- Email: [email protected]
Young Children and Parental Physical Activity Levels
Young Children and Parental Physical Activity Levels Findings from the Canadian Health Measures Survey Kristi B. Adamo, PhD, Kellie A. Langlois, MSc, Kendra E. Brett, MSc, Rachel C. Colley, PhD Background: Physical inactivity is a global public health concern. The relationship between dependent children in the home and parental physical activity has not been quantifıed using objective measures, nor has the relative association of the physical activity levels of mothers and fathers been examined.
Purpose: To investigate the association of children of different ages in the home on two measures of parental physical activity: daily moderate-to-vigorous physical activity (MVPA) and likelihood of meeting the guideline of 150 minutes of MVPA per week accumulated in 10-minute bouts. Methods: Data were from the 2007–2009 Canadian Health Measures Survey (n⫽2315), and analyses were conducted between February and December 2011. MVPA was measured directly using accelerometry. Linear (minutes of MVPA) and logistic (meeting physical activity guidelines) regression models were performed to determine if the presence, number of children, or the age of the youngest child at home was associated with parental physical activity. All models were adjusted for parental age, marital status, household income, employment, and BMI. Results: Mothers whose youngest child was aged ⬍6 years and fathers whose youngest was aged 6 –11 years engaged in fewer minutes of daily MVPA than those without dependent children. Linear regression results identifıed that in comparison to those without children, women whose youngest child in the home was aged ⬍6 years participated in 7.7 minutes less activity per day (p⫽0.007) whereas men engaged in 5.7 fewer minutes per day, or 54 and 40 minutes per week less, respectively. Similarly, logistic regression analyses indicated that both women and men were less likely to meet guidelines if their youngest child in the home was aged ⬍6 years (OR⫽0.31, 95% CI⫽0.11, 0.87; OR⫽0.34, 95% CI⫽0.13, 0.93). Conclusions: The physical activity level of parents with young children present in the home was lower than that of those without children. Given the many physiologic, psychological, and social benefıts of healthy active living, research efforts should continue to focus on strategies to encourage parents with young children to establish or re-engage in a physically active lifestyle, not only for their own health but to model healthy behavior for the next generation. (Am J Prev Med 2012;43(2):168 –175) © 2012 American Journal of Preventive Medicine
Introduction
From the Healthy Active Living and Obesity Research Group, Children’s Hospital of Eastern Ontario Research Institute (Adamo, Brett, Colley); the Faculty of Medicine, Department of Pediatrics (Adamo, Colley); the Faculty of Health Sciences, School of Human Kinetics (Adamo, Brett, Colley), University of Ottawa; and the Health Analysis Division (Langlois, Colley), Statistics Canada, Ottawa, Ontario, Canada Address correspondence to: Kristi B. Adamo, PhD, Children’s Hospital of Eastern Ontario Research Institute, 401 Smyth Road, Ottawa, ON K1H 8L1, Canada. E-mail: [email protected]. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2012.02.032
168 Am J Prev Med 2012;43(2):168 –175
S
ubstantive research has profıled the health-related risks of physical inactivity and a sedentary lifestyle.1 Physical activity is associated with reductions in both all-cause mortality and risk for multiple chronic diseases.2 Despite public awareness of the benefıts of an active lifestyle, physical activity participation continues to be low in adults in Canada3 and worldwide.4 Many factors are associated with low levels of physical activity in the population (such as age, income, gender, BMI), and having dependent children in the home has been sug-
© 2012 American Journal of Preventive Medicine • Published by Elsevier Inc.
Adamo et al / Am J Prev Med 2012;43(2):168 –175 5
gested as a contributor to lowered levels in adults, particularly women.6,7 Having children aged ⬍5 years, in particular, is associated negatively with self-reported exercise.8 Clarifying the relationship between parenthood and physical activity is diffıcult given that young children affect parental levels of physical activity differently than older children because of changing degrees of dependency on parental time and attention. A recent systematic review noted that parents, specifıcally mothers, are less active than those without children; however, the literature is mixed in regard to the impact of the age and number of children at home as predictors of parental physical activity.5 In addition, research concerning parental physical activity habits is limited by the use of self-reported data, with only one study using pedometers as a direct measure of physical activity.9 There is also limited evidence on physical activity levels in fatherhood, with the majority of the research targeting mothers.10 –16 Further, previous research has focused primarily on qualifying the barriers to physical activity,5,11,17,18 rather than quantifying the effects of number and age of dependent children at home on levels of parental physical activity. The need to examine the impact of children on physical activity patterns of both parents using objective measures has been called for.5 The present study examines the influence of dependent children (presence, number and age) on parental physical activity levels measured by accelerometry. It is hypothesized that parents will be less active than nonparents and that parents whose youngest child in the home is aged ⬍6 years will be less active than those with older, lessdependent children. The strength of the relationships between the number of dependent children at home and the ages of children on parental physical activity levels was examined using Canadian Health Measures Survey (CHMS) data. In addition, the differences between parents and nonparents, between mothers and fathers, and the contribution of multiple children on physical activity were examined.
Methods Data Source Data for the current study are from the CHMS Cycle 1 (2007– 2009). The CHMS is a comprehensive, direct health measures survey that involved a household interview followed by a visit to a mobile examination center for a series of direct physical measurements. The data were collected from a nationally representative sample of the Canadian population aged 6 –79 years. Approximately 96% of Canadians are represented. Participation was voluntary; respondents could opt out at any time.19 To be nationally representative, the proposed sample size for Cycle 1 of the CHMS was 5000 respondents equally distributed by set age groups August 2012
169
(6 –11 years, 12–19 years, 20 –39 years, 40 –59 years, and 60 –79 years) and gender.20 Using the Labour Force Survey frame clusters, data were collected in 15 collection sites in fıve regions within Canada (British Columbia, the Prairies, Ontario, Quebec, and the Atlantic Provinces). Data collection began in March of 2007 and was continuous until February 2009 with sites ordered to take into account seasonality by region and the temporal effect. A detailed description of the sampling strategy is reported elsewhere.20 Of the selected households, 69.6% agreed to provide a roster of all residents (age and gender). In each responding household, one or two members were chosen to participate in the CHMS based on allocation needs for each age group in the sampling plan; 87.2% of selected adults aged 20 – 65 years completed the household questionnaire, and 83.3% of this group participated in the mobile examination center component. Of the adults for whom an activity monitor was available, 83.0% had at least 4 valid days, defıned as ⱖ10 hours of wear time each day for ⱖ4 days.3,21 After adjustments to account for the sampling strategy,20 the fınal response rate for having at least 4 valid days was 42.0% (69.6% X 87.2% X 83.3% X 83.0%). Additional information on the CHMS can be found in previously published reports.20,22,23
