 times greater for workers with 3 reports of passive jobs than for those reporting no passive jobs.
Hellerstedt and Jeffery (1997) Level 6	N = 3,843 (mostly white, professional workers; 49% men) M = 39 yr (men); 38 yr (women) 32 public/ private worksites in Minnesota, USA	Exercise sessions: sum of the responses to 12 items about the weekly frequency of specific heavy and moderate intensity PA	Job demands Lowest 2^nd lowest 2^nd highest Highest Job control Lowest 2^nd lowest 2^nd highest Highest Job strain High strain Low strain Active job Passive job	Adj means 8.2/9.2 F/M 8.2/9.3 F/M 8.7/9.5 F/M 8.7/9.5 F/M 8.1/8.7 F/M 8.5/9.4 F/M 8.7/9.6 F/M 8.8/9.6 F/M 8.3/9.6 F/M 8.0/9.6 F/M 9.2/9.5 F/M 8.3/8.8* F/M	After adjusting for covariates, job control was positively associated with PA in men and women. Men in passive jobs were less likely to engage in PA than other men combined (p = .005). Women in active jobs were more likely to engage in PA than other women combined (p < .001).
Kouvonen et al. (2005) Level 6	N = 46,573 (mostly white-collar workers; 19% men) M = 45 yr (men); 45 yr (women) 31 Finnish public sectors (10 municipalities and 21 hospitals)	LTPA: average amount of time spent per week on leisure and on the journey to and from work in PA corresponding to the activity intensity of walking, vigorous walking, jogging, and running	Job demands Low (Ref) Intermediate High Job control High (Ref) Intermediate Low Job strain Low strain (Ref) Active job Passive job High strain	Adj means 31.2/35.6 F/M 30.6/35.9 F/M 30.8/34.0 F/M 32.5/37.4 F/M 30.9*/35.6 F/M 29.2/32.5 F/M 32.1/38.3 F/M 31.8/34.9 F/M 29.5*/33.1* F/M 29.6*/33.9* F/M	After adjusting for covariates, low job control, high-strain jobs, and passive jobs were associated with lower LTPA in men and women. Active jobs were associated with lower LTPA only in men and older workers. High job demands were associated with lower LTPA only in upper nonmanual workers.
Lallukka et al. (2004) Level 6	N = 6,243 staff (20% men) Age = 40-60 yr Helsinki, the capital city of Finland	LTPA during the past 12 mo An activity MET was computed based on reported hours of different exercise intensities per week Recommended PA level (>30 MET × h/wk) was a dependent variable in logistic regression	Job strain High strain Passive job Low strain Active job Work fatigue Work overtime Physically strenuous work Mentally strenuous work Work-home interface	Adj OR [95% CI] for women only Reference group 0.82 [0.65, 1.02] 1.11 [0.90, 1.36] 1.05 [0.85, 1.30] 0.77 [0.63, 0.93] 1.08 [0.89, 1.30] 1.18 [1.02, 1.35] 0.96 [0.82, 1.13] 1.27 [1.11, 1.44]	After adjusting for covariates, physically active women reported less work fatigue, but more physically strenuous work and satisfaction with work-home interface. In men, none of the work conditions were related to PA in the fully adjusted model.
Lallukka et al. (2008) Level 6	Data from 3 studies n₁ = 3,397 (72% men; Britain) n₂ = 6,070 (20% men; Finland) n₃ = 2,213 (71% men; Japan) Age = 45–60 yr 20 civil service departments in London; city of Helsinki; a local government in Japan	Physical inactivity (assessed by reported amount and intensity of LTPA): ≤1 h/wk of LTPA or the lowest quintile of the distribution of the sum used as a cutoff point	(Britain) Job strain Low strain Active job Passive job High strain Work overtime (Finland) Job strain Low strain Active job Passive job High strain Work overtime	Adj OR (95% CI) for men only Reference group 1.20 [0.85, 1.69] 1.64 [1.18, 2.27] 1.88 [1.30, 2.72] 0.97 [0.75, 1.24] for women only Reference group 0.96 [0.76, 1.20] 1.31 [1.05, 1.62] 1.25 [1.00, 1.57] 1.05 [0.83, 1.31]	After adjusting for covariates, passive jobs and high-strain jobs were associated with physical inactivity among men in Britain and women in Finland. None of the job strain variables were related to physical inactivity in Japanese men and women, British women, and Finnish men.
Landsbergis et al. (1998) Level 6	N = 202 men Age = 30-60 yr 9 public/private sectors in New York, USA	Sedentary behavior (exercise <2 h/wk) assessed by 3 items about regularity of exercise, frequency, and duration	Job control Job demands High job strain	Adj SOR/OR [95% CI] 1.0 [0.7, 1.6] 0.8 [0.5, 1.2] 1.3 [0.4, 4.0]	After adjusting for covariates, sedentary behavior was not associated with job control, job demands, and job strain.
Payne et al. (2002) Level 6	n₁ = 241 (baseline; 70% men) Age = 16–55+ yr n₂ = 213 (one week) A British computer company	Exercise (single item): “What forms of exercise did you do last week and how many hours did you devote to each?”	High-strain jobs (n = 54) Low-strain jobs (n = 53)	M = 1.07 M = 1.93	Workers in high-strain jobs did less exercise than those in low-strain jobs, F(1, 96) = 4.39, p = .04. Job demands and job control had no direct effect on exercise, but intenders who did not exercise reported higher job demands than intenders who did exercise, F(1, 147) = 7.84, p < .01.
Payne et al. (2005) Level 6	n₁ = 331 (baseline; 70% men) Age = 16-64 yr n₂ = 289 (one week) A British computer company	Exercise (single item): “What forms of exercise did you do last week and how many hours did you devote to each?”	Job demands Job control Demands × Control	β = −.15 β = −.15 β = .15	Job demands and job control had no direct effect on exercise, but job demands affected exercise indirectly by reducing perceptions of control over exercise.

