Occupational physical activity and body mass index (BMI) among Canadian adults: does physical activity at work help to explain the socio-economic patterning of body weight?

Barberio, Amanda ; McLaren, Lindsay

Canada is one of the many industrialized nations that have recorded an increasing prevalence of overweight and obesity. The prevalence of adult obesity (i.e., BMI [greater than or equal to] 30 kg/[m.sup.2]) in Canada has risen from 13.8% in 1978/79 to 23.1% in 2004. (1) Obesity has been associated with a broad spectrum of physical, psychological, social and economic consequences. (2-4)

Body mass index (BMI) and obesity are associated with socioeconomic position (SEP), (5) which refers to one's social and economic position in society. (6) The nature of this association varies by sex and by country, being predominantly inverse for women in the developed world (i.e., higher SEP, lower BMI) and less consistent for men in these countries. (5) The association appears to further vary by indicator or aspect of SEP; (1) for example, the education-BMI association is predominantly inverse (particularly for women), while some studies have shown a positive income-BMI association for men but not women. (5,7)

Some Canadian studies have examined possible mediators of the associations between SEP and BMI/obesity. For example, both Kuhle and Veugelers (7) and Ward and colleagues (8) reported that leisure time physical activity (LTPA) and fruit and vegetable consumption helped to explain the SEP-obesity association among Canadian adults. Research suggests that LTPA does not fully capture one's total energy expenditure (9,10) and thus, the amount of physical activity accumulated in other life domains should also be considered.

Although recent technological advancements and automation have reduced the amount of energy expenditure for many occupations, (11) it is nonetheless important to consider occupational physical activity (OPA) because individuals spend a large portion of their waking hours at work. It is plausible that, for example, the positive income-BMI association in men referred to above (5,7) reflects, in part, that higher-income men are less active at work.

We located ten population-based studies that investigated OPA in relation to BMI. (12-21) Findings were varied; however, among the six studies that used measured (versus self-reported) height and weight, four reported an inverse OPA-BMI association (12-15) (with the exception of a positive association for women in the study by Barengo et al. (13)) and two found no significant association. (16,17) Importantly, in most (i.e., nine of ten) studies, OPA was based on employee self-report, which is prone to reporting biases, and is thus a limitation of existing research.

To our knowledge, there are no Canadian population-based studies that examine OPA in relation to BMI. Thus, it is not currently known in Canada whether, or the extent to which: a) OPA is associated with BMI; and b) OPA mediates the association between SEP (income, education) and BMI, among Canadian working adults. Our aim is to fill these knowledge gaps.

METHODS

We analyzed microdata from the 2008 Canadian Community Health Survey (CCHS), accessed at the Prairie Regional Research Data Centre. Full details about the 2008 CCHS are available online at www.statcan.gc.ca, but briefly, the CCHS is a nationally representative, cross-sectional survey that collects health information for the Canadian population age 12 and over. The 2008 CCHS covers approximately 98% of the population in the provinces, and approximately 71-97% of the population living in the Territories. The vast majority (99%) of the sample was drawn using a multistage stratified cluster design and the overall survey response rate was 77.6%.

We focused on respondents 25-64 years of age who indicated working at a job or business in the previous week. Use of the 2008 CCHS is ideal for this research for two reasons: first, measured height and weight (used to calculate BMI) was collected by a trained interviewer using validated instruments for a subsample of 4,688 respondents (of whom 2,841 were age 25-64). The target population for this subsample was the same as that of the full CCHS survey with the exception of the Territories, according to the 2008 CCHS User Guide. Second, detailed occupational data were available for 2,550 respondents who reported working at a job or business (of whom 2,135 were age 25-64), and this information was used to assign each respondent an occupational code (of which there are 520) based on the National Occupational Classification for Statistics 2006 (NOC-S) system.