Accelerometry Data After completion of their mobile examination center visit, respondents were asked to wear an Actical accelerometer during waking hours, which provides a minute-by-minute measurement of the intensity of physical activity for 7 consecutive days. The Actical has been validated to measure physical activity and step counts in adults24 and children.25–27 Details of measurement validity and cut-points are described elsewhere.28 Briefly, any observations above 1535 counts per minute (cpm) were identifıed as “moderate,” and any observations above 3962 cpm were identifıed as “vigorous.”29 Moderate-to-vigorous activity was summed each day counting any MVPA accumulated in 10-minute bouts. Published guidelines were followed to identify and remove days with incomplete accelerometer wear time.21,28,30 Wear time was defıned by subtracting nonwear time from 24 hours. Standard accelerometry-derived variables used for these analyses included minutes of MVPA per day and proportion meeting the recommendation of 150 minutes per week or more of MVPA accumulated in 10-minute bouts,31 and the analytic approach was harmonized with that used in the analysis of the 2003–2004 NHANES accelerometry data.21
Population The current study is based on 2315 CHMS respondents (1078 men and 1237 women) aged 20 – 65 years who completed their mobile examination center visit, had measured height and weight information, and valid Actical accelerometry data. Pregnant women were excluded (n⫽28). Based on the household roster information, four derived variables detailing the number of children in the household were provided with the data: the number of children aged ⬍6 years, ⬍12 years, ⬍16 years, and 16 or 17 years. These derived variables made it possible to defıne the number and ages of children within each household.32 The study population was divided into groups based on the presence of a dependent child in the home (yes/no); the number of dependent children in the home (0, 1, ⱖ2); and age of the youngest child (no dependent children in the home, youngest child aged
Adamo et al / Am J Prev Med 2012;43(2):168 –175
170
Table 1. Descriptive characteristics of Canadian adults, by gender, living with or without dependent children Women (n⫽1237)
Men (n⫽1078)
Has a dependent child in the home
Has a dependent child in the home
No
No n
33.0 (28.8, 37.6)
133
35.9 (30.8, 41.3)
29.7 (24.2, 35.8*) 268
51.3 (43.4, 59.2)
306
37.3 (31.9, 43.1*)
54
12.8a (8.6, 18.6)
310 42.3 (34.3, 50.7*) 121 18.2 (13.4, 24.2) 283
41.4 (34.3, 48.9*)
46
8.4a (5.5, 12.5)
58.6 (51.1, 65.7*) 409
91.6 (87.5, 94.5)
n
% (95% CI)
Yes
% (95% CI)
n
% (95% CI)
Yes n
% (95% CI)
Age, years 20–35
152 30.4 (26.2, 34.9)
36–50
153 24.3 (19.4, 30.0*) 351 57.9 (52.1, 63.6) 162
51–65
351 45.3 (37.4, 53.5*)
192 34.2 (29.7, 38.9) 155
38
a
7.9 (4.4, 13.6)
Marital status Single/divorced
Married/common-law 345 57.7 (49.3, 65.7*) 459 81.8 (75.8, 86.6) 340 Education ⬍Degree/diploma
240 39.5 (30.6, 49.1)
167 32.0 (23.2, 42.3) 224
38.3 (27.6, 50.3*) 131
27.1 (19.6, 36.2)
Degree/diploma
409 60.5 (50.9, 69.4)
411 68.0 (57.7, 76.8) 391
61.7 (49.7, 72.4*) 320
72.9 (63.8, 80.4)
Highest
273 48.7 (42.4, 55.1)
263 48.0 (40.7, 55.4) 324
56.8 (51.0, 62.5)
243
57.2 (48.8, 65.1)
Lowest to uppermiddle
354 51.3 (44.9, 57.6)
291 52.0 (44.6, 59.3) 275
43.2 (37.5, 49.0)
199
42.8 (34.9, 51.2)
141 18.8 (14.9, 23.5)
108 17.7 (12.5, 24.4) 104
9.8 (8.0, 11.9)
Income
Employment statusb Do not have a job
61
a
10.8 (7.2, 16.0*)
13
N/A a
Part-time (⬍30 hours 112 17.3 (12.5, 23.6) per week)
131 22.6 (18.2, 27.7)
Full-time (ⱖ30 hours 396 63.8 (56.6, 70.5) per week)
340 59.7 (54.1, 65.1) 454
79.4 (75.0, 83.2*) 421
94.7 (91.4, 96.8)
21
3.9 (2.1, 7.2)
BMI classificationb Normal or underweight
290 48.4 (41.1, 55.9)
280 49.3 (38.8, 59.8) 197
33.3 (27.2, 40.1)
108
26.1 (19.8, 33.7)
Overweight
222 30.4 (26.7, 34.4)
171 27.7 (21.4, 35.1) 283
45.8 (38.6, 53.1)
229
50.3 (40.5, 60.1)
Obese
144 21.1 (15.9, 27.5)
130 23.0 (16.7, 30.8) 143
20.9 (15.4, 27.6)
118
23.5 (16.5, 32.5)
Data should be interpreted with caution because of high sampling variability (coefficient of variation ⱖ16.6 and ⬍33.3). Employment status and BMI are shown here as categoric variables for descriptive purposes only. *Significantly different from those of the same gender with dependent children in the home (p⬍0.05) N/A, not available (estimate not provided because of extreme sampling variability or small sample size)
a b
⬍6 years, youngest child aged 6 –11 years, and youngest child aged 12–17 years). This categorization also coincides with the child’s presumed level of independence, and the available self-report data suggest that the physical activity of parents with children aged ⬍5 years is particularly affected.8
Covariates Marital status was categorized into single/divorced (including widowed and separated) versus married/common-law. Highest education of the respondent also was grouped in two: having a postsecondary degree/diploma versus less than a postsecondary degree/ diploma (e.g., high school diploma or not, some postsecondary education) as the vast majority of the sample had completed post-
secondary education and very few had less than a high school diploma. Total household income, adjusted for number of people in the household, was collapsed into two categories because of limited sample sizes in the lower-income categories: highest income versus upper-middle, lower-middle, and lowest combined.33 Although respondent employment status was categorized for descriptive purposes (Table 1) as those who do not work (assigned 0 hours); those who work part-time (⬍30 hours per week); and those who work full-time (ⱖ30 hours per week); employment information was included in the models as usual hours of work per week on a continuous scale. Finally, BMI was calculated based on direct measures of height and weight and is presented descriptively in Table 1 as underweight (BMI ⬍18.5) and normal weight www.ajpmonline.org
Adamo et al / Am J Prev Med 2012;43(2):168 –175
Table 2. Average minutes of daily MVPA, by gender Women (n⫽1237) Average (95% CI)
n
Men (n⫽1078) n
Average (95% CI)
Has a dependent child in the home No
656
22.9 (19.2, 26.6*)
623
30.8 (25.4, 36.3)
Yes
581
21.1 (18.1, 24.0)
455
25.4 (20.7, 30.0)
Number of dependent children in the home 0
656
22.9 (19.2, 26.6*)
623
30.8 (25.4, 36.3)
1
230
21.3 (17.7, 24.9)
197
25.7 (21.4, 30.0)
ⱖ2
351
20.9 (17.3, 24.5)
258
25.0 (19.3, 30.7)
Age of youngest dependent child in the home (years) 0 (no dependent children)
656
22.9 (19.2, 26.6*)
623
30.8 (25.4, 36.3)
⬍6
231
17.7** (13.8, 21.6*)
210
25.6 (23.9, 27.2)
6–11
221
23.8 (20.3, 27.3)
139
23.7** (16.6, 30.9)
12–17
129
22.9 (18.5, 27.2)
106
26.3a (15.8, 36.8)
171 the independent effects of having dependent children in the home: fırst, whether presence of a dependent child (yes/no) is associated with the dependent variables (Model 1); second, examining the potential association with having multiple dependent children in the home (Model 2); and last, examining whether age of the youngest dependent child in the home is associated with the dependent variables (Model 3).