Note: yr = year; wk = week; mo = month; PA = physical activity; Adj = adjusted; OR = odds ratio; F/M = females/males; LTPA = leisure-time physical activity; CI = confidence interval;

Ref = reference; SOR = standardized odds ratio; M = mean; β = standardized coefficient.

Rated according to Melnyk and Fineout-Overholt (2011).

The odds ratio was statistically significant.

†

Borderline statistical significance.

The prevalence ratio was statistically significant.

p < .05. **p < .01. ***p < .001.

All studies reviewed relied on self-report measures of all variables, with the exception of one study (Crespo, Sallis, Conway, Saelens, & Frank, 2011) that included an objective measure of physical activity (accelerometer). Seven studies examined associations of work environment factors with leisure-time physical activity. Workplace policies and resources relevant to physical activity were usually assessed using measures of workplace policies (n = 5) and physical environments (n = 5), followed by social environments (n = 2). Job strain factors were most commonly assessed with the job demands and job control scales derived from Karasek’s (1979) Job Content Questionnaire (JCQ; n = 5) or questions similar to JCQ (n = 5). Tables 1 and 2 summarize the level of evidence, samples, type of physical activity outcome, work environment factors measured and their association with physical activity, and major findings. Table 3 presents the level of empirical support, as well as significant and nonsignificant associations between job strain factors and physical activity, stratified by gender.

Table 3.

Association of Job Strain Factors With Physical Activity Among White-Collar Workers, Research Yields Inconsistent Results.

