Hostname: page-component-7b7567c8c-kwt9d Total loading time: 0 Render date: 2026-11-15T05:22:38.372Z Has data issue: false hasContentIssue false
  • English
  • Français

Dietary and physical activity behaviours related to obesity-specific quality of life and work productivity: baseline results from a worksite trial

Published online by Cambridge University Press:  06 December 2011

Stephanie Whisnant Cash,
Shirley A. A. Beresford,
Jo Ann Henderson,
Anne McTiernan,
Liren Xiao,
C. Y. Wang and
Donald L. Patrick
Stephanie Whisnant Cash
Affiliation:
Department of Epidemiology, School of Public Health, University of Washington, 1959 NE Pacific Street, Box 357236, Seattle, WA98195, USA
Shirley A. A. Beresford*
Affiliation:
Department of Epidemiology, School of Public Health, University of Washington, 1959 NE Pacific Street, Box 357236, Seattle, WA98195, USAFred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Mailstop M3-B232, Seattle, WA98109, USA
Jo Ann Henderson
Affiliation:
Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Mailstop M3-B232, Seattle, WA98109, USA
Anne McTiernan
Affiliation:
Department of Epidemiology, School of Public Health, University of Washington, 1959 NE Pacific Street, Box 357236, Seattle, WA98195, USAFred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Mailstop M3-B232, Seattle, WA98109, USA
Liren Xiao
Affiliation:
Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Mailstop M3-B232, Seattle, WA98109, USA
C. Y. Wang
Affiliation:
Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Mailstop M3-B232, Seattle, WA98109, USA
Donald L. Patrick
Affiliation:
School of Public Health, University of Washington, Box 359455, 4333 Brooklyn Avenue NE, Room 14-101, Seattle, WA98195, USA
*
*Corresponding author: S. A. A. Beresford, fax +1 206 685 9651, email beresfrd@u.washington.edu
Rights & Permissions [Opens in a new window]

Abstract

Obesity is associated with impaired health-related quality of life (QOL) and reduced productivity; less is known about the effect of dietary factors. The present study investigated how dietary behaviours, physical activity and BMI relate to weight-specific QOL and work productivity. The study was conducted in thirty-one small blue-collar and service industry worksites in Seattle. Participants were 747 employees (33·5 % non-White). Measures included self-reported servings of fruits and vegetables, dietary behaviours such as fast food consumption, Godin free-time physical activity scores, measured height and weight, Obesity and Weight-Loss QOL (OWLQOL) scores, and Work Limitations Questionnaire scores. Baseline data were analysed using linear mixed models separately for men (n 348) and women (n 399), since sex modified the effects. BMI was negatively associated with OWLQOL in both women (P < 0·001) and men (P < 0·001). The linear effect estimate for OWLQOL scores associated with a one-category increase in BMI was 30 (95 % CI 25, 44) % for women and 14 (95 % CI 10, 17) % for men. BMI was positively associated with productivity loss only in women (exp(slope) = 1·46, 95 % CI 1·02, 2·11, P = 0·04). Eating while doing another activity was negatively associated with OWLQOL scores in men (P = 0·0006, independent of BMI) and with productivity in women (P = 0·04, although the effect diminished when adjusting for BMI). Fast-food meals were associated with decreased productivity in men (P = 0·038, independent of BMI). The results suggest that obesogenic dietary behaviours and higher BMI are associated with decreased QOL and productivity to different degrees in women and men.

Keywords

Dietary behavioursObesityQuality of lifeWork productivity

Information

Type
Full Papers
Information
British Journal of Nutrition , Volume 108 , Issue 6 , 28 September 2012, pp. 1134 - 1142
Copyright
Copyright © The Authors 2011

Obesity has become a major public health problem. In the USA, the prevalence of obesity (BMI of 30 kg/m2 or higher) has increased drastically, from 15 % in 1985 to the current 34·3 %(1). It is well established that obesity is associated with a number of adverse health outcomes including hypertension, type 2 diabetes, heart disease, stroke, respiratory problems and certain types of cancer(2). Often overlooked are the equally important relationships between obesity and non-physical health outcomes, such as well-being, productivity and psychosocial functioning.