Results
This CHMS sample included 2315 representative Canadians aged a Data should be interpreted with caution because of high sampling variability (coefficient of variation ⱖ16.6 20 – 65 years (Table 1). and ⬍33.3). The age group with the *Significantly different from men in the same category/column (p⬍0.05) **Significantly different from those of the same gender with no dependent children in the home (p⬍0.05) highest proportion of MVPA, moderate-to-vigorous physical activity dependent children in the home was the group aged 36 –50 years (57.9% in women and 51.3% in men) 34 (18.5 to ⬍25); overweight (25 to ⬍30); and obese (BMI ⱖ30) ; and of those with children at home, more were married/ however, similar to hours worked per week, regression models common-law compared to those without dependent chilexamined BMI on a continuous scale. dren (81.8% vs 18.2% women and 91.6% vs 8.4% for men, Statistical Analysis p⬍0.05). Among men with at least one dependent child All analyses were conducted in SUDAAN v.10, between February in the home, a greater proportion had a college degree/ and December 2011, using weighted data to represent the Canadiploma (72.9%) and worked full-time (94.7%), whereas dian population. Given the low response rate for the CHMS accelfewer had less than a college degree/diploma and worked erometry population, special weights were assigned for this particpart-time, compared to men without dependent children ular subpopulation. Variance estimation (95% CIs) and (27.1% vs 38.3% and 3.9% vs 10.8%, respectively). signifıcance testing (t-tests) on differences between estimates were calculated using the bootstrap weights provided with the data, in For women, daily MVPA was lower among those order to account for the complex sampling design.35,36 Given the whose youngest child was aged ⬍6 years compared to limited number of df in the CHMS available for variance estimathose with no dependent children in the home (Table 2). tion (df⫽11; because of only 15 primary sampling units minus four For men, those whose youngest child was aged 6 –11 years 37 strata), Satterwaite-adjusted statistics were used to test the signifhad lower average daily MVPA than men with no depenicance of each regression model’s coeffıcients. Signifıcance was defıned as a p-value of ⬍0.05. dent children (Table 2). Although not a within-household The present study had two dependent variables: (1) average mincomparison, it is worth noting that the average daily utes of MVPA per day and (2) meeting the current Canadian physical MVPA for women with no dependent children in the activity guidelines of 150 minutes per week of MVPA in bouts of ⱖ10 home was lower than that for men with no dependent minutes (yes/no). Descriptive statistics were computed to estimate the children in the home (p⬍0.05); however, this gender dependent variables by number and age of dependent children in the difference was not evident when there was at least one home, and by other subject characteristics. Linear regression models were conducted to examine the relationship between the covariates dependent child in the home. and minutes of MVPA per day. Logistic regression models examined When adjusted for known covariates, having one child in the odds of meeting the current physical activity guidelines (150 minthe home was not associated with fewer daily minutes of utes per week in bouts ⱖ10 minutes). MVPA in women (⫽ ⫺3.09, p⫽0.106) or men All models were conducted separately by gender and con(⫽ ⫺4.92, p⫽0.089). Age of the youngest child in the trolled for parental age, marital status, total household income, usual hours of work per week, and BMI. Each model examined home (Table 3; Model 3) was negatively associated with August 2012
Adamo et al / Am J Prev Med 2012;43(2):168 –175
172 a
Table 3. Linear regression models estimating minutes of daily MVPA among women and men Model 1b Variable
Model 2c
Model 3d
 (95% CI)
p-value
 (95% CI)
p-value
 (95% CI)
p-value
0.110
—
0.110
—
0.126
—
WOMEN R2 Children
0.180
No
—
—
—
—
—
—
Yes
⫺2.57 (⫺6.51, 1.38)
0.180
—
—
—
—
Number of children
0.263
0
—
—
—
—
—
—
1
—
—
⫺3.09 (⫺6.95, 0.77)
0.106
—
—
ⱖ2
—
—
⫺2.10 (⫺6.87, 2.67)
0.352
—
—
Age of youngest child (years)
0.015
0 (no dependent children)
—
—
—
—
—
—
⬍6
—
—
—
—
6⫺11
—
—
—
—
1.13 (⫺3.93, 6.19)
0.632
12–17
—
—
—
—
0.39 (⫺3.68, 4.46)
0.836
0.125
—
0.125
—
0.127
—
⫺7.74 (⫺12.95, ⫺2.53) 0.007
MEN R2 Children
0.138
No
—
—
—
—
—
—
Yes
⫺4.47 (⫺10.61, 1.68)
0.138
—
—
—
—
Number of children
0.185
0
—
—
—
—
—
—
1
—
—
⫺4.92 (⫺10.73, 0.89)
0.089
—
—
ⱖ2
—
—
⫺3.95 (⫺11.24, 3.35)
0.259
—
—
Age of youngest child (years)
0.245
0 (No dependent children)
—
—
—
—
—
—
⬍6
—
—
—
—
⫺5.67 (⫺11.79, 0.44)
0.066
6–11
—
—
—
—
⫺5.10 (⫺12.26, 2.05)
0.145
12–17
—
—
—
—
⫺2.47 (⫺11.46, 6.51)
0.557
a
All models are adjusted for parental age, marital status, household income, usual hours of work each week and BMI. Model 1 explores the association of having a dependent child in the home. c Model 2 explores the association of having more than one dependent child in the home. d Model 3 explores the association of age of youngest dependent child in the home. b
minutes of MVPA among women (p⫽0.015); specifıcally, women whose youngest child was aged ⬍6 years engaged in 7.7 fewer minutes of daily MVPA (54 minutes per week less) compared to those without dependent children (p⫽0.007). Among men, age of youngest child was not associated with minutes of MVPA (p⫽0.245); however, having a dependent child in the home aged ⬍6 years was marginally associated with 5.7 minutes less daily (40 minutes per week less) MVPA than men without dependent children (p⫽0.066).
Fewer mothers and fathers (⬍10%) with children aged ⬍6 years were meeting recommendations compared to those without dependent children (mothers: 7.9% vs 18.1%, fathers: 9.8% vs 23%, respectively). According to a logistic regression analysis (Figure 1), age of the youngest dependent child in the home was associated with meeting the physical activity guidelines (p⬍0.05). Specifıcally, parents whose youngest dependent child was aged ⬍6 years were less likely to meet the guidelines than those www.ajpmonline.org
Adamo et al / Am J Prev Med 2012;43(2):168 –175 Age of youngest child (years) Mothers <6 11 12 17 Fathers <6 11 12 17
173
AOR (95% CI) 0.31 (0.11, 0.87) 0.57 (0.16, 2.07) 0.58 (0.28, 1.19) 0.34 (0.13, 0.93) 0.41 (0.09, 1.82) 0.52 (0.20, 1.35) 0
0.2
0.4 0.6 0.8 1.0 1.2 1.4 1.6 Decreased likelihood Increased likelihood
1.8
2.0
2.2
Figure 1. AORs and CIs for meeting physical activity guidelines based on age of youngest dependent child Note: Values are adjusted for parental age, marital status, income, usual hours of work per week and BMI.
with no dependent children in the home (OR⫽0.31, 95% CI⫽0.11, 0.87 (p⫽0.030) for women, OR⫽0.34, 95% CI⫽0.13, 0.93 (p⫽0.037) for men).
Discussion Compared to those without children, daily minutes of MVPA were lower among mothers whose youngest child was aged ⬍6 years and fathers whose youngest child was aged 6 –11 years. In addition, fewer parents than nonparents were meeting the recommended guidelines of 150 minutes per week, in bouts of ⱖ10 minutes. These data illustrate that it is not simply the presence or the number of dependent children in the home but rather the age of the youngest child that is associated with parental levels of MVPA. Importantly, this fınding persisted in mothers and marginally among fathers when the analyses were adjusted for parental age, marital status, income, employment, and BMI. The magnitude of difference in MVPA accumulated per week between parents of children aged ⬍6 years and nonparents (⫺54 minutes per week in women and ⫺40 minutes per week in men; Table 3) is important to consider given that the current physical activity guidelines recommend that adults accumulate 150 minutes per week of MVPA. In other words, parents of young children are missing out on approximately one third of their weekly MVPA needs. To the authors’ knowledge, this is the fırst study to examine the relationship between age of the youngest dependent child in the home and directly measured physical activity levels of mothers and fathers. Consistent with self-report data from the U.S.,8 the accelerometry results presented here indicated that among mothers, when compared to those without children in the home, having a dependent child in the home was signifıcantly associated with physical activity only when the youngest child was aged ⬍6 years. Thus, the hypothesis that parents of young children would be less active than those with older (less-dependent) children was supported. When looking at unadjusted data August 2012
(Table 2), the physical activity levels of women and men were affected by age of the youngest child at home; specifıcally, women whose youngest child was aged ⬍6 years and men whose youngest child was aged 6 –11 years had lower MVPA than those without dependent children. The linear regression models (Table 3), which accounted for known covariates, found that the association between age of youngest child at home remained signifıcant for women, with the youngest age group (⬍6 years) driving this trend. For men, although there was no association overall between age of youngest child and minutes of MVPA, men with a dependent child in the youngest age category had marginally fewer minutes of MVPA than men with no dependent children. The logistic regression examining the likelihood of meeting physical activity guidelines confırmed these fındings with an effect in parents regardless of gender. Overall, the fındings of the current study suggest that having dependent children at home has some influence on parental levels of MVPA; however, the degree of influence appears to be slightly more powerful in women compared to men, especially when the child is very young (⫺54 vs ⫺40 minutes per week or ⫺7.7 vs ⫺5.7 minutes per day). The CHMS does not collect the more-qualitative data that would be needed to offer potential explanations for this gender difference; however, evidence from previous sociology-based research provides arguments about why such a trend may have emerged in our data. For instance, there is some evidence to suggest that women spend more time, and incur greater responsibility, for child care compared to men.38 – 42 In particular, the early years of child-rearing (i.e., when children are aged ⬍6 years) has been identifıed as a time when this may be particularly true,43– 45 a trend that is probably due to factors including— but not limited to—a greater restriction on a mother’s leisure activities during this time, interrupted sleep patterns, and the substantive time required for exclusive breastfeeding.11
174
Adamo et al / Am J Prev Med 2012;43(2):168 –175
These data suggest that adults with young children at home face additional barriers to engaging in regular physical activity when compared to adults without children. Strategies aimed at increasing population levels of physical activity should acknowledge this reality and consider including alternative and creative solutions to help parents overcome these barriers. For example, communities could create physical activity–friendly environments, making it simple and safe for parents to adopt active lifestyles with or without their children (e.g., parks within walking/biking distance to residential areas and daycare centers, well-lit paths or sidewalks, community fıtness centers or programs with child care).