		Significant (p ≤ .05)			Nonsignificant
Factor	Level of empirical support	Women	Men	Both sexes	Women	Men	Both sexes
Job demands	Weak	Brisson (2000)			Hellerstedt (1997) Kouvonen (2005)	Brisson (2000) Hellerstedt (1997) Kouvonen (2005) Landsbergis (1998)	Payne (2002) Payne (2005) Choi (2010)
Job control	Equivocal	Hellerstedt (1997) Kouvonen (2005)	Hellerstedt (1997) Kouvonen (2005)	Choi (2010)	Brisson (2000)	Landsbergis (1998) Brisson (2000)	Payne (2002) Payne (2005)
Low strain	Equivocal	Hellerstedt (1997)	Hellerstedt (1997)	Choi (2010)	Lallukka (2004)	Lallukka (2004)
Active job	Weak	Hellerstedt (1997)	Kouvonen (2005)	Choi (2010)	Brisson (2000)Kouvonen (2005)Lallukka (2004)Lallukka (2008)-BLallukka (2008)-FLallukka (2008)-J	Brisson (2000)Hellerstedt (1997)Lallukka (2004)Lallukka (2008)-BLallukka (2008)-FLallukka (2008)-J
Passive job	Equivocal	Hellerstedt (1997) Kouvonen (2005) Lallukka (2008)-F	Gimeno (2009) Hellerstedt (1997) Kouvonen (2005) Lallukka (2008)-B		Brisson (2000)Gimeno (2009)Lallukka (2004)Lallukka (2008)-BLallukka (2008)-J	Brisson (2000)Lallukka (2004)Lallukka (2008)-FLallukka (2008)-J
High strain	Equivocal	Hellerstedt (1997) Kouvonen (2005) Lallukka (2008)-F	Hellerstedt (1997) Kouvonen (2005) Lallukka (2008)-B		Brisson (2000)Lallukka (2008)-BLallukka (2008)-J	Brisson (2000)Landsbergis (1998)Lallukka (2008)-FLallukka (2008)-J	Choi (2010)

Note: Articles are identified by the first author. The author’s name in italics under the column of significant represents a negative relationship between a job strain factor and leisure-time physical activity or exercise. The article by Lallukka et al. (2008) consisted of three studies conducted in Britain, Finland, and Japan. B = Britain; F = Finland; J = Japan.

Workplace Physical Activity Policies/Resources and Physical Activity

Four studies examined supportive workplace policies/resources for physical activity, and results generally suggest a weak positive relationship with level of physical activity. Supportive workplace policies/resources for physical activity were positively related to moderate-to-vigorous physical activity (Crespo et al., 2011; Dodson, Lovegreen, Elliott, Haire-Joshu, & Brownson, 2008), leisure-time physical activity (Crespo et al., 2011; Lucove, Huston, & Evenson, 2007; Prodaniuk, Plotnikoff, Spence, & Wilson, 2004), and work-break/workplace physical activity (Lucove et al., 2007; Prodaniuk et al., 2004). For workplace policies supporting physical activity, subsidies for health club use were positively related to leisure-time physical activity and work-break physical activity, and paid time for non-work-related physical activity was positively related to work-break physical activity. For workplace resources supporting physical activity, having access to onsite exercise facility was positively related to leisure-time physical activity, and a safe place to walk outside work was positively related to work-break physical activity (Lucove et al., 2007). Workers with accessible exercise facilities and equipment, stairways, and personal services (e.g., fitness testing and counseling) at worksites were more likely to meet physical activity recommendations than those without these resources. For the cumulative effects, the number of supportive elements in the environment was associated with an increased likelihood of meeting physical activity recommendations (Dodson et al., 2008). In addition, there is evidence that the relationship between supportive workplace policies/resources and workplace physical activity was partially mediated by self-efficacy. However, outcome expectations did not mediate the effects of supportive policies/resources on physical activity (Prodaniuk et al., 2004).

A few results from this body of research were inconsistent with the proposed relationship between supportive workplace policies/resources and physical activity. For example, Lucove et al. (2007) did not find a relationship between supportive policies/resources and recommended levels of physical activity. Crespo et al. (2011) found that supportive workplace policies/resources were related positively to higher total sedentary behavior and negatively to job-related physical activity.

One study (Kaczynski, Bopp, & Wittman, 2010) examined the association between workplace physical and cultural supports for active commuting and employee active commuting behavior. Factors associated with walking or bicycling to work included having coworkers who actively commute to work and bicycle parking and storage policies in the workplace. In the cumulative effects, workers reporting more supports were more likely to actively commute at least once a week, especially women (Table 1).