To accommodate this oversight, Healthy People 2010(3) expanded its 2000 goal of longevity to include increased quality of life (QOL) and overall well-being(Reference Karlsson, Taft and Ryden4); researchers have started collecting such data, and the body of literature concerning obesity and outcomes such as QOL and productivity is rapidly growing. Health-related QOL (HRQOL) has been defined variously, but most often concerns concepts of physical, psychological, and social functioning and well-being. Measures of perceived QOL are confined to those aspects of life known only to the person and are perceptions about how a specific condition affects a person's life in relation to their goals, standards and concerns(Reference Patrick and Erickson5).

Several studies have found that obesity is associated with a diminished HRQOL(Reference Fontaine, Cheskin and Barofsky6, Reference Kolotkin, Crosby and Williams7), and higher BMI has been shown to be associated with physical, social and emotional domains of HRQOL(Reference Karlsson, Taft and Ryden4). Recently, researchers have also examined obesity-specific subscales of HRQOL. Kolotkin et al. (Reference Kolotkin, Crosby and Williams7), using the Impact of Weight on Quality of Life-Lite instrument, showed that individuals with higher BMI were more likely to report impaired obesity-specific HRQOL. Similar impairments in energy and vitality, in addition to increased body pain, have been observed in obese patients(Reference Fontaine, Cheskin and Barofsky6).

Obesity is also associated with reduced productivity(Reference Wolf and Colditz8), with an estimated $73·1 billion in productivity losses due to obesity among full-time employees(Reference Finkelstein, Dibonaventura and Burgess9). While some of this effect may be due to co-morbid conditions such as diabetes and hypertension, evidence suggests that lower productivity is exacerbated by the presence of obesity(Reference Sullivan, Ghushchyan and Ben-Joseph10). BMI appears to independently predict absenteeism, with obese employees missing more days of work than their normal-weight counterparts(Reference Bungum, Satterwhite and Jackson11). With high estimated costs lost due to sickness or disability related to obesity, the prevalence of obesity is a serious economic concern.

Evidence suggests that HRQOL and productivity are modifiable by targeting diet and exercise. Randomised controlled trials have shown that a diet rich in fruits and vegetables is associated with higher HRQOL(Reference Plaisted, Lin and Ard12). Data also suggest that physical activity is associated with improved HRQOL(Reference Vuillemin, Boini and Bertrais13) and fewer physical and mental unhealthy days(Reference Brown, Balluz and Heath14, Reference Brown, Brown and Heath15). In addition, data from the longitudinal Swedish Obesity Study, which used the Swedish Obesity Study Quality of Life Survey, showed that weight loss was associated with improved scores on the obesity-related problems subscale(Reference Karlsson, Taft and Ryden4). Worksite interventions have proven successful in increasing fruit and vegetable consumption and physical activity(Reference Beresford, Thompson and Feng16), and may indirectly increase HRQOL and productivity.

The Move and Moderate in Balance (MOVE `M) study is a worksite-randomised trial to encourage healthy dietary choices and physical activity and maintain or reduce the weight of employees at worksites. The present study explores baseline associations between BMI, fruit and vegetable consumption, physical activity and weight-related QOL and proxy measures for productivity among employees of small South Seattle worksites.

Methods

Recruitment

Worksites in South Seattle were recruited from a defined geographical area and were identified using US Standard Industrial Classification two-digit codes(17), including construction, manufacturing, transportation and service industries. Worksites with between fifteen and eighty employees were recruited. During a screening call, additional criteria were checked: a high proportion of employees with sedentary or stationary occupations (>50 %), a low turnover rate ( < 30 %); a low proportion of non-English-speaking employees ( < 30 %); and a breakdown of ≤ 33 % management/professional staff and ≥ 30 % in blue- or pink-collar positions.

Eligible interested worksites were involved in a run-in phase that required a letter of intent to participate, an employee list for survey administration and a high baseline survey response rate. Worksites were recruited on a rolling basis, constituting four waves with thirty-two worksites from 2006 to 2008. Each wave contains six to ten worksites (three to five pairs) randomised to either intervention or control arms. The present study includes data from thirty-one worksites (one worksite was not randomised) at baseline only.

The study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Institutional Review Boards at the University of Washington and the Fred Hutchinson Cancer Research Center. Written informed consent was obtained from all participants.

Baseline measurements

Baseline surveys were solicited from all employees at participating worksites on an index date agreed with the company. Each site was given an additional 2 weeks to collect surveys from those who were not present at the proctored group meeting on the index date. If a site did not reach the 70 % enrolment goal, two additional attempts were made to collect surveys from non-responders. The third and final attempts to collect survey data used a shorter version of the survey containing fifteen vital information questions. Additional measures (i.e. blood draw and pedometer assessment) were taken on a randomly selected subsample of about twenty individuals who had agreed on the survey to be re-contacted (or a minimum of 50 % of survey respondents).