Strengths and Limitations Important strengths of the present study include the use of directly measured MVPA in a large, representative sample of the Canadian population. Additionally, the relationship between dependent children and age of the youngest child in the home and physical activity levels of both mothers and fathers was examined, with the recognition that a child’s age is an important variable with regard to parental responsibility. This secondary analysis is limited by its cross-sectional design, which does not allow for investigation of changes in MVPA over time in individuals as they transition through various stages of parenthood. It would have been useful to further categorize those children aged ⬍6 years, but the CHMS survey design does not allow this. It is possible that some respondents were misclassifıed according to the presence or absence of dependent children in the home; however, 99% of the individuals identifıed as having dependent children also described themselves as parents in a separate question. Although the CHMS provides information on household type, according to the survey data, there is no appropriate indication of the level of additional child care or support and thus could not be controlled for in these analyses. Nevertheless, when examining the data controlling for number of people in the household rather than marital status, the results did not change. The low response rate and accelerometry compliance of the CHMS (42%) is another limitation. The subsample of respondents with at least 4 valid days of accelerometry data was weighted to represent the Canadian population and has been corrected for nonresponse bias. Nonrespondents on the accelerometry fıle tended to be younger, male, and more obese than respondents and therefore may differ in terms of physical activity levels and/or family status.3 Although well-known contributing factors were included, the low R2 of the models tested suggested that some key explanatory variables were missing. How-
ever, the 11 df in the CHMS limited the number of parameters included in each model.
Conclusion Age of the youngest dependent child at home is associated with physical activity levels of Canadian parents. The association was stronger in women compared to men; however, overall it appears that when parents have a child aged ⬍6 years in the home, this is a critical time for parental physical activity engagement. Future research should explore further breakdowns of the ⬍6 years age group (i.e., infants, toddlers, preschoolers) to examine whether the relationship with parental physical activity differs within this early childhood window, and focus on strategies to encourage parents with young children to establish or re-engage in a physically active lifestyle, not only for their own health but to model healthy behavior for their children. The authors thank the Physical Health Measures Division and the Health Analysis Division of Statistics Canada for providing scientifıc and IRB approval of this manuscript. Special thanks to Didier Garriguet for his assistance in guiding the data analyses and for his critical review of the manuscript. KBA and RCC receive research funding from various governmental funding agencies, nongovernmental (not-for-profıt) organizations, and philanthropic groups. No fınancial disclosures were reported by the authors of this paper.
References 1. Tremblay MS, Colley RC, Saunders TJ, Healy GN, Owen N. Physiological and health implications of a sedentary lifestyle. Appl Physiol Nutr Metab 2010;35(6):725– 40. 2. Schoenborn CA, Stommel M. Adherence to the 2008 adult physical activity guidelines and mortality risk. Am J Prev Med 2011;40(5): 514 –21. 3. Colley RC, Garriguet D, Janssen I, Craig CL, Clarke J, Tremblay MS. Physical activity of Canadian adults: accelerometer results from the 2007 to 2009 Canadian Health Measures Survey. Health Rep 2011;22(1):7–14. 4. Guthold R, Ono T, Strong KL, Chatterji S, Morabia A. Worldwide variability in physical inactivity a 51-country survey. Am J Prev Med 2008;34(6):486 –94. 5. Bellows-Riecken KH, Rhodes RE. A birth of inactivity? A review of physical activity and parenthood. Prev Med 2008;46(2):99 –110. 6. Pereira MA, Rifas-Shiman SL, Kleinman KP, Rich-Edwards JW, Peterson KE, Gillman MW. Predictors of change in physical activity during and after pregnancy: Project Viva. Am J Prev Med 2007;32(4):312–9. 7. Gennaro S, Fehder W. Health behaviors in postpartum women. Fam Community Health 2000;22(4):16 –26. 8. Nomaguchi KM, Bianchi SM. Exercise time: gender differences in the effects of marriage, parenthood, and employment. J Marriage Fam 2004;66:413–30.
www.ajpmonline.org
Adamo et al / Am J Prev Med 2012;43(2):168 –175 9. Fahrenwald NL, Atwood JR, Walker SN, Johnson DR, Berg K. A randomized pilot test of “Moms on the Move”: a physical activity intervention for WIC mothers. Ann Behav Med 2004;27(2):82–90. 10. Brown WJ, Mishra G, Lee C, Bauman A. Leisure time physical activity in Australian women: relationship with well being and symptoms. Res Q Exerc Sport 2000;71(3):206 –16. 11. Brown PR, Brown WJ, Miller YD, Hansen V. Perceived constraints and social support for active leisure among mothers with young children. Leisure Sci 2001;23:131– 44. 12. Brown WJ, Trost SG. Life transitions and changing physical activity patterns in young women. Am J Prev Med 2003;25(2):140 –3. 13. Andajani-Sutjahjo S, Ball K, Warren N, Inglis V, Crawford D. Perceived personal, social and environmental barriers to weight maintenance among young women: a community survey. Int J Behav Nutr Phys Act 2004;1(1):15. 14. Verhoef MJ, Love EJ, Rose MS. Women’s social roles and their exercise participation. Women Health 1992;19(4):15–29. 15. Verhoef MJ, Love EJ. Women and exercise participation: the mixed blessings of motherhood. Health Care Women Int 1994;15(4):297–306. 16. Verhoef MJ, Love EJ. Women’s exercise participation: the relevance of social roles compared to non-role-related determinants. Can J Public Health 1992;83(5):367–70. 17. Hamilton K, White KM. Identifying key belief-based targets for promoting regular physical activity among mothers and fathers with young children. J Sci Med Sport 2011;14(2):135– 42. 18. Jones C, Burns S, Howat P, Jancey J, McManus A, Carter O. Playgroups as a setting for nutrition and physical activity interventions for mothers with young children: exploratory qualitative fındings. Health Promot J Austr 2010;21(2):92– 8. 19. Tremblay M, Wolfson M, Gorber SC. Canadian Health Measures Survey: rationale, background and overview. Health Rep 2007;18(S):7–20. 20. Giroux S. Canadian Health Measures Survey: sampling strategy overview. Health Rep 2007;18(S):31– 6. 21. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc 2008;40(1):181– 8. 22. Bryan S, St-Denis M, Wojtas D. Canadian Health Measures Survey: clinic operations and logistics. Health Rep 2007;18(S):53–70. 23. Day B, Langlois R, Tremblay M, Knoppers BM. Canadian Health Measures Survey: ethical, legal and social issues. Health Rep 2007; 18(S):37–51. 24. Heil DP. Predicting activity energy expenditure using the Actical activity monitor. Res Q Exerc Sport 2006;77(1):64 – 80. 25. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of two objective measures of physical activity for children. J Sports Sci 2008;26(14):1557– 65. 26. Puyau MR, Adolph AL, Vohra FA, Zakeri I, Butte NF. Prediction of activity energy expenditure using accelerometers in children. Med Sci Sports Exerc 2004;36(9):1625–31.