In summary, our findings provide empirical support for the importance of supportive workplace physical activity policies and resources in promoting employee physical activity. Five studies found a significant positive relationship between overall supportive policies/resources (Crespo et al., 2011; Prodaniuk et al., 2004) or cumulative policies/resources (Dodson et al., 2008; Kaczynski et al., 2010; Lucove et al., 2007) and physical activity (including leisure-time, moderate-to-vigorous, work-break/workplace, and recommended physical activity or active commuting). In reality, however, workplace physical activity policies and resources are not yet generally available. Dodson et al. (2008) found that 69% of 977 U.S. workers reported no policies or resources that support physical activity in their workplace. The most commonly reported policies or resources were accessible stairways (57%) and a safe place to walk outside work (56%), whereas paid time for non-work-related physical activity (15%), facilities (15%), and equipment (11%) were the least commonly reported (Dodson et al., 2008; Lucove et al., 2007).

Job Strain Factors and Physical Activity

The relationships between job strain factors and physical activity were inconsistent and varied across gender and countries (Tables 2 and 3). The level of support for these relationships is presented in Table 3. For example, U.S. women with active jobs were more likely to be physically active (Hellerstedt & Jeffery, 1997), while Finnish men and older workers with active jobs were less likely to be active (Kouvonen et al., 2005). In the United States, however, highly educated men with active jobs were more likely to be physically active during leisure time (Choi et al., 2010). Men (in higher education) with low-strain jobs were more likely to be active (Choi et al., 2010; Hellerstedt & Jeffery, 1997), while women with low-strain jobs were less likely to be active (Hellerstedt & Jeffery, 1997). However, less educated women with low-strain jobs were more likely to be physically active during leisure time (Choi et al., 2010). White-collar workers from Britain, Finland, and Japan demonstrated inconsistent and weak associations between job strain factors and physical inactivity across gender and all three countries (Lallukka et al., 2008).

Nevertheless, in some studies, high job control had a relatively consistent positive relationship with leisure-time physical activity or exercise (Choi et al., 2010; Hellerstedt & Jeffery, 1997; Kouvonen et al., 2005), whereas passive jobs and high-strain jobs had a relatively consistent negative relationship with leisure-time physical activity or exercise, controlling for covariates (Gimeno et al., 2009; Hellerstedt & Jeffery, 1997; Kouvonen et al., 2005; Lallukka et al., 2008). In other studies, job demands and job control were not associated with exercise (Landsbergis et al., 1998; Payne, Jones, & Harris, 2002, 2005). However, Payne et al. (2002) indicated that people who intended to exercise but did not exercise reported higher job demands than intenders who did exercise, and Payne et al. (2005) further suggested that job demands affected exercise indirectly by reducing perceptions of control over exercise. In brief, the level of empirical support for job strain factors suggests only equivocal (e.g., job control, passive job, high strain, and low strain) or weak (e.g., job demands and active job) support.

Discussion

In this review, organizational factors include work environment factors, specifically workplace physical activity policies/resources and job strain factors. To our knowledge, this is the first review to synthesize and evaluate the extant literature examining the relationship between work environment factors and physical activity in predominantly white-collar workers. Overall, positive, albeit weak, associations between supportive workplace policies/resources and different types of physical activity were identified, providing relatively consistent empirical support. Job strain factors had some independent, but relatively small, effects on leisure-time physical activity or exercise, with job control, passive jobs, and high-strain jobs demonstrating stronger, though equivocal evidence. Other job strain factors were observed to have insufficient evidence for predicting physical activity, such as low-strain jobs, active jobs, and job demands. These merit further study. Moreover, we examine job strain as one of the psychosocial work environment factors, there might be other psychosocial work environment factors (e.g., low social support and low reward) that require further exploration. Finally, this is what we know about the organizational factors; future research may consider individual factors to better understand the whole picture.