Assessment of BMI

BMI (kg/m2) was calculated using measured height and weight and was grouped according to standard categories: ‘underweight’; ‘normal’; ‘overweight’; ‘obese’(18). Measurements were conducted at the baseline survey administration for everyone who completed a survey, and at the date of a blood draw at the site for those missing these objective measures.

Assessment of dietary behaviours

Daily servings of fruits and vegetables were assessed via the survey, which provided pictorial representations of typical fruit and vegetable servings. Summary fruit and vegetable questions in a FFQ format(Reference Thompson and Byers19) assessed consumption of fruit, 100 % fruit juice, other fruit juice, non-French-fried potatoes, non-potato vegetables and green salad. Fruit and vegetable consumption has consistently been shown to be associated with reduced BMI(Reference Lin and Morrison20). Several eating behaviours have been shown to be associated with the risk of obesity, including consumption of fast food(Reference French, Harnack and Jeffery21) and soft drinks(Reference Liebman, Pelican and Moore22), television viewing(Reference Liebman, Pelican and Moore22), and eating while doing other activities(Reference Liebman, Pelican and Moore22). The questionnaire included items on frequency of eating at fast-food restaurants, frequency of drinking sodas, and eating while doing another activity such as watching television.

Assessment of physical activity

Physical activity was quantified as free-time physical activity of at least 10 min and was assessed using a modified Godin Leisure-Time Exercise Questionnaire(Reference Godin and Shephard23), which used a cut-point of 15 min. The change, made in common with other centres in the Obesity Prevention in Worksites consortium, is consistent with the new recommendations(24). Godin & Shephard(Reference Godin and Shephard25) established its reliability (test–retest correlation coefficient ranging from 0·48 for light activity to 0·94 for strenuous activity) and validity in relation to maximal oxygen consumption. At least one additional study has confirmed reliability and validity(Reference Jacobs, Ainsworth and Hartman26). Their published composite free-time exercise score was computed for each individual.

Assessment of obesity-specific quality of life

Obesity-specific QOL (HRQOL) was measured using the Obesity and Weight-Loss QOL (OWLQOL) Questionnaire(Reference Niero, Martin and Finger27, Reference Patrick, Bushnell and Rothman28). The OWLQOL is a seventeen-item self-administered questionnaire that was developed cross-culturally to address shortcomings of pre-existing instruments that focus on functional status or behaviours associated with obesity. The theoretical foundation for this instrument is Maslow's needs hierarchy(Reference Maslow29). All items tap unobservable needs such as freedom from stigma and attainment of culturally appropriate goals. For example, respondents are asked the degree to which weight affects their energy with statements such as ‘I feel frustrated that I have less energy because of my weight’ and ‘My weight prevents me from doing what I want to do’. Reliability and validity are well established and have exceeded recommended minimums within obese populations(Reference Patrick, Bushnell and Rothman28). Guttman–Cronbach's α was 0·96, indicating a high internal consistency, and factor analyses supported an overall score(Reference Patrick, Bushnell and Rothman28). Although Cronbach's α of 0·96 is very high, it refers to the total scale and 6000 data points indicated a high α for each of many combinations of items, but the seventeen that were chosen reflected the best content(Reference Patrick and Whisnant30). Excluding several items reduced the α below 0·90, indicating reduced internal consistency(Reference Patrick and Whisnant30). The composite OWLQOL score was computed for each individual and used as the obesity-specific QOL outcome.

Assessment of productivity

Productivity was assessed using the eight-item short form of the Work Limitations Questionnaire (WLQ)(Reference Lerner, Amick and Rogers31). The WLQ is a self-administered questionnaire designed to measure the degree to which health problems may interfere with performing job functions(Reference Lerner, Amick and Rogers31). For instance, respondents are asked ‘In the past 2 weeks, how much of the time did your physical health or emotional problems make it difficult for you to do the following? (1) Get going easily at the beginning of the workday, (2) Handle the workload’. The reliability and validity of the WLQ are well established(Reference Lerner, Amick and Lee32). The WLQ is considered to be a good measure of self-reported work productivity, as self-reported work limitations have been shown to be significantly (P < 0·0001) associated with employee work productivity(Reference Lerner, Amick and Lee32). The overall WLQ Productivity Loss Score(Reference Lerner, Rogers and Chang33) was computed, where a higher score indicates greater limitations in work productivity.