We used a novel approach for characterizing OPA based on information from the NOC Career Handbook (http://www5.hrsdc.gc.ca/NOC/English/CH/2001/Welcome.aspx). For each occupation indicated by a NOC 4-digit unit group, the Handbook provides descriptive information about the job derived through systematic, field-based research by the former Occupational and Career Information Branch of Human Resources Development Canada. We used information from three items describing each occupation's physical exertion: body position, limb coordination, and strength. This classification scheme has considerable face validity; for example, a firefighter has the highest scores in all three dimensions whereas a university professor has the lowest scores in all three dimensions. We generated an overall score for OPA for each CCHS respondent by summing the scores from the three dimensions of physical activity and dividing into three approximately equal-sized groups (high, medium, low). To further explore the validity of our OPA measure, we located a self-reported work physical exertion variable that was present as optional content for the main sample of the 2008 CCHS ("WST_408"). Among adults age 25-64 in the main 2008 CCHS sample for whom both occupational physical activity variables were available, we computed the correlation coefficient between the two variables.

Other variables were: highest level of education achieved by the respondent and the respondent's income adequacy group * (these were used as indicators of SEP); socio-demographic covariates (age; hours worked during past week; birth place); behaviours (smoking status; alcohol consumption; LTPA; fruit/vegetable consumption) and heath status variables (self-rated health; chronic disease diagnosis). These covariates have elsewhere shown associations with obesity. (22-25)

Analyses were performed using STATA software (version 11), and appropriate sampling weights were applied as directed by Statistics Canada. Ordinary least squares (OLS) regression was used. To examine the association between OPA and BMI (first research objective), we ran an unadjusted model in which BMI was regressed on each variable individually and a fully adjusted model in which BMI was simultaneously regressed on OPA, income adequacy, education, and all covariates. To test mediation (our second research objective), we followed the procedure outlined by Baron and Kenny (1986), see Figure 1. (26) In our case, the "independent variable" is SEP (i.e., income adequacy or education), the "mediator" is OPA, and the "outcome variable" is BMI. According to Baron and Kenny, a variable may function as a mediator when paths "a", "b", and "c" exist (i.e., statistically significant unadjusted association). Finally, when the mediator (OPA) is adjusted for, path "c" (SEP-BMI association) must be reduced, ideally to non-significance.

RESULTS

Our sample included n=1,972 (n=1,036 males and n=936 females) respondents age 25-64 who had complete data on all variables. Please refer to Table 1 for the descriptive statistics for our sample. A total of 163 (7.6%) eligible respondents were excluded due to missing data on one or more variables used in this study. Excluded versus included respondents were significantly (at p<0.05) heavier and more likely to be female, but did not differ on any other study variables. We observed a statistically significant positive correlation between our NOC-based OPA variable and the "WST_408" variable among the main 2008 CCHS sample (men: Pearson's r=0.52, p<0.001, n=1,436; women: Pearson's r=0.50, p<0.01, n=1,365), thus supporting the validity of our constructed OPA variable.

[FIGURE 1 OMITTED]

Table 2 presents the results of the OLS regression models (unadjusted and adjusted) for both males and females. OPA was not associated with BMI in the unadjusted models for either males or females (Model A). In the fully adjusted model for females (Model B), a marginal effect was observed whereby those with a medium level of OPA were marginally lighter than those with a low level of OPA (coef.= -1.09, 95% CI -2.24 to 0.07, p=0.07). For males in the fully adjusted model, the effect of OPA remained non-significant.

In terms of testing mediation (Figure 1), we confirmed that conditions "a" ** (SEP-OPA) and "c" (SEP-BMI) (see Table 1) were fulfilled for both men (income adequacy) and women (education); however, condition "b" (OPA-BMI) was not satisfied in any case, and thus, we did not proceed with testing mediation.

DISCUSSION

Aside from a small effect observed in women, BMI was not found to be clearly or strongly related to OPA. As such, our findings differ from four of six reasonably comparable studies identified in the literature (12-15) that reported a significant inverse association.

Both conceptual and methodological explanations must be considered to account for our non-significant findings. Perhaps there really is no clear association between OPA and BMI in Canadian adults working at a job or business, and our methods accurately captured this. Alternatively, perhaps a weak association between OPA and BMI exists, but it is overwhelmed by other aspects of lifestyle. For example, it is possible that the greater energy expenditure of individuals working at a highly active occupation may be counterbalanced by a higher caloric consumption. Giving support to this notion, a 1998 study found that individuals whose occupations required a greater amount of OPA reported a higher caloric intake than their counterparts whose occupations required a lower amount of OPA. (27) To the extent that an individual's caloric intake exceeds their expenditure, caloric intake may be a more important predictor of BMI/obesity than OPA, even among individuals who are highly active at work.