Much of the research on worker physical activity to date has focused on job strain factors, but researchers are beginning to explore the potential influence of workplace physical activity policies and resources on worker physical activity. The paucity of research examining the effects of supportive workplace policies/resources on physical activity is likely influenced by the challenge of operationalizing key constructs and the lack of theoretically based and psychometrically sound measures of workplace physical activity policies/resources. Only one study (Prodaniuk et al., 2004) used a theoretical framework (social cognitive theory and ecological workplace physical activity model) to guide their research endeavors and used a reliable and valid measure of perceived workplace environment. In contrast, the theory and measurement of job strain constructs are well-developed, such as Karasek’s (1979) demand/control model and JCQ. More effort is needed to establish a shared understanding of the conceptual and operational definitions and theoretical frameworks to advance empirical examination of the concept of workplace physical activity policies and resources.

None of the reviewed studies examined the combined effects of workplace physical activity policies/resources and job strain factors on worker physical activity. Examining both the perspectives simultaneously will contribute to our understanding. Furthermore, only one study (Prodaniuk et al., 2004) used mediator analyses, which allowed the researchers to disentangle the causal direction and meaning of the relationship between the predictor variable and the outcome variable. Although limited literature prevents a definitive conclusion, the self-efficacy construct may have use as a mediator of physical activity. There have been recent calls to examine the potential mediation effects of psychosocial factors on the relationship of environmental factors and physical activity, in an attempt to guide development of interventions (Baranowski, Cullen, Nicklas, Thompson, & Baranowski, 2003; Brug, Oenema, & Ferreira, 2005; Kremers et al., 2006).

The research exhibits diversity in the population groups and research settings across studies that warrant careful consideration. The vast majority of studies reviewed were conducted in North America (52.9%) and Europe (41.2%); only one study was conducted in Asia (Japan). Although the target population in this review is predominantly white-collar workers (mostly well-educated, office-based, and sedentary workers), variations in job characteristics and workplace environments are expected within and across studies. These variations can affect workers’ levels of physical activity. For example, nurses’ job tasks are very different from that of other workers studied (Kouvonen et al., 2005). Nurses are likely more physically active than office workers during the workday, whereas on a nonwork day, office workers are likely more active than nurses. These differences may account for some inconsistent findings in some of the studies reviewed here.

Methodological and conceptual issues related to sample size, potential confounding factors, and the demand/control model may have contributed to the mixed results for job strain factors. In two studies, an inadequate subgroup sample size may have contributed to nonsignificant results (Lallukka et al., 2004, 2008; Landsbergis et al., 1998). In another study, results may have been confounded by the fact that subjects who participated in the study were more physically active than those who did not participate, thereby affecting the observed relationship between job strain factors and physical activity (Gimeno et al., 2009). Hellerstedt and Jeffery (1997) noted that the demand/control model weights demands and control equally and assumes linear relationships between demands and control and health behaviors, which may have produced misclassification errors in the job strain group, leading to an underestimation of the true association.

Methodological issues must be considered when synthesizing diverse sources of evidence. There was consistency in the way similar variables were defined in individual studies. Yet, there was less consistency in the specific measures used to assess the variables. The variables of physical activity and workplace physical activity policies/resources were measured differently across studies. The studies assessed job strain factors using Karasek’s (1979) job demands and job control scales or similar ones, but there were individual variations in the number of items, response options, item content, and reliability. These differences in measures make synthesis and generalizations across studies difficult and highlight the need for standardized and validated measures to be used across studies to determine the theoretically consistent patterns of findings.

It was interesting to note that supportive workplace policies/resources had a positive effect on leisure-time physical activity and sedentary behavior but a negative effect on job-related physical activity (Crespo et al., 2011). As suggested by the authors, it may be that supportive policies/resources were most likely to influence workers’ leisure-time physical activity (at work and away from work), and employers were more likely to provide supportive policies/resources for physical activity for their sedentary workers.