Statistical methods

Baseline demographic characteristics of the study subjects with no missing values for sex were computed. Associations between the log-transformed outcomes (OWLQOL scores and productivity loss scores) and the predictors were assessed. Predictors included log-transformed BMI, physical activity (Godin score), daily servings of fruits and vegetables, doing something while eating, fast-food restaurant meals per month and soft drink consumption. Multiplicative interaction terms were generated by creating a cross-product term between each predictor and sex, and tested for significance in a univariate model using a likelihood ratio test. Sex was found to modify the relationship between OWLQOL and all six predictors (each P < 0·001), and between some predictors and WLQ productivity scores, so the OWLQOL and WLQ scores were analysed by sex. For both outcomes, adjusted geometric means and 95 % CI were obtained using linear mixed models for fixed age, race, education, and income effects and random worksite effects. The geometric mean is a common alternative way of presenting an average with an emphasis on central tendency, and is the appropriate choice when an outcome is log-transformed. It is similar to the arithmetic mean but is calculated by multiplying the numbers together (not summing, as in the arithmetic mean), and then dividing the resulting product by the nth root (where n is the count of numbers in the set). Regression analyses were based on individuals with complete data on age, race, education, income, and the predictor and outcome of interest for each regression; thus, the number of individuals varied for each analysis. The outcomes were log-transformed to reduce skewness and back-transformed for ease of interpretation.

Next, OWLQOL scores and productivity loss scores were examined for a linear trend within each of the ordered categorical predictor variables described above, adjusting for age, race, education and income. Race was collapsed into five categories (non-Hispanic White, Hispanic, Asian, Black or African American, and Other) when used as an adjustment variable. Logarithmically transformed slope estimates and their corresponding standard errors were obtained, and then back-transformed by taking the exponential of the slopes and corresponding 95 % CI. The exponential of the slope represented the relative change of the outcome when the predictor increased one unit. Finally, the effects of physical activity and obesogenic dietary behaviours were assessed when controlling for BMI, which was logarithmically transformed to reduce the effect of skewness.

All statistical tests were two-sided. Statistical analyses were conducted using Statistical Analysis Systems software (version 9.1; SAS Institute, Inc.).

Results

Results include all four randomisation waves of MOVE `M (thirty-one worksites). The average worksite had twenty-seven employees, similar to the non-randomised companies in the recruiting pool. The most common worksite size was fifteen to twenty-four employees for randomised companies and ten to fourteen for non-randomised companies. Randomised worksites reflected the Standard Industrial Classification code distribution of the underlying recruiting pool.

The original dataset of employees in randomised companies who completed the baseline questionnaire (n 753) was restricted to individuals with non-missing values for sex (n 747, 99·2 % of original dataset). A slightly greater proportion of males had a college degree, whereas a slightly greater proportion of females had a postgraduate or professional degree (Table 1). The majority of both female and male employees were non-Hispanic White (65·4 and 66·8 %, respectively), reflecting the general population composition in the Pacific Northwest. Approximately 61 % (n 459) of employees were either overweight (25·8 %) or obese (35·6 %). The average BMI, servings of fruits and vegetables, free-time physical activity scores and WLQ scores were similar between males and females; however, a higher proportion of men were overweight when compared with women. Males had a slightly higher average OWLQOL score (86·0) than females (69·5).

Table 1Baseline demographic characteristics of employees at Move and Moderate in Balance worksites

(Number of participants and percentages; mean values and standard deviations)

GED, General Equivalency Diploma.

* Percentages may not sum to 100 % due to rounding or missing data.

† Other includes Native American or Alaskan Native, Hawaiian or Pacific Islander, or self-specified.

‡ Arithmetic mean.

Examining OWLQOL as the outcome of interest, for all predictors, there were significant interactions with sex (each P < 0·0001). Thus, all OWLQOL analyses were presented as sex-specific (Table 2). OWLQOL scores were negatively associated with BMI for both females (P < 0·0001) and males (P < 0·0001). The negative impact of obesity on OWLQOL scores was more pronounced for women. An additional unit in BMI was associated with a 30 % lower (95 % CI 25, 34 %) geometric mean OWLQOL for women and a 14 % lower (95 % CI 10, 17 %) geometric mean OWLQOL for men.