August 2012
175
27. Esliger DW, Probert A, Gorber SC, Bryan S, Laviolette M, Tremblay MS. Validity of the Actical accelerometer step-count function. Med Sci Sports Exerc 2007;39(7):1200 – 4. 28. Colley R, Gorber SC, Tremblay MS. Quality control and data reduction procedures for accelerometry-derived measures of physical activity. Health Rep 2010;21(1):63–9. 29. Colley RC, Tremblay MS. Moderate and vigorous physical activity intensity cut-points for the Actical accelerometer. J Sports Sci 2011; 29(8):783–9. 30. National Cancer Institute. Risk factor monitoring and methods: SAS programs for analyzing NHANES 2003–2004 accelerometer data. 8-16-2007. 9-8-2010 31. Tremblay MS, Warburton ER, Janssen I, et al. New Canadian Physical Activity Guidelines. Appl Physiol Nutr Metab 2011;36:36 – 46. 32. Statistics Canada. Canadian Health Measures Survey: Cycle 1. 2007– 2009. Household Questionnaire. 2009. 33. Statistics Canada. Canadian Health Measures Survey Cycle 1 Wave 1 Derived Variable (DV) Specifıcations. 2009. 34. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard defınition for child overweight and obesity worldwide: international survey. BMJ 2000;320(7244):1240 –3. 35. Rao JNK, Wu CFJ, Yue K. Some recent work on resampling methods for complex surveys. Survey Methodol 1992;18(2):209 –17. 36. Rust KF, Rao JN. Variance estimation for complex surveys using replication techniques. Stat Methods Med Res 1996;5(3):283–310. 37. Statistics Canada. Canadian Health Measures Survey (CHMS) Data User guide: Cycle 1. 2010. 38. Bailey WT. A longitudinal study of fathers’ involvement with young children: infancy to age 5 years. J Genet Psychol 1994;155(3):331–9. 39. Craig L. Does father care mean fathers share? A comparison of how mothers and fathers in intact families spend time with children. Gender Soc 2006;20(2):259 – 81. 40. Jackson EL, Henderson KA. Gender-based analysis of leisure constraints. Leisure Sci 1995;17:31–51. 41. McBride BA, Mills G. A comparison of mother and father involvement with their preschool age children. Early Childhood Res Q 1993;8: 457–77. 42. Miller YD, Brown WJ. Determinant of active leisure for women with young children—an “Ethic of Care” prevails. Leisure Sci 2005;27: 405–20. 43. Baxter J. Patterns of change and stability in the gender division of household labour in Australia, 1996 –1997. J Sociol 2002;38(4): 399 – 424. 44. Casper L, Biachi S. Continuity & change in the American family. Thousand Oaks CA: Sage, 2002. 45. Craig L, Bittman M. The effect of children on adults’ time-use: analysis of the incremental time costs of children in Australia. Sydney: Social Policy Research Centre, 2005. SPRC discussion paper no. 143.
Purpose: To investigate the association of children of different ages in the home on two measures of parental physical activity: daily moderate-to-vigorous physical activity (MVPA) and likelihood of meeting the guideline of 150 minutes of MVPA per week accumulated in 10-minute bouts. Methods: Data were from the 2007–2009 Canadian Health Measures Survey (n⫽2315), and analyses were conducted between February and December 2011. MVPA was measured directly using accelerometry. Linear (minutes of MVPA) and logistic (meeting physical activity guidelines) regression models were performed to determine if the presence, number of children, or the age of the youngest child at home was associated with parental physical activity. All models were adjusted for parental age, marital status, household income, employment, and BMI. Results: Mothers whose youngest child was aged ⬍6 years and fathers whose youngest was aged 6 –11 years engaged in fewer minutes of daily MVPA than those without dependent children. Linear regression results identifıed that in comparison to those without children, women whose youngest child in the home was aged ⬍6 years participated in 7.7 minutes less activity per day (p⫽0.007) whereas men engaged in 5.7 fewer minutes per day, or 54 and 40 minutes per week less, respectively. Similarly, logistic regression analyses indicated that both women and men were less likely to meet guidelines if their youngest child in the home was aged ⬍6 years (OR⫽0.31, 95% CI⫽0.11, 0.87; OR⫽0.34, 95% CI⫽0.13, 0.93). Conclusions: The physical activity level of parents with young children present in the home was lower than that of those without children. Given the many physiologic, psychological, and social benefıts of healthy active living, research efforts should continue to focus on strategies to encourage parents with young children to establish or re-engage in a physically active lifestyle, not only for their own health but to model healthy behavior for the next generation. (Am J Prev Med 2012;43(2):168 –175) © 2012 American Journal of Preventive Medicine
Introduction
From the Healthy Active Living and Obesity Research Group, Children’s Hospital of Eastern Ontario Research Institute (Adamo, Brett, Colley); the Faculty of Medicine, Department of Pediatrics (Adamo, Colley); the Faculty of Health Sciences, School of Human Kinetics (Adamo, Brett, Colley), University of Ottawa; and the Health Analysis Division (Langlois, Colley), Statistics Canada, Ottawa, Ontario, Canada Address correspondence to: Kristi B. Adamo, PhD, Children’s Hospital of Eastern Ontario Research Institute, 401 Smyth Road, Ottawa, ON K1H 8L1, Canada. E-mail: [email protected]. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2012.02.032
168 Am J Prev Med 2012;43(2):168 –175
S
ubstantive research has profıled the health-related risks of physical inactivity and a sedentary lifestyle.1 Physical activity is associated with reductions in both all-cause mortality and risk for multiple chronic diseases.2 Despite public awareness of the benefıts of an active lifestyle, physical activity participation continues to be low in adults in Canada3 and worldwide.4 Many factors are associated with low levels of physical activity in the population (such as age, income, gender, BMI), and having dependent children in the home has been sug-
© 2012 American Journal of Preventive Medicine • Published by Elsevier Inc.
Adamo et al / Am J Prev Med 2012;43(2):168 –175 5
gested as a contributor to lowered levels in adults, particularly women.6,7 Having children aged ⬍5 years, in particular, is associated negatively with self-reported exercise.8 Clarifying the relationship between parenthood and physical activity is diffıcult given that young children affect parental levels of physical activity differently than older children because of changing degrees of dependency on parental time and attention. A recent systematic review noted that parents, specifıcally mothers, are less active than those without children; however, the literature is mixed in regard to the impact of the age and number of children at home as predictors of parental physical activity.5 In addition, research concerning parental physical activity habits is limited by the use of self-reported data, with only one study using pedometers as a direct measure of physical activity.9 There is also limited evidence on physical activity levels in fatherhood, with the majority of the research targeting mothers.10 –16 Further, previous research has focused primarily on qualifying the barriers to physical activity,5,11,17,18 rather than quantifying the effects of number and age of dependent children at home on levels of parental physical activity. The need to examine the impact of children on physical activity patterns of both parents using objective measures has been called for.5 The present study examines the influence of dependent children (presence, number and age) on parental physical activity levels measured by accelerometry. It is hypothesized that parents will be less active than nonparents and that parents whose youngest child in the home is aged ⬍6 years will be less active than those with older, lessdependent children. The strength of the relationships between the number of dependent children at home and the ages of children on parental physical activity levels was examined using Canadian Health Measures Survey (CHMS) data. In addition, the differences between parents and nonparents, between mothers and fathers, and the contribution of multiple children on physical activity were examined.