The possibility that certain methodological flaws masked significant associations between OPA and BMI cannot be ruled out. One example is that an individual's BMI (the method we used to capture one's weight status) does not provide an accurate indication of their fat mass versus their fat-free mass. (28) This could result in a misclassification of BMI for a small number of individuals (e.g., those who are muscular because of their OPA, and who may thus be misclassified as overweight/obese) which might dilute our association of interest. Furthermore, it is possible that our measure of OPA failed to capture physical exertion at work with the same degree of accuracy as, for example, an electronic device to measure calorie expenditure (although this approach could be expensive and impractical for large samples). However, the high and statistically significant correlation between our OPA measure and a separate self-report variable on work physical exertion, coupled with the ethnographic approach used to gather the NOC data, suggests reasonable validity of our variable.

We explored two other possible reasons for not detecting a clear association between OPA and BMI, in a post hoc manner. First, it is possible that, by combining three dimensions of OPA (body position, limb coordination, and strength) into an overall index, we may have missed an important effect of any one dimension on its own. We tested this possibility by regressing BMI on each OPA dimension separately, adjusting for covariates. We observed two marginally significant associations in adjusted models: for men, a negative association between the highest category of strength and BMI (coef. = -0.87, 95% CI -0.18 to 0.03, p=0.058), and for women, a negative association between the third highest (of four) category of body position and BMI (coef. = -1.03, 95% CI -2.18 to 0.13, p=0.08). These marginal findings suggest that we did not miss important effects by creating an overall score for OPA, though they raise interesting ideas about gender differences in dimensions of importance (strength, e.g., heavy lifting, for men; body position, e.g., combination of sitting, standing, walking, for women) to be pursued in future research.

Second, it is possible that our sample size, though reasonably large, rendered our analysis under-powered to detect an OPA-BMI association if one exists. A post hoc power calculation using alpha = 0.05, current sample size (n=936 for women; n=1,036 for men), and mean (SD) values of BMI from our own data indicated that we had limited (<40%) power to detect a mean difference in BMI of 0.5 kg/[m.sup.2] between low and high OPA groups. To have 80% (acceptable) power to detect such a difference, a sample size of 1,570 in each group (high and low OPA) would be needed. Thus, we reran analyses using the main sample of the 2008 CCHS, for which a much larger sample was available (n=11,628 for women; n=12,057 for men), but for which BMI was computed from self-reported rather than measured height and weight. Towards improving the BMI variable, we applied the correction factor outlined by Connor-Gorber et al. (29) *** For women, main sample results were generally consistent with the subsample results: high OPA was positively associated with BMI (coefficient = 0.58, 95% CI 0.13 to 1.02, p<0.01), medium OPA was not significantly associated with BMI but the coefficient was negative in direction (coefficient = -0.12, 95% CI -0.46 to 0.23, NS), in adjusted models. For men, in contrast to the subsample findings, the main sample findings indicated a positive association between high OPA and BMI (coefficient = 0.33, 95% CI 0.00 to 0.68, p<0.05) and between medium OPA and BMI (coefficient = 0.36, 95% CI 0.04 to 0.69, p<0.05), in adjusted analyses. Although we cannot determine whether different findings between the two samples reflect sample size differences or the different outcome variables (measured vs. self-report BMI), we might sum up our results as indicating that the OPA-BMI association for men is unclear, while for women the association appears to be nonlinear (inverse for medium OPA; positive for high OPA). Barengo et al. (13) also detected a positive relationship between OPA and BMI for women. One possible explanation is that highly active jobs attract women who are larger and/or more muscular to begin with, perhaps because the larger physique is suited to the job (e.g., job requires strength) and/or because the job environment is more accepting of larger women.

Naturally, our research has some limitations; we focus here on two. First, the duration of exposure to OPA was unknown because information on the length of time in one's job was not available in the data source used. Second, due to its cross-sectional design, this study was unable to discern causality or temporality of associations.

Several strengths are present in this study. First, we built on existing population-based studies of OPA by taking advantage of a relatively objective form of OPA measurement, thereby reducing the likelihood of recall and social desirability biases that can plague self-report data. Second, our primary data source (from the 2008 CCHS subsample) contained measured height and weight, thereby eliminating the potential of reporting biases that occur with self-reported data. (30) Finally, our research was based on a reasonably large, nationally representative sample, and the number of participants excluded due to missing data was small and our analysis indicated only minimal missing data bias.