Table 2Mean Obesity and Weight-Loss Quality-of-Life (OWLQOL) scores by baseline reported dietary behaviours, physical activity and BMI

(Mean values and 95 % confidence intervals)

Physical activity was positively associated with OWLQOL only in women (P = 0·0011). An additional unit in free-time physical activity score was associated with a 13 % positive difference (95 % CI 5, 22 %) in geometric mean OWLQOL scores. However, this effect became non-significant when controlling for BMI. Servings of fruits and vegetables, soda consumption and fast food consumption were not significantly associated with OWLQOL scores for males or females. Eating while doing other activities was significantly associated with OWLQOL scores in men only (P = 0·0006). The geometric mean OWLQOL score for men who ate more frequently while doing other activities was 7 % lower (95 % CI 3, 11 %) than men who ate less frequently while doing other activities. This association remained significant after controlling for BMI but became attenuated (exp(slope) = 0·95, 95 % CI 0·92, 0·98, P = 0·0028). All other effects remained the same when controlling for BMI. All associations and effect sizes persisted after controlling for age, race, education and income.

Interactions were again present when examining the WLQ score as the outcome, so these analyses were also presented as sex-specific (Table 3). BMI was positively associated with WLQ scores only in women (P = 0·04). A one-category difference in BMI was associated with a 1·46-fold (95 % CI 1·02, 2·11) positive difference in geometric mean WLQ scores. The association between BMI and WLQ scores was non-significant in men (P = 0·06). Eating while doing another activity was positively associated with WLQ scores only in women (P = 0·04). A one-category difference was associated with a 1·59-fold (95 % CI 1·01, 2·50) positive difference in geometric mean WLQ scores; however, this effect diminished when controlling for BMI. Fast food consumption was significantly positively related to WLQ scores in men (P = 0·038). A one-category difference was associated with a 1·56-fold (95 % CI 1·03, 2·37) positive difference in geometric mean WLQ scores. This association remained significant when controlling for BMI. None of the remaining predictors was associated with productivity loss scores for males or females. All other effects remained the same when controlling for BMI, and all associations and effect sizes persisted after controlling for age, race, education and income.

Table 3Mean Work Limitations Questionnaire (WLQ) Productivity Loss Score by baseline reported dietary behaviours, physical activity and BMI

(Mean values and 95 % confidence intervals)

Discussion

Results from the present study suggest that BMI is inversely associated with self-reported, obesity-specific HRQOL and work productivity. Certain obesogenic behaviours were also related to obesity-specific HRQOL and productivity. For women, free-time physical activity was associated with higher obesity-specific QOL, and eating while doing another activity was associated with lower productivity, but these associations appeared to operate through BMI. For men, eating while doing another activity was inversely associated with obesity-specific QOL and fast-food meals were inversely associated with productivity; however, unlike in women, these effects did not appear to operate through BMI.

The suggestion of a sex difference appears to be supported by the research in this area. As was expected from the literature(Reference Patrick, Bushnell and Rothman28, Reference Kolotkin, Head and Hamilton34–Reference Kolotkin, Crosby and Kosloski36), women in the present study were more likely to report lower obesity-specific QOL. In addition, the results presented agree with previous studies that found the impact of obesity on HRQOL to be more severe for women than for men(Reference Kolotkin, Head and Hamilton34, Reference Larsson, Karlsson and Sullivan37). Similarly, prior results have found an effect of physical activity on HRQOL for both men and women(Reference Brown, Balluz and Heath14, Reference Brown, Brown and Heath15), and within women alone(Reference Brown, Mishra and Lee38). The findings also complement a study by Muennig et al. (Reference Muennig, Jia and Lee39) who found the desire to lose weight was a stronger predictor of unhealthy days for women than for men, concluding that negative body image contributes to the morbidity of obesity, particularly for women. Indeed, one potential explanation for this sex difference is body image. The social acceptability of body size can vary between women and men; at least one study has shown that overweight men are significantly more likely to report that their body size was socially acceptable than overweight women(Reference Rand and Resnick40). Women also tend to be more heavily affected by body image psychologically than men, associating body dissatisfaction with self-esteem(Reference Furnham, Badmin and Sneade41), and by selecting body images thinner than their own as being more desirable to the opposite sex(Reference Demarest and Allen42). Thus, it is possible that the positive effects of healthier dietary and physical activity behaviours on QOL and productivity are conditional on body size perceptions only among women.