There are several limitations to this review. First, the review was limited to published articles. This may be subject to a publication bias in which studies with larger effects are more likely to be reported than those with smaller effects. Given the relatively small effects in the findings of this review, the bias may be small. Second, the review was limited to journal articles in English. Thus, not all relevant research papers were included. Third, the review was limited to the search terms, databases, and other search sources described in the methods section. Fourth, all but one of the studies reviewed were descriptive or exploratory (relatively low level of evidence). The use of a cross-sectional design can help identify the most salient constructs for further study, but questions of causality cannot be demonstrated. Finally, this review focused on white-collar workers, and the subjects in each study were “mostly” white-collar workers but not exclusively white-collar workers. Thus, one must be cautious about generalizing the findings to other groups of white-collar workers.

The results of this study offer implications for clinical practitioners. First, practitioners should consider creating supportive policies/resources for physical activity for inclusion in workplace physical activity interventions. Second, supportive policies/resources related to employee physical activity should be considered for inclusion in workplace policies, guidelines, and/or regulations. Third, redesigning jobs to reduce adverse psychosocial work environment (job strain) factors, such as high psychological work demands and low decision latitude, may influence work culture and reduce barriers to physical activity or exercise. The information provided in this review may provide some guidance to occupational health nurses in terms of a limited number of potentially effective work environment factors that could be addressed to help increase physical activity of white-collar workers. The results may also be used to identify features of supportive policies/resources for physical activity and to develop strategies that integrate physical activity promotion into the workplace culture.

In conclusion, workplace physical activity policies and resources appear to have a more consistent association with different types of physical activity than job strain factors. This finding supports the rising trend of creating supportive policies/resources as a sustainable approach to promoting employee health or physical activity in the workplace, aiming to maintain a healthy and productive workforce and workplace. Nevertheless, given the evidence we reviewed, more research is needed to confirm the effect of work environment factors on physical activity. Specifically, to clarify the direction of causality with regard to these relationships, we recommend a longitudinal design with a large population of white-collar workers that would reduce the confounding between predictor and outcome variables. Furthermore, future research should use reliable and valid measures to more accurately evaluate predictor and outcome variables. Additional research needs to clearly define and improve measurements of workplace physical activity policies and resources to facilitate an empirical examination of how these policies and resources affect workers’ physical activity.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Andersen

L. B.

Schnohr

Schroll

Hein

H. O.

(2000). All-cause mortality associated with physical activity during leisure time, work, sports, and cycling to work. Archives of Internal Medicine, 160, 1621-1628.

Baranowski

Cullen

K. W.

Nicklas

Thompson

Baranowski

(2003). Are current health behavioral change models helpful in guiding prevention of weight gain efforts? Obesity Research, 11 Suppl, 23S-43S. doi:10.1038/oby.2003.222

Brisson

Larocque

Moisan

Vezina

Dagenais

G. R.

(2000). Psychosocial factors at work, smoking, sedentary behavior, and body mass index: A prevalence study among 6995 white-collar workers. Journal of Occupational & Environmental Medicine, 42, 40-46.

Brug

Oenema

Ferreira

(2005). Theory, evidence and intervention mapping to improve behavior nutrition and physical activity interventions. International Journal of Behavioral Nutrition and Physical Activity, 2, 2.

Campbell

G. B.

Matthews

J. T.

(2010). An integrative review of factors associated with falls during poststroke rehabilitation. Journal of Nursing Scholarship, 42, 395-404. doi:10.1111/j.1547-5069.2010.01369.x

Choi

Schnall

P. L.

Yang

Dobson

Landsbergis

Israel

. . . Baker

(2010). Psychosocial working conditions and active leisure-time physical activity in middle-aged U.S. workers. International Journal of Occupational Medicine and Environmental Health, 23, 239-253. doi:10.2478/v10001-010-0029-0