Future research attempting to explain the association between SEP and BMI could explore other plausible variables (e.g., psychosocial stress, which has been linked with lower SEP (31,32) and with accumulation of visceral adipose tissue (33)), adopt a prospective design, and if possible clarify duration of exposure to plausible explanatory variables.

Acknowledgements: This work stems from Amanda Barberio's honours thesis project, submitted in partial fulfillment of the degree of Bachelor of Health Sciences Honours at the University of Calgary and successfully defended in April 2010. Lindsay McLaren is supported by a Population Health Investigator Award from Alberta Innovates--Health Solutions (formerly the Alberta Heritage Foundation for Medical Research). We thank Dr. Jenny Godley for helpful comments on an earlier version of this manuscript.

Conflict of Interest: None to declare.

Received: August 20, 2010

Accepted: November 22, 2010

REFERENCES

(1.) Tjepkema M. Adult obesity. Health Rep 2006;17:9-25.

(2.) Bray GA. Medical consequences of obesity. J Clin Endocrinol Metab 2004;89:2583-89.

(3.) Brownell KD, Puhl R, Schwartz MB, Rudd L. Weight Bias: Nature, Consequences, and Remedies. New York, NY: The Guilford Press, 2005.

(4.) Katzmarzyk PT, Janssen I. The economic costs associated with physical inactivity and obesity in Canada: An update. Appl Physiol Nutr Metab 2004;29:90-115

(5.) McLaren L. Socioeconomic status and obesity. Epidemiol Rev 2007;29:29-48.

(6.) Shaw M, Galobardes B, Lawlor DA, Lynch J, Wheeler B, Smith GD. The Handbook of Inequality and Socioeconomic Position: Concepts and Measures. Bristol, UK: The Policy Press, 2008.

(7.) Kuhle S, Veugelers PJ. Why does the social gradient in health not apply to overweight? Health Rep 2008;19:7-15.

(8.) Ward H, Tarasuk V, Mendelson R. Socioeconomic patterns of obesity in Canada: Modeling the role of health behaviour. Appl Physiol Nutr Metab 2007;32:206-16.

(9.) Macera CA, Pratt M. Public health surveillance of physical activity. Res Q Exercise Sport 2000;71: 97-103.

(10.) Craig CL, Russell SJ, Cameron CC, Bauman A. Twenty-year trends in physical activity among Canadian adults. Can J Public Health 2004;95(1):59-63.

(11.) Pratt M, Macera CA, Sallis JF, O'Donnell M, Frank L. Economic interventions to promote physical activity: Application of the SLOTH model. Am J Prev Med 2004;27:136-45.

(12.) King GA, Fitzhugh EC, Bassett Jr. DR, McLaughlin JE, Strath SJ, Swartz AM, et al. Relationship of leisure-time physical activity and occupational activity to the prevalence of obesity. Int J Obesity 2001;25:606-12.

(13.) Barengo NC, Kastarinen M, Lakka T, Nissinen A, Tuomilehto J. Different forms of physical activity and cardiovascular risk factors among 24-64-year-old men and women in Finland. Eur J Cardiovasc Prev Rehabil 2006;13:51-59.

(14.) Larsson I, Lissner L, Naslund I, Lindroos AK. Leisure and occupational physical activity in relation to body mass index in men and women. Scand J Nutr 2004;48:165-72.

(15.) Monda KL, Adair LS, Zhai F, Popkin BM. Longitudinal relationships between occupational and domestic physical activity patterns and body weight in China. Eur J Clin Nutr 2007;62:1318-25.

(16.) Sofi F, Capalbo A, Marcucci R, Gori AM, Fedi S, Macchi C, et al. Leisure time but not occupational physical activity significantly affects cardiovascular risk factors in an adult population. Eur J Clin Invest 2007;37:947-53.

(17.) Ball K, Owen N, Salmon J, Bauman A, Gore CJ. Associations of physical activity with body weight and fat in men and women. Int J Obes 2001;25:914-19.