The evidence presented, combined with current economic evidence, may help to convince management to invest in the health of their employees. In particular, the findings may have important implications for companies and organisations, which may want to consider offering healthier choices and lunch breaks for employees in order to maintain or increase productivity. It is especially important for the MOVE `M study, as this could provide incentive for worksites to join the study if they believe that decreasing employees' BMI as a result of the intervention's effort may increase productivity.

The present study has several limitations. All variables, with the exception of BMI, were assessed using self-report survey data. However, it is important to note that QOL can only be measured by reports by the individual and perceived productivity may be an equally or more important measure compared with productivity measures such as absenteeism, lost wages and replacement staff from the employer. Due to the cross-sectional nature of these data, one cannot draw conclusions concerning causality. Thus, it is unclear whether BMI and physical activity may influence obesity-specific QOL, or if this relationship is reversed. Weight-related QOL could well be a predictor of weight-related behaviours. Due to the small number of underweight individuals, the data could not be used to verify previous research(Reference Ford, Moriarty and Zack43, Reference Jia and Lubetkin44), suggesting that individuals at either end of the spectrum, underweight or overweight/obese, tend to have compromised self-reported HRQOL. Nonetheless, whereas current studies have focused on general HRQOL, the present study confirms a similar negative impact of obesity on obesity-specific HRQOL. Finally, although the most common worksite size of randomised companies was slightly larger than non-randomised companies, this is unlikely to affect the generalisability of these results.

The study has several noteworthy strengths. The large sample size allows us to draw strong conclusions. In addition, height and weight were directly measured for all employees, resulting in BMI calculations that are more accurate than those obtained from self-reported height and weight. Using linear mixed models, worksite-level effects were controlled for in order to examine individual-level associations. Finally, the study design will allow us to evaluate the effect of the intervention, particularly in relation to the measures described in the present study. The hypothesis is that the intervention will increase fruit and vegetable consumption and physical activity and decrease BMI, and that a related improvement in OWLQOL and productivity scores will be observed. Nonetheless, this determination cannot be made based on cross-sectional data, and evaluation of these outcomes at 2-year follow-up may shed more light on this potential mediating mechanism.

These results suggest that obesity-specific HRQOL and productivity may be particularly impaired in obese men and women, with the impact being more pronounced in women. Physical activity and obesogenic behaviours may operate through BMI to affect obesity-specific QOL and productivity in women, but obesogenic behaviours appear to affect obesity-specific QOL and productivity independent of BMI in men. The implication of sex differences in the effects of obesity, physical activity and obesogenic behaviours on obesity-specific HRQOL and productivity is important and should be further researched. Such information may prove useful in addressing the mental health and overall well-being of obese women and men.

Acknowledgements

The authors wish to acknowledge the employees at the worksites who took the time to complete the baseline survey and physical measurements for the MOVE `M study. Their voluntary participation made this investigation possible. The present study was supported by funding from the US National Cancer Institute as part of U54 CA116847. All authors contributed to this manuscript. S. W. C. conceived the study and wrote the majority of the manuscript. S. A. A. B. is the principal investigator of the MOVE `M project, wrote the protocol, guided the assessment development, and monitored the research process for this paper from its inception by providing methodological and statistical guidance and participating in manuscript revisions. J. A. H. assisted in writing the Methods section, particularly concerning the recruitment of worksites. A. M. edited the manuscript and acted as the principal investigator of the collection of studies of which this study was a part. L. X. conducted all analyses and assisted in the write-up of the Statistical methods section. C. Y. W. contributed to the statistical methods section and aided in analyses. D. L. P. provided guidance on patient-reported outcomes and assisted in writing the sections on the measurement of obesity-related QOL (the OWLQOL) and the discussion and interpretation of the results. The authors declare that they have no competing interests.