Methods Data Source Data for the current study are from the CHMS Cycle 1 (2007– 2009). The CHMS is a comprehensive, direct health measures survey that involved a household interview followed by a visit to a mobile examination center for a series of direct physical measurements. The data were collected from a nationally representative sample of the Canadian population aged 6 –79 years. Approximately 96% of Canadians are represented. Participation was voluntary; respondents could opt out at any time.19 To be nationally representative, the proposed sample size for Cycle 1 of the CHMS was 5000 respondents equally distributed by set age groups August 2012
169
(6 –11 years, 12–19 years, 20 –39 years, 40 –59 years, and 60 –79 years) and gender.20 Using the Labour Force Survey frame clusters, data were collected in 15 collection sites in fıve regions within Canada (British Columbia, the Prairies, Ontario, Quebec, and the Atlantic Provinces). Data collection began in March of 2007 and was continuous until February 2009 with sites ordered to take into account seasonality by region and the temporal effect. A detailed description of the sampling strategy is reported elsewhere.20 Of the selected households, 69.6% agreed to provide a roster of all residents (age and gender). In each responding household, one or two members were chosen to participate in the CHMS based on allocation needs for each age group in the sampling plan; 87.2% of selected adults aged 20 – 65 years completed the household questionnaire, and 83.3% of this group participated in the mobile examination center component. Of the adults for whom an activity monitor was available, 83.0% had at least 4 valid days, defıned as ⱖ10 hours of wear time each day for ⱖ4 days.3,21 After adjustments to account for the sampling strategy,20 the fınal response rate for having at least 4 valid days was 42.0% (69.6% X 87.2% X 83.3% X 83.0%). Additional information on the CHMS can be found in previously published reports.20,22,23
Accelerometry Data After completion of their mobile examination center visit, respondents were asked to wear an Actical accelerometer during waking hours, which provides a minute-by-minute measurement of the intensity of physical activity for 7 consecutive days. The Actical has been validated to measure physical activity and step counts in adults24 and children.25–27 Details of measurement validity and cut-points are described elsewhere.28 Briefly, any observations above 1535 counts per minute (cpm) were identifıed as “moderate,” and any observations above 3962 cpm were identifıed as “vigorous.”29 Moderate-to-vigorous activity was summed each day counting any MVPA accumulated in 10-minute bouts. Published guidelines were followed to identify and remove days with incomplete accelerometer wear time.21,28,30 Wear time was defıned by subtracting nonwear time from 24 hours. Standard accelerometry-derived variables used for these analyses included minutes of MVPA per day and proportion meeting the recommendation of 150 minutes per week or more of MVPA accumulated in 10-minute bouts,31 and the analytic approach was harmonized with that used in the analysis of the 2003–2004 NHANES accelerometry data.21
Population The current study is based on 2315 CHMS respondents (1078 men and 1237 women) aged 20 – 65 years who completed their mobile examination center visit, had measured height and weight information, and valid Actical accelerometry data. Pregnant women were excluded (n⫽28). Based on the household roster information, four derived variables detailing the number of children in the household were provided with the data: the number of children aged ⬍6 years, ⬍12 years, ⬍16 years, and 16 or 17 years. These derived variables made it possible to defıne the number and ages of children within each household.32 The study population was divided into groups based on the presence of a dependent child in the home (yes/no); the number of dependent children in the home (0, 1, ⱖ2); and age of the youngest child (no dependent children in the home, youngest child aged
Adamo et al / Am J Prev Med 2012;43(2):168 –175
170
Table 1. Descriptive characteristics of Canadian adults, by gender, living with or without dependent children Women (n⫽1237)
Men (n⫽1078)
Has a dependent child in the home
Has a dependent child in the home
No
No n
33.0 (28.8, 37.6)
133
35.9 (30.8, 41.3)
29.7 (24.2, 35.8*) 268
51.3 (43.4, 59.2)
306
37.3 (31.9, 43.1*)
54
12.8a (8.6, 18.6)
310 42.3 (34.3, 50.7*) 121 18.2 (13.4, 24.2) 283
41.4 (34.3, 48.9*)
46
8.4a (5.5, 12.5)
58.6 (51.1, 65.7*) 409
91.6 (87.5, 94.5)
n
% (95% CI)
Yes
% (95% CI)
n
% (95% CI)
Yes n
% (95% CI)
Age, years 20–35
152 30.4 (26.2, 34.9)
36–50
153 24.3 (19.4, 30.0*) 351 57.9 (52.1, 63.6) 162
51–65
351 45.3 (37.4, 53.5*)
192 34.2 (29.7, 38.9) 155
38
a
7.9 (4.4, 13.6)
Marital status Single/divorced
Married/common-law 345 57.7 (49.3, 65.7*) 459 81.8 (75.8, 86.6) 340 Education ⬍Degree/diploma
240 39.5 (30.6, 49.1)
167 32.0 (23.2, 42.3) 224
38.3 (27.6, 50.3*) 131
27.1 (19.6, 36.2)
Degree/diploma
409 60.5 (50.9, 69.4)
411 68.0 (57.7, 76.8) 391
61.7 (49.7, 72.4*) 320
72.9 (63.8, 80.4)
Highest
273 48.7 (42.4, 55.1)
263 48.0 (40.7, 55.4) 324
56.8 (51.0, 62.5)
243
57.2 (48.8, 65.1)
Lowest to uppermiddle
354 51.3 (44.9, 57.6)
291 52.0 (44.6, 59.3) 275
43.2 (37.5, 49.0)
199
42.8 (34.9, 51.2)
141 18.8 (14.9, 23.5)
108 17.7 (12.5, 24.4) 104
9.8 (8.0, 11.9)
Income
Employment statusb Do not have a job
61
a
10.8 (7.2, 16.0*)
13
N/A a
Part-time (⬍30 hours 112 17.3 (12.5, 23.6) per week)
131 22.6 (18.2, 27.7)
Full-time (ⱖ30 hours 396 63.8 (56.6, 70.5) per week)
340 59.7 (54.1, 65.1) 454
79.4 (75.0, 83.2*) 421
94.7 (91.4, 96.8)
21
3.9 (2.1, 7.2)
BMI classificationb Normal or underweight
290 48.4 (41.1, 55.9)
280 49.3 (38.8, 59.8) 197
33.3 (27.2, 40.1)
108
26.1 (19.8, 33.7)
Overweight
222 30.4 (26.7, 34.4)
171 27.7 (21.4, 35.1) 283
45.8 (38.6, 53.1)
229
50.3 (40.5, 60.1)
Obese
144 21.1 (15.9, 27.5)
130 23.0 (16.7, 30.8) 143
20.9 (15.4, 27.6)
118
23.5 (16.5, 32.5)
Data should be interpreted with caution because of high sampling variability (coefficient of variation ⱖ16.6 and ⬍33.3). Employment status and BMI are shown here as categoric variables for descriptive purposes only. *Significantly different from those of the same gender with dependent children in the home (p⬍0.05) N/A, not available (estimate not provided because of extreme sampling variability or small sample size)
a b
⬍6 years, youngest child aged 6 –11 years, and youngest child aged 12–17 years). This categorization also coincides with the child’s presumed level of independence, and the available self-report data suggest that the physical activity of parents with children aged ⬍5 years is particularly affected.8
Covariates Marital status was categorized into single/divorced (including widowed and separated) versus married/common-law. Highest education of the respondent also was grouped in two: having a postsecondary degree/diploma versus less than a postsecondary degree/ diploma (e.g., high school diploma or not, some postsecondary education) as the vast majority of the sample had completed post-
secondary education and very few had less than a high school diploma. Total household income, adjusted for number of people in the household, was collapsed into two categories because of limited sample sizes in the lower-income categories: highest income versus upper-middle, lower-middle, and lowest combined.33 Although respondent employment status was categorized for descriptive purposes (Table 1) as those who do not work (assigned 0 hours); those who work part-time (⬍30 hours per week); and those who work full-time (ⱖ30 hours per week); employment information was included in the models as usual hours of work per week on a continuous scale. Finally, BMI was calculated based on direct measures of height and weight and is presented descriptively in Table 1 as underweight (BMI ⬍18.5) and normal weight www.ajpmonline.org
Adamo et al / Am J Prev Med 2012;43(2):168 –175
Table 2. Average minutes of daily MVPA, by gender Women (n⫽1237) Average (95% CI)
n
Men (n⫽1078) n
Average (95% CI)
Has a dependent child in the home No
656
22.9 (19.2, 26.6*)
623
30.8 (25.4, 36.3)
Yes
581
21.1 (18.1, 24.0)
455
25.4 (20.7, 30.0)
Number of dependent children in the home 0
656
22.9 (19.2, 26.6*)
623
30.8 (25.4, 36.3)
1
230
21.3 (17.7, 24.9)
197
25.7 (21.4, 30.0)
ⱖ2
351
20.9 (17.3, 24.5)
258
25.0 (19.3, 30.7)
Age of youngest dependent child in the home (years) 0 (no dependent children)
656
22.9 (19.2, 26.6*)
623
30.8 (25.4, 36.3)
⬍6
231
17.7** (13.8, 21.6*)
210
25.6 (23.9, 27.2)
6–11
221
23.8 (20.3, 27.3)
139
23.7** (16.6, 30.9)
12–17
129
22.9 (18.5, 27.2)
106
26.3a (15.8, 36.8)
171 the independent effects of having dependent children in the home: fırst, whether presence of a dependent child (yes/no) is associated with the dependent variables (Model 1); second, examining the potential association with having multiple dependent children in the home (Model 2); and last, examining whether age of the youngest dependent child in the home is associated with the dependent variables (Model 3).