(18.) Mummery WK, Schofield GM, Steele R, Eakin EG, Brown WJ. Occupational sitting time and overweight and obesity in Australian workers. Am J Prev Med 2005;29(2):91-97.

(19.) Gutierrez-Fisac JL, Guallar-Castillon P, Diez-Ganan L, Garcia EL, Banegas JR, Artalejo FR. Work-related physical activity is not associated with body mass index and obesity. Obes Res 2002;10(4):270-76.

(20.) Kaleta D, Makowiec-Dabrowska T, Jegier A. Occupational and leisure-time energy expenditure and body mass index. Int J Occup Med Environ Health 2007;20(1):9-16.

(21.) Abu-Omar K, Rutten A. Relation of leisure time, occupational, domestic, and commuting physical activity to health indicators in Europe. Prev Med 2008;47:319-23.

(22.) Raine KD. Overweight and obesity in Canada: A population health perspective. Ottawa, ON: Canadian Institute for Health Information, 2004.

(23.) Healton CG, Vallone D, McCausland KL, Xiao H, Green MP. Smoking, obesity and their co-occurrence in the United States: Cross-sectional analysis. Br Med J 2006;33:25-26.

(24.) Suter PM, Tremblay A. Is alcohol consumption a risk factor for weight gain and obesity? Crit Rev Clin Lab Sci 2005;42:197-227.

(25.) Shields M. Long working hours and health. Health Rep 1999;11:33-38.

(26.) Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. J Pers Soc Psychol 1986;51:1173-82.

(27.) Paeratakul S, Popkin BM, Keyou G, Adair LS, Stevens J. Changes in diet and physical activity affect the body mass index of Chinese adults. Int J Obes Rel Metab Disord 1998;22:424-31.

(28.) Prentice AM, Jebb SA. Beyond body mass index. Obes Rev 2001;2:141-47.

(29.) Connor Gorber S, Shields M, Tremblay MS, McDowell I. The feasibility of establishing correction factors to adjust self-reported estimates of obesity. Health Rep 2008;19:71-82.

(30.) Gorber CS, Tremblay M, Moher D, Gorber B. A comparison of direct vs. self report measures for assessing height, weight and body mass index: A systematic review. Obesity Rev 2007;8:307-26.

(31.) Marmot M. Social determinants of health inequalities. Lancet 2005;365:1099-104

(32.) Raphael D. Social Determinants of Health: Canadian Perspectives, 2nd ed. Toronto, ON: Canadian Scholars' Press Inc., 2009.

(33.) Bjorntorp J. Do stress reactions cause abdominal obesity and comorbidities? Obes Rev 2001;2:73-86.

* Income adequacy is a derived Statistics Canada variable that takes into account annual household income and the weighted size of the household.

** Statistically significant association between SEP and OPA for men (e.g., highest income adequacy: coefficient= -1.17, 95% CI 1.83 to -0.52, p=0.00; highest education [post-sec grad] coef.= -1.77, 95% CI -2.32 to -1.23, p=0.00) and women (e.g., highest income adequacy: coefficient = -2.06, 95% CI -2.66 to -1.46, p=0.00; highest education [post-sec grad] coef.= -0.89, 95% CI -1.65 to -0.14, p=0.02).

*** The findings did not differ substantively between the corrected and the uncorrected data.

Amanda Barberio, BHSc (hons), Lindsay McLaren, PhD

Department of Community Health Sciences, University of Calgary, Calgary, AB

Correspondence: Dr. Lindsay McLaren, Department of Community Health Sciences, University of Calgary, 3rd Floor TRW, 3280 Hospital Dr. NW, Calgary, AB T2N 4Z6, Tel: 403-210-9424, Fax: 403-270-7307, E-mail: lmclaren@ucalgary.ca

Table 1. Weighted Descriptive Statistics for Study Sample, Stratified
by Sex

Variable                                           Males (n=1,036)
                                                  Mean (SD) or n (%)