Conn

V. S.

Hafdahl

A. R.

Cooper

P. S.

Brown

L. M.

Lusk

S. L.

(2009). Meta-analysis of workplace physical activity interventions. American Journal of Preventive Medicine, 37, 330-339. doi:10.1016/j.amepre.2009.06.008

Crespo

N. C.

Sallis

J. F.

Conway

T. L.

Saelens

B. E.

Frank

L. D.

(2011). Worksite physical activity policies and environments in relation to employee physical activity. American Journal of Health Promotion, 25, 264-271. doi:10.4278/ajhp.081112-QUAN-280

Dishman

R. K.

Oldenburg

O’Neal

Shephard

R. J.

(1998). Worksite physical activity interventions. American Journal of Preventive Medicine, 15, 344-361.

10.

Dodson

E. A.

Lovegreen

S. L.

Elliott

M. B.

Haire-Joshu

Brownson

R. C.

(2008). Worksite policies and environments supporting physical activity in Midwestern communities. American Journal of Health Promotion, 23, 51-55. doi:10.4278/ajhp.07031626

11.

Engbers

L. H.

van Poppel

M. N.

Chin

Paw

M. J.

van Mechelen

(2005). Worksite health promotion programs with environmental changes: A systematic review. American Journal of Preventive Medicine, 29, 61-70. doi:10.1016/j.amepre.2005.03.001

12.

Gates

Brehm

Hutton

Singler

Poeppelman

(2006). Changing the work environment to promote wellness: A focus group study. AAOHN Journal, 54, 515-520.

13.

Gimeno

Elovainio

Jokela

De Vogli

Marmot

M. G.

Kivimaki

(2009). Association between passive jobs and low levels of leisure-time physical activity: The Whitehall II cohort study. Occupational and Environmental Medicine, 66, 772-776. doi:10.1136/oem.2008.045104

14.

Hamilton

M. T.

Hamilton

D. G.

Zderic

T. W.

(2007). Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Diabetes, 56, 2655-2667. doi:10.2337/db07-0882

15.

Hellerstedt

W. L.

Jeffery

R. W.

(1997). The association of job-strain and health behaviors in men and women. International Journal of Epidemiology, 26, 575-583.

16.

Jans

M. P.

Proper

K. I.

Hildebrandt

V. H.

(2007). Sedentary behavior in Dutch workers: Differences between occupations and business sectors. American Journal of Preventive Medicine, 33, 450-454.

17.

Kaczynski

A. T.

Bopp

M. J.

Wittman

(2010). Association of workplace supports with active commuting. Preventing Chronic Disease, 7, A127.

18.

Karasek

R. A.

(1979). Job demands, job-decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285-308.

19.

Karasek

Brisson

Kawakami

Houtman

Bongers

Amick

(1998). The Job Content Questionnaire (JCQ): An instrument for internationally comparative assessments of psychosocial job characteristics. Journal of Occupational Health Psychology, 3, 322-355.

20.

Kouvonen

Kivimaki

Elovainio

Virtanen

Linna

Vahtera

(2005). Job strain and leisure-time physical activity in female and male public-sector employees. Preventive Medicine, 41, 532-539. doi:10.1016/j.ypmed.2005.01.004

21.

Kremers

S. P.

de Bruijn

G. J.

Visscher

T. L.

van Mechelen

de Vries

N. K.

Brug

(2006). Environmental influences on energy balance-related behaviors: A dual-process view. International Journal of Behavioral Nutrition and Physical Activity, 3, 9. doi:10.1186/1479-5868-3-9

22.

Lallukka

Lahelma

Rahkonen

Roos

Laaksonen

Martikainen

. . . Kagamimori

(2008). Associations of job strain and working overtime with adverse health behaviors and obesity: Evidence from the Whitehall II Study, Helsinki Health Study, and the Japanese Civil Servants Study. Social Science & Medicine, 66, 1681-1698. doi:10.1016/j.socscimed.2007.12.027

23.

Lallukka

Sarlio-Lahteenkorva

Roos

Laaksonen

Rahkonen

Lahelma

(2004). Working conditions and health behaviors among employed women and men: The Helsinki Health Study. Preventive Medicine, 38, 48-56.