Results
This CHMS sample included 2315 representative Canadians aged a Data should be interpreted with caution because of high sampling variability (coefficient of variation ⱖ16.6 20 – 65 years (Table 1). and ⬍33.3). The age group with the *Significantly different from men in the same category/column (p⬍0.05) **Significantly different from those of the same gender with no dependent children in the home (p⬍0.05) highest proportion of MVPA, moderate-to-vigorous physical activity dependent children in the home was the group aged 36 –50 years (57.9% in women and 51.3% in men) 34 (18.5 to ⬍25); overweight (25 to ⬍30); and obese (BMI ⱖ30) ; and of those with children at home, more were married/ however, similar to hours worked per week, regression models common-law compared to those without dependent chilexamined BMI on a continuous scale. dren (81.8% vs 18.2% women and 91.6% vs 8.4% for men, Statistical Analysis p⬍0.05). Among men with at least one dependent child All analyses were conducted in SUDAAN v.10, between February in the home, a greater proportion had a college degree/ and December 2011, using weighted data to represent the Canadiploma (72.9%) and worked full-time (94.7%), whereas dian population. Given the low response rate for the CHMS accelfewer had less than a college degree/diploma and worked erometry population, special weights were assigned for this particpart-time, compared to men without dependent children ular subpopulation. Variance estimation (95% CIs) and (27.1% vs 38.3% and 3.9% vs 10.8%, respectively). signifıcance testing (t-tests) on differences between estimates were calculated using the bootstrap weights provided with the data, in For women, daily MVPA was lower among those order to account for the complex sampling design.35,36 Given the whose youngest child was aged ⬍6 years compared to limited number of df in the CHMS available for variance estimathose with no dependent children in the home (Table 2). tion (df⫽11; because of only 15 primary sampling units minus four For men, those whose youngest child was aged 6 –11 years 37 strata), Satterwaite-adjusted statistics were used to test the signifhad lower average daily MVPA than men with no depenicance of each regression model’s coeffıcients. Signifıcance was defıned as a p-value of ⬍0.05. dent children (Table 2). Although not a within-household The present study had two dependent variables: (1) average mincomparison, it is worth noting that the average daily utes of MVPA per day and (2) meeting the current Canadian physical MVPA for women with no dependent children in the activity guidelines of 150 minutes per week of MVPA in bouts of ⱖ10 home was lower than that for men with no dependent minutes (yes/no). Descriptive statistics were computed to estimate the children in the home (p⬍0.05); however, this gender dependent variables by number and age of dependent children in the difference was not evident when there was at least one home, and by other subject characteristics. Linear regression models were conducted to examine the relationship between the covariates dependent child in the home. and minutes of MVPA per day. Logistic regression models examined When adjusted for known covariates, having one child in the odds of meeting the current physical activity guidelines (150 minthe home was not associated with fewer daily minutes of utes per week in bouts ⱖ10 minutes). MVPA in women (⫽ ⫺3.09, p⫽0.106) or men All models were conducted separately by gender and con(⫽ ⫺4.92, p⫽0.089). Age of the youngest child in the trolled for parental age, marital status, total household income, usual hours of work per week, and BMI. Each model examined home (Table 3; Model 3) was negatively associated with August 2012
Adamo et al / Am J Prev Med 2012;43(2):168 –175
172 a
Table 3. Linear regression models estimating minutes of daily MVPA among women and men Model 1b Variable
Model 2c
Model 3d
 (95% CI)
p-value
 (95% CI)
p-value
 (95% CI)
p-value
0.110
—
0.110
—
0.126
—
WOMEN R2 Children
0.180
No
—
—
—
—
—
—
Yes
⫺2.57 (⫺6.51, 1.38)
0.180
—
—
—
—
Number of children
0.263
0
—
—
—
—
—
—
1
—
—
⫺3.09 (⫺6.95, 0.77)
0.106
—
—
ⱖ2
—
—
⫺2.10 (⫺6.87, 2.67)
0.352
—
—
Age of youngest child (years)
0.015
0 (no dependent children)
—
—
—
—
—
—
⬍6
—
—
—
—
6⫺11
—
—
—
—
1.13 (⫺3.93, 6.19)
0.632
12–17
—
—
—
—
0.39 (⫺3.68, 4.46)
0.836
0.125
—
0.125
—
0.127
—
⫺7.74 (⫺12.95, ⫺2.53) 0.007
MEN R2 Children
0.138
No
—
—
—
—
—
—
Yes
⫺4.47 (⫺10.61, 1.68)
0.138
—
—
—
—
Number of children
0.185
0
—
—
—
—
—
—
1
—
—
⫺4.92 (⫺10.73, 0.89)
0.089
—
—
ⱖ2
—
—
⫺3.95 (⫺11.24, 3.35)
0.259
—
—
Age of youngest child (years)
0.245
0 (No dependent children)
—
—
—
—
—
—
⬍6
—
—
—
—
⫺5.67 (⫺11.79, 0.44)
0.066
6–11
—
—
—
—
⫺5.10 (⫺12.26, 2.05)
0.145
12–17
—
—
—
—
⫺2.47 (⫺11.46, 6.51)
0.557
a
All models are adjusted for parental age, marital status, household income, usual hours of work each week and BMI. Model 1 explores the association of having a dependent child in the home. c Model 2 explores the association of having more than one dependent child in the home. d Model 3 explores the association of age of youngest dependent child in the home. b
minutes of MVPA among women (p⫽0.015); specifıcally, women whose youngest child was aged ⬍6 years engaged in 7.7 fewer minutes of daily MVPA (54 minutes per week less) compared to those without dependent children (p⫽0.007). Among men, age of youngest child was not associated with minutes of MVPA (p⫽0.245); however, having a dependent child in the home aged ⬍6 years was marginally associated with 5.7 minutes less daily (40 minutes per week less) MVPA than men without dependent children (p⫽0.066).