Body mass index (BMI)                             27.7 (4.5)
  (kg/[m.sup.2])
Occupational physical  Low                         273 (26.3%)
  activity (OPA)       Medium                      425 (41.0%)
                       High                        338 (32.7%)
Income adequacy        Lowest                      110 (10.6%)
                       Low-middle                  283 (27.3%)
                       High-middle                 216 (20.8%)
                       Highest                     427 (41.2%)
Education              Less than secondary         115 (11.1%)
                       Secondary school graduate   152 (14.7%)
                       Some post-secondary          62 (6.0%)
                       Post-secondary graduate     707 (68.2%)
Alcohol consumption    Regular                     804 (77.6%)
                       Occasional                  104 (10.0%)
                       Never                       128 (12.3%)
Smoking status         Current                     298 (28.7%)
                       Former                      421 (40.7%)
                       Never                       317 (30.6%)
Fruit and vegetable    Eats fruit and vegetables   661 (63.8%)
  consumption            <5 times/day
                       Eats fruit and vegetables   375 (36.2%)
                         [greater than or equal
                         to] 5 times/day
Leisure-time physical  Inactive                    546 (52.7%)
  activity (LTPA)      Moderately Active           267 (25.8%)
                       Active                      223 (21.5%)
Hours Worked (past     0 to 35 hrs/week            126 (12.2%)
  week)                36 to 40 hrs/week           431 (41.6%)
                       41+ hrs/week                479 (46.2%)
Birth place            Born outside of Canada      274 (26.5%)
                       Born in Canada              762 (73.5%)
Self-rated health      Poor or fair                 59 (5.7%)
                       Good                        318 (30.7%)
                       Very good or excellent      659 (63.6%)
Chronic disease        Diagnosed with one or       220 (21.2%)
  diagnosis              more chronic diseases
                       Not diagnosed with any      816 (78.8%)
                         chronic diseases

Variable                                           Females (n=936)
                                                  Mean (SD) or n (%)

Body mass index (BMI)                             26.8 (6.4)
  (kg/[m.sup.2])
Occupational physical  Low                         415 (44.3%)
  activity (OPA)       Medium                      363 (38.8%)
                       High                        158 (16.9%)
Income adequacy        Lowest                       95 (10.2%)
                       Low-middle                  277 (29.6%)
                       High-middle                 236 (25.2%)
                       Highest                     328 (35.1%)
Education              Less than secondary          48 (5.1%)
                       Secondary school graduate   161 (17.2%)
                       Some post-secondary          63 (6.7%)
                       Post-secondary graduate     664 (71.0%)
Alcohol consumption    Regular                     536 (57.3%)
                       Occasional                  236 (25.2%)
                       Never                       164 (17.5%)
Smoking status         Current                     186 (19.9%)
                       Former                      367 (39.2%)
                       Never                       383 (40.9%)
Fruit and vegetable    Eats fruit and vegetables   485 (51.9%)
  consumption            <5 times/day
                       Eats fruit and vegetables   451 (48.1%)
                         [greater than or equal
                         to] 5 times/day
Leisure-time physical  Inactive                    534 (57.0%)
  activity (LTPA)      Moderately Active           225 (24.1%)
                       Active                      177 (18.9%)
Hours Worked (past     0 to 35 hrs/week            387 (41.3%)
  week)                36 to 40 hrs/week           383 (41.0%)
                       41+ hrs/week                166 (17.7%)
Birth place            Born outside of Canada      249 (26.6%)
                       Born in Canada              687 (73.4%)
Self-rated health      Poor or fair                 78 (8.3%)
                       Good                        233 (24.9%)
                       Very good or excellent      625 (66.8%)
Chronic disease        Diagnosed with one or       239 (25.6%)
  diagnosis              more chronic diseases
                       Not diagnosed with any      697 (74.4%)
                         chronic diseases

Note: Chronic diseases considered were: arthritis, asthma,
bronchitis, cancer, diabetes and heart disease.

Table 2. Results of Ordinary Least Squares Regression Analysis
for Males (n=1,036) and Females (n=936), with BMI (continuous
variable), Regressed on Occupational Physical Activity (OPA),
Socio-economic Variables (income adequacy and education) and
Covariates

Predictor Variable                            Model A--Males
                                               (unadjusted)
                                           Coefficient (95% CI)