References

1US Department of Health and Human Services (2010) Overweight and obesity: introduction. http://www.cdc.gov/obesity/data/trends.html (accessed 22 February 2010).Google Scholar
2US Department of Health and Human Services (2010) Overweight and obesity: health consequences. http://www.cdc.gov/obesity/causes/health.html (accessed 22 February 2010).Google Scholar
3US Department of Health and Human Services (2001) Office of Disease Prevention and Health Promotion. Healthy People 2010. Washington, DC. http://www.healthypeople.gov/2010/ (accessed 22 February 2010).Google Scholar
4Karlsson, J, Taft, C, Ryden, A, et al. (2007) Ten-year trends in health-related quality of life after surgical and conventional treatment for severe obesity: the SOS intervention study. Int J Obes (Lond) 31, 1248–1261.CrossRefGoogle ScholarPubMed
5Patrick, DL & Erickson, P (1993) Health Status and Health Policy: Quality of Life in Health Care Evaluation and Resource Allocation. New York: Oxford University Press.Google Scholar
6Fontaine, KR, Cheskin, LJ & Barofsky, I (1996) Health-related quality of life in obese persons seeking treatment. J Fam Pract 43, 265–270.Google Scholar
7Kolotkin, RL, Crosby, RD & Williams, GR (2002) Health-related quality of life varies among obese subgroups. Obes Res 10, 748–756.Google Scholar
8Wolf, AM & Colditz, GA (1998) Current estimates of the economic cost of obesity in the United States. Obes Res 6, 97–106.Google Scholar
9Finkelstein, EA, Dibonaventura, MD, Burgess, SM, et al. (2010) The costs of obesity in the workplace. J Occup Environ Med 52, 971–976.CrossRefGoogle ScholarPubMed
10Sullivan, PW, Ghushchyan, V & Ben-Joseph, RH (2008) The effect of obesity and cardiometabolic risk factors on expenditures and productivity in the United States. Obesity 16, 2155–2162.CrossRefGoogle ScholarPubMed
11Bungum, T, Satterwhite, M, Jackson, AW, et al. (2003) The relationship of body mass index, medical costs, and job absenteeism. Am J Health Behav 27, 456–462.Google Scholar
12Plaisted, CS, Lin, PH, Ard, JD, et al. (1999) The effects of dietary patterns on quality of life: a substudy of the Dietary Approaches to Stop Hypertension trial. J Am Diet Assoc 99, Suppl. 8, S84–S89.Google Scholar
13Vuillemin, A, Boini, S, Bertrais, S, et al. (2005) Leisure time physical activity and health-related quality of life. Prev Med 41, 562–569.Google Scholar
14Brown, DW, Balluz, LS, Heath, GW, et al. (2003) Associations between recommended levels of physical activity and health-related quality of life. Findings from the 2001 Behavioral Risk Factor Surveillance System (BRFSS) survey. Prev Med 37, 520–528.Google Scholar
15Brown, DW, Brown, DR, Heath, GW, et al. (2004) Associations between physical activity dose and health-related quality of life. Med Sci Sports Exerc 36, 890–896.CrossRefGoogle ScholarPubMed
16Beresford, SA, Thompson, B, Feng, Z, et al. (2001) Seattle 5 a Day worksite program to increase fruit and vegetable consumption. Prev Med 32, 230–238.Google Scholar
17US Census Bureau (1997) SIC matched to 1997 NAICS major groups (2-Digit). http://www.census.gov/epcd/naics/nsic2ndx.htm (accessed 4 July 2008).Google Scholar
18US Centers for Diseases Control and Prevention (2008) How is BMI calculated and interpreted?http://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html#Interpreted (accessed 28 July 2008).Google Scholar
19Thompson, FE & Byers, T (1994) Dietary assessment resource manual. J Nutr 124, Suppl. 11, 2245S–2317S.Google Scholar
20Lin, BH & Morrison, RM (2002) Higher fruit consumption linked with lower body mass index. Food Rev 25, 28–32.Google Scholar
21French, SA, Harnack, L & Jeffery, RW (2000) Fast food restaurant use among women in the Pound of Prevention study: dietary, behavioral and demographic correlates. Int J Obes Relat Metab Disord. 24, 1353–1359.Google Scholar
22Liebman, M, Pelican, S, Moore, SA, et al. (2003) Dietary intake, eating behavior, and physical activity-related determinants of high body mass index in rural communities in Wyoming, Montana, and Idaho. Int J Obes Relat Metab Disord. 27, 684–692.CrossRefGoogle ScholarPubMed