24.

Landsbergis

P. A.

Schnall

P. L.

Deitz

D. K.

Warren

Pickering

T. G.

Schwartz

J. E.

(1998). Job strain and health behaviors: Results of a prospective study. American Journal of Health Promotion, 12, 237-245.

25.

Lucove

J. C.

Huston

S. L.

Evenson

K. R.

(2007). Workers’ perceptions about worksite policies and environments and their association with leisure-time physical activity. American Journal of Health Promotion, 21, 196-200.

26.

Matson-Koffman

D. M.

Brownstein

J. N.

Neiner

J. A.

Greaney

M. L.

(2005). A site-specific literature review of policy and environmental interventions that promote physical activity and nutrition for cardiovascular health: What works? American Journal of Health Promotion, 19, 167-193.

27.

McLeroy

K. R.

Bibeau

Steckler

Glanz

(1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15, 351-377.

28.

Melnyk

B. M.

Fineout-Overholt

(2011). Evidence-based practice in nursing & healthcare: A guide to best practice. Philadelphia, PA: Lippincott Williams & Wilkins.

29.

Miller

Brown

(2004). Steps and sitting in a working population. International Journal of Behavioral Medicine, 11, 219-224. doi:10.1207/s15327558ijbm1104_5

30.

Patel

A. V.

Bernstein

Deka

Feigelson

H. S.

Campbell

P. T.

Gapstur

S. M.

. . . Thun

M. J.

(2010). Leisure time spent sitting in relation to total mortality in a prospective cohort of U.S. adults. American Journal of Epidemiology, 172, 419-429. doi:10.1093/aje/kwq155

31.

Payne

Jones

Harris

P. R.

(2002). The impact of working life on health behavior: The effect of job strain on the cognitive predictors of exercise. Journal of Occupational Health Psychology, 7, 342-353.

32.

Payne

Jones

Harris

P. R.

(2005). The impact of job strain on the predictive validity of the theory of planned behavior: An investigation of exercise and healthy eating. British Journal of Health Psychology, 10, 115-131. doi:10.1348/135910704x14636

33.

Prodaniuk

T. R.

Plotnikoff

R. C.

Spence

J. C.

Wilson

P. M.

(2004). The influence of self-efficacy and outcome expectations on the relationship between perceived environment and physical activity in the workplace. International Journal of Behavioral, Nutrition and Physical Activity, 1, 7. doi:10.1186/1479-5868-1-7

34.

Pronk

N. P.

(2009). Physical activity promotion in business and industry: Evidence, context, and recommendations for a national plan. Journal of Physical Activity and Health, 6(Suppl. 2), S220-S235.

35.

Pronk

N. P.

Kottke

T. E.

(2009). Physical activity promotion as a strategic corporate priority to improve worker health and business performance. Preventive Medicine, 49, 316-321. doi:10.1016/j.ypmed.2009.06.025

36.

Proper

K. I.

Koning

van der Beek

A. J.

Hildebrandt

V. H.

Bosscher

R. J.

van Mechelen

(2003). The effectiveness of worksite physical activity programs on physical activity, physical fitness, and health. Clinical Journal of Sport Medicine, 13, 106-117.

37.

Sallis

J. F.

Bauman

Pratt

(1998). Environmental and policy interventions to promote physical activity. American Journal of Preventive Medicine, 15, 379-397.

38.

U.S. Department of Health and Human Services. (2008). 2008 physical activity guidelines for Americans. Retrieved from http://www.health.gov/PAGuidelines/pdf/paguide.pdf

39.

Whittemore

Knafl

(2005). The integrative review: Updated methodology. Journal of Advanced Nursing, 52, 546-553. doi:10.1111/j.1365-2648.2005.03621.x

40.

World Health Organization. (2002). The world health report 2002: Reducing risks, promoting healthy life. Retrieved from http://www.who.int/whr/2002/en/whr02_en.pdf