Occupational physical activity
    (OPA) (Reference: Low)
  Medium                             -0.05 (-1.02 to 0.91)
  High                               -0.19 (-1.17 to 0.80)
Income adequacy (Reference:
    Lowest)
  Middle-low                          1.03 (-0.53 to 2.59)
  Middle-high                         1.75 (0.14 to 3.36) *
  Highest                             1.98 (0.39 to 3.57) **
Education (Reference: Less than
    secondary)
  Secondary school graduate           0.72 (-0.56 to 2.01)
  Some post-secondary                 2.29 (0.60 to 3.99) **
  Post-secondary graduate             0.59 (-0.33 to 1.52)
Alcohol consumption (Reference:
    Regular)
  Occasional                          0.18 (-1.17 to 1.54)
  Never                              -0.29 (-1.33 to 0.76)
Smoking status (Reference:
    Current)
  Former                              1.20 (0.25 to 2.16) **
  Never                               0.51 (-0.54 to 1.57)
Fruit and vegetable consumption
  (Reference: Eats fruit and
    vegetables <5 times/day)
  Eats fruit and vegetables
    [greater than or equal to]
    5 times/day                      -0.34 (-1.11 to 0.43)
Leisure time physical activity
    (LTPA) (Reference: Inactive)
  Moderately active                  -0.56 (-1.38 to 0.27)
  Active                             -0.95 (-2.05 to 0.16) ([dagger])
Hours worked/week (Reference:
    0 to 35 hrs/week )
  36 to 40 hrs/week                   0.31 (-0.89 to 1.50)
  41+ hrs/week                        0.92 (-0.32 to 2.16)
Birth place (Reference: Born
    outside of Canada)
  Born in Canada                      2.59 (1.81 to 3.38) **
Self-rated health (Reference:
    Poor or fair)
  Good                               -1.46 (-3.88 to 0.96)
  Very good or excellent             -2.12 (-4.45 to 0.21) ([dagger])
Chronic disease diagnosis
  (Reference: Not diagnosed
  with any chronic diseases)
Diagnosed with one or more
  chronic diseases                    1.13 (0.22 to 2.05) *

Predictor Variable                             Model B--Males
                                              (fully adjusted)
                                            Coefficient (95% CI)

Occupational physical activity
    (OPA) (Reference: Low)
  Medium                              0.23 (-0.66 to 1.13)
  High                               -0.01 (-1.01 to 0.98)
Income adequacy (Reference:
    Lowest)
  Middle-low                          0.74 (-0.66 to 2.15)
  Middle-high                         1.00 (-0.47 to 2.48)
  Highest                             1.13 (-0.32 to 2.57)
Education (Reference: Less than
    secondary)
  Secondary school graduate           0.62 (-0.58 to 1.81)
  Some post-secondary                 1.74 (0.13 to 3.34) *
  Post-secondary graduate             0.78 (-0.18 to 1.74)
Alcohol consumption (Reference:
    Regular)
  Occasional                          0.60 (-0.66 to 1.86)
  Never                               0.23 (-0.75 to 1.21)
Smoking status (Reference:
    Current)
  Former                              1.31 (0.52 to 2.09) **
  Never                               1.13 (0.22 to 2.04) *
Fruit and vegetable consumption
  (Reference: Eats fruit and
    vegetables <5 times/day)
  Eats fruit and vegetables
    [greater than or equal to]
    5 times/day                      -0.20 (-0.90 to 0.51)
Leisure time physical activity
    (LTPA) (Reference: Inactive)
  Moderately active                  -0.68 (-1.45 to 0.09) ([dagger])
  Active                             -0.93 (-1.93 to 0.08) ([dagger])
Hours worked/week (Reference:
    0 to 35 hrs/week )
  36 to 40 hrs/week                   0.57 (-0.45 to 1.59)
  41+ hrs/week                        0.93 (-0.10 to 1.96) ([dagger])
Birth place (Reference: Born
    outside of Canada)
  Born in Canada                      2.49 (1.66 to 3.32) **
Self-rated health (Reference:
    Poor or fair)
  Good                               -1.51 (-3.86 to 0.85)
  Very good or excellent             -2.11 (-4.51 to 0.29) ([dagger])
Chronic disease diagnosis
  (Reference: Not diagnosed
  with any chronic diseases)
Diagnosed with one or more
  chronic diseases                    0.90 (0.07 to 1.72) *

Predictor Variable                       Model A--Females
                                           (unadjusted)
                                       Coefficient (95% CI)