23Godin, G & Shephard, RJ (1997) Godin leisure-time exercise questionnaire. Med Sci Sports Exer 29, S36–S38.Google Scholar
24US Department of Agriculture (2010) How much physical activity is needed?http://www.choosemyplate.gov/foodgroups/physicalactivity_amount.html (accessed 7 November 2010).Google Scholar
25Godin, G & Shephard, RJ (1985) A simple method to assess exercise behavior in the community. Can J Appl Sport Sci 10, 141–146.Google Scholar
26Jacobs, DR Jr, Ainsworth, BE, Hartman, TJ, et al. (1993) A simultaneous evaluation of 10 commonly used physical activity questionnaires. Med Sci Sports Exerc 25, 81–91.Google Scholar
27Niero, M, Martin, M, Finger, T, et al. (2002) A new approach to multicultural item generation in the development of two obesity-specific measures: the Obesity and Weight Loss Quality of Life (OWLQOL) questionnaire and the Weight-Related Symptom Measure (WRSM). Clin Ther 24, 690–700.Google Scholar
28Patrick, DL, Bushnell, DM & Rothman, M (2004) Performance of two self-report measures for evaluating obesity and weight loss. Obes Res 12, 48–57.Google Scholar
29Maslow, AH (1943) A theory of human motivation. Psychol Rev 50, 370–396.CrossRefGoogle Scholar
30Patrick, D (2010) Personal communication via e-mail. [Whisnant, S, editor].Google Scholar
31Lerner, D, Amick, BC 3rd, Rogers, WH, et al. (2001) The work limitations questionnaire. Med Care 39, 72–85.Google Scholar
32Lerner, D, Amick, BC 3rd, Lee, JC, et al. (2003) Relationship of employee-reported work limitations to work productivity. Med Care 41, 649–659.Google Scholar
33Lerner, D, Rogers, WH & Chang, H (2005) Technical report: Scoring the Work Limitations Questionnaire (WLQ) scales and the WLQ index for estimating work productivity loss. Unpublished report.Google Scholar
34Kolotkin, RL, Head, S, Hamilton, M, et al. (1995) Assessing impact of weight on quality of life. Obes Res 3, 49–56.Google Scholar
35Katz, DA, McHorney, CA & Atkinson, RL (2000) Impact of obesity on health-related quality of life in patients with chronic illness. J Gen Intern Med 15, 789–796.Google Scholar
36Kolotkin, RL, Crosby, RD, Kosloski, KD, et al. (2001) Development of a brief measure to assess quality of life in obesity. Obes Res 9, 102–111.Google Scholar
37Larsson, U, Karlsson, J & Sullivan, M (2002) Impact of overweight and obesity on health-related quality of life – a Swedish population study. Int J Obes Relat Metab Disord 26, 417–424.CrossRefGoogle ScholarPubMed
38Brown, WJ, Mishra, G, Lee, C, et al. (2000) Leisure time physical activity in Australian women: relationship with well being and symptoms. Res Q Exerc Sport 71, 206–216.Google Scholar
39Muennig, P, Jia, H, Lee, R, et al. (2008) I think therefore I am: perceived ideal weight as a determinant of health. Am J Public Health 98, 501–506.Google Scholar
40Rand, CS & Resnick, JL (2000) The “good enough” body size as judged by people of varying age and weight. Obes Res 8, 309–316.Google Scholar
41Furnham, A, Badmin, N & Sneade, I (2002) Body image dissatisfaction: gender differences in eating attitudes, self-esteem, and reasons for exercise. J Psychol 136, 581–596.Google Scholar
42Demarest, J & Allen, R (2000) Body image: gender, ethnic, and age differences. J Soc Psychol 140, 465–472.Google Scholar
43Ford, ES, Moriarty, DG, Zack, MM, et al. (2001) Self-reported body mass index and health-related quality of life: findings from the Behavioral Risk Factor Surveillance System. Obes Res 9, 21–31.CrossRefGoogle ScholarPubMed
44Jia, H & Lubetkin, EI (2005) The impact of obesity on health-related quality-of-life in the general adult US population. J Public Health (Oxf) 27, 156–164.Google Scholar
Figure 0

Table 1Baseline demographic characteristics of employees at Move and Moderate in Balance worksites(Number of participants and percentages; mean values and standard deviations)

Figure 1

Table 2Mean Obesity and Weight-Loss Quality-of-Life (OWLQOL) scores by baseline reported dietary behaviours, physical activity and BMI(Mean values and 95 % confidence intervals)

Figure 2

Table 3Mean Work Limitations Questionnaire (WLQ) Productivity Loss Score by baseline reported dietary behaviours, physical activity and BMI(Mean values and 95 % confidence intervals)