Occupational physical activity
    (OPA) (Reference: Low)
  Medium                             -0.83 (-2.00 to 0.34)
  High                                1.08 (-0.62 to 2.79)
Income adequacy (Reference:
    Lowest)
  Middle-low                         -0.33 (-1.98 to 1.32)
  Middle-high                         0.51 (-1.41 to 2.43)
  Highest                            -1.19 (-2.86 to 0.48)
Education (Reference: Less than
    secondary)
  Secondary school graduate          -0.40 (-2.45 to 1.65)
  Some post-secondary                -0.74 (-2.85 to 1.38)
  Post-secondary graduate            -1.76 (-3.17 to 0.34) *
Alcohol consumption (Reference:
    Regular)
  Occasional                         -0.03 (-1.35 to 1.30)
  Never                              -0.55 (-1.82 to 0.73)
Smoking status (Reference:
    Current)
  Former                              0.98 (-0.38 to 2.34)
  Never                              -0.62 (-1.87 to 0.64)
Fruit and vegetable consumption
  (Reference: Eats fruit and
    vegetables <5 times/day)
  Eats fruit and vegetables
    [greater than or equal to]
    5 times/day                       0.17 (-0.95 to 1.29)
Leisure time physical activity
    (LTPA) (Reference: Inactive)
  Moderately active                  -0.82 (-2.06 to 0.41)
  Active                             -1.19 (-2.72 to 0.35)
Hours worked/week (Reference:
    0 to 35 hrs/week )
  36 to 40 hrs/week                  -0.08 (-1.31 to 1.16)
  41+ hrs/week                        0.03 (-1.47 to 1.54)
Birth place (Reference: Born
    outside of Canada)
  Born in Canada                      2.28 (1.20 to 3.36) **
Self-rated health (Reference:
    Poor or fair)
  Good                                0.61 (-2.02 to 3.24)
  Very good or excellent             -1.27 (-3.67 to 1.13)
Chronic disease diagnosis
  (Reference: Not diagnosed
  with any chronic diseases)
Diagnosed with one or more
  chronic diseases                    2.40 (0.86 to 3.93) **

Predictor Variable                           Model B--Females
                                             (fully adjusted)
                                           Coefficient (95% CI)

Occupational physical activity
    (OPA) (Reference: Low)
  Medium                             -1.09 (-2.24 to 0.07) ([dagger])
  High                                0.72 (-0.98 to 2.41)
Income adequacy (Reference:
    Lowest)
  Middle-low                         -0.07 (-1.79 to 1.65)
  Middle-high                         0.37 (-1.65 to 2.38)
  Highest                            -1.09 (-3.00 to 0.83)
Education (Reference: Less than
    secondary)
  Secondary school graduate          -0.67 (-2.79 to 1.45)
  Some post-secondary                -0.93 (-3.15 to 1.29)
  Post-secondary graduate            -1.45 (-3.25 to 0.35)
Alcohol consumption (Reference:
    Regular)
  Occasional                         -0.01 (-1.22 to 1.21)
  Never                               0.22 (-1.11 to 1.56)
Smoking status (Reference:
    Current)
  Former                              1.65 (0.27 to 3.03) *
  Never                               0.87 (-0.57 to 2.32)
Fruit and vegetable consumption
  (Reference: Eats fruit and
    vegetables <5 times/day)
  Eats fruit and vegetables
    [greater than or equal to]
    5 times/day                       0.92 (-0.17 to 2.01) ([dagger])
Leisure time physical activity
    (LTPA) (Reference: Inactive)
  Moderately active                  -0.87 (-2.05 to 0.32)
  Active                             -1.14 (-2.67 to 0.39)
Hours worked/week (Reference:
    0 to 35 hrs/week )
  36 to 40 hrs/week                   0.48 (-0.68 to 1.63)
  41+ hrs/week                        0.73 (-0.68 to 2.14)
Birth place (Reference: Born
    outside of Canada)
  Born in Canada                      2.76 (1.64 to 3.87) **
Self-rated health (Reference:
    Poor or fair)
  Good                                1.06 (-1.38 to 3.49)
  Very good or excellent             -0.62 (-3.01 to 1.78)
Chronic disease diagnosis
  (Reference: Not diagnosed
  with any chronic diseases)
Diagnosed with one or more
  chronic diseases                    2.15 (0.49 to 3.81) **

** p<0.01; * p<0.05; ([dagger]) p<0.10