Fewer mothers and fathers (⬍10%) with children aged ⬍6 years were meeting recommendations compared to those without dependent children (mothers: 7.9% vs 18.1%, fathers: 9.8% vs 23%, respectively). According to a logistic regression analysis (Figure 1), age of the youngest dependent child in the home was associated with meeting the physical activity guidelines (p⬍0.05). Specifıcally, parents whose youngest dependent child was aged ⬍6 years were less likely to meet the guidelines than those www.ajpmonline.org
Adamo et al / Am J Prev Med 2012;43(2):168 –175 Age of youngest child (years) Mothers <6 11 12 17 Fathers <6 11 12 17
173
AOR (95% CI) 0.31 (0.11, 0.87) 0.57 (0.16, 2.07) 0.58 (0.28, 1.19) 0.34 (0.13, 0.93) 0.41 (0.09, 1.82) 0.52 (0.20, 1.35) 0
0.2
0.4 0.6 0.8 1.0 1.2 1.4 1.6 Decreased likelihood Increased likelihood
1.8
2.0
2.2
Figure 1. AORs and CIs for meeting physical activity guidelines based on age of youngest dependent child Note: Values are adjusted for parental age, marital status, income, usual hours of work per week and BMI.
with no dependent children in the home (OR⫽0.31, 95% CI⫽0.11, 0.87 (p⫽0.030) for women, OR⫽0.34, 95% CI⫽0.13, 0.93 (p⫽0.037) for men).
Discussion Compared to those without children, daily minutes of MVPA were lower among mothers whose youngest child was aged ⬍6 years and fathers whose youngest child was aged 6 –11 years. In addition, fewer parents than nonparents were meeting the recommended guidelines of 150 minutes per week, in bouts of ⱖ10 minutes. These data illustrate that it is not simply the presence or the number of dependent children in the home but rather the age of the youngest child that is associated with parental levels of MVPA. Importantly, this fınding persisted in mothers and marginally among fathers when the analyses were adjusted for parental age, marital status, income, employment, and BMI. The magnitude of difference in MVPA accumulated per week between parents of children aged ⬍6 years and nonparents (⫺54 minutes per week in women and ⫺40 minutes per week in men; Table 3) is important to consider given that the current physical activity guidelines recommend that adults accumulate 150 minutes per week of MVPA. In other words, parents of young children are missing out on approximately one third of their weekly MVPA needs. To the authors’ knowledge, this is the fırst study to examine the relationship between age of the youngest dependent child in the home and directly measured physical activity levels of mothers and fathers. Consistent with self-report data from the U.S.,8 the accelerometry results presented here indicated that among mothers, when compared to those without children in the home, having a dependent child in the home was signifıcantly associated with physical activity only when the youngest child was aged ⬍6 years. Thus, the hypothesis that parents of young children would be less active than those with older (less-dependent) children was supported. When looking at unadjusted data August 2012
(Table 2), the physical activity levels of women and men were affected by age of the youngest child at home; specifıcally, women whose youngest child was aged ⬍6 years and men whose youngest child was aged 6 –11 years had lower MVPA than those without dependent children. The linear regression models (Table 3), which accounted for known covariates, found that the association between age of youngest child at home remained signifıcant for women, with the youngest age group (⬍6 years) driving this trend. For men, although there was no association overall between age of youngest child and minutes of MVPA, men with a dependent child in the youngest age category had marginally fewer minutes of MVPA than men with no dependent children. The logistic regression examining the likelihood of meeting physical activity guidelines confırmed these fındings with an effect in parents regardless of gender. Overall, the fındings of the current study suggest that having dependent children at home has some influence on parental levels of MVPA; however, the degree of influence appears to be slightly more powerful in women compared to men, especially when the child is very young (⫺54 vs ⫺40 minutes per week or ⫺7.7 vs ⫺5.7 minutes per day). The CHMS does not collect the more-qualitative data that would be needed to offer potential explanations for this gender difference; however, evidence from previous sociology-based research provides arguments about why such a trend may have emerged in our data. For instance, there is some evidence to suggest that women spend more time, and incur greater responsibility, for child care compared to men.38 – 42 In particular, the early years of child-rearing (i.e., when children are aged ⬍6 years) has been identifıed as a time when this may be particularly true,43– 45 a trend that is probably due to factors including— but not limited to—a greater restriction on a mother’s leisure activities during this time, interrupted sleep patterns, and the substantive time required for exclusive breastfeeding.11
174
Adamo et al / Am J Prev Med 2012;43(2):168 –175
These data suggest that adults with young children at home face additional barriers to engaging in regular physical activity when compared to adults without children. Strategies aimed at increasing population levels of physical activity should acknowledge this reality and consider including alternative and creative solutions to help parents overcome these barriers. For example, communities could create physical activity–friendly environments, making it simple and safe for parents to adopt active lifestyles with or without their children (e.g., parks within walking/biking distance to residential areas and daycare centers, well-lit paths or sidewalks, community fıtness centers or programs with child care).
Strengths and Limitations Important strengths of the present study include the use of directly measured MVPA in a large, representative sample of the Canadian population. Additionally, the relationship between dependent children and age of the youngest child in the home and physical activity levels of both mothers and fathers was examined, with the recognition that a child’s age is an important variable with regard to parental responsibility. This secondary analysis is limited by its cross-sectional design, which does not allow for investigation of changes in MVPA over time in individuals as they transition through various stages of parenthood. It would have been useful to further categorize those children aged ⬍6 years, but the CHMS survey design does not allow this. It is possible that some respondents were misclassifıed according to the presence or absence of dependent children in the home; however, 99% of the individuals identifıed as having dependent children also described themselves as parents in a separate question. Although the CHMS provides information on household type, according to the survey data, there is no appropriate indication of the level of additional child care or support and thus could not be controlled for in these analyses. Nevertheless, when examining the data controlling for number of people in the household rather than marital status, the results did not change. The low response rate and accelerometry compliance of the CHMS (42%) is another limitation. The subsample of respondents with at least 4 valid days of accelerometry data was weighted to represent the Canadian population and has been corrected for nonresponse bias. Nonrespondents on the accelerometry fıle tended to be younger, male, and more obese than respondents and therefore may differ in terms of physical activity levels and/or family status.3 Although well-known contributing factors were included, the low R2 of the models tested suggested that some key explanatory variables were missing. How-
ever, the 11 df in the CHMS limited the number of parameters included in each model.
Conclusion Age of the youngest dependent child at home is associated with physical activity levels of Canadian parents. The association was stronger in women compared to men; however, overall it appears that when parents have a child aged ⬍6 years in the home, this is a critical time for parental physical activity engagement. Future research should explore further breakdowns of the ⬍6 years age group (i.e., infants, toddlers, preschoolers) to examine whether the relationship with parental physical activity differs within this early childhood window, and focus on strategies to encourage parents with young children to establish or re-engage in a physically active lifestyle, not only for their own health but to model healthy behavior for their children. The authors thank the Physical Health Measures Division and the Health Analysis Division of Statistics Canada for providing scientifıc and IRB approval of this manuscript. Special thanks to Didier Garriguet for his assistance in guiding the data analyses and for his critical review of the manuscript. KBA and RCC receive research funding from various governmental funding agencies, nongovernmental (not-for-profıt) organizations, and philanthropic groups. No fınancial disclosures were reported by the authors of this paper.
References 1. Tremblay MS, Colley RC, Saunders TJ, Healy GN, Owen N. Physiological and health implications of a sedentary lifestyle. Appl Physiol Nutr Metab 2010;35(6):725– 40. 2. Schoenborn CA, Stommel M. Adherence to the 2008 adult physical activity guidelines and mortality risk. Am J Prev Med 2011;40(5): 514 –21. 3. Colley RC, Garriguet D, Janssen I, Craig CL, Clarke J, Tremblay MS. Physical activity of Canadian adults: accelerometer results from the 2007 to 2009 Canadian Health Measures Survey. Health Rep 2011;22(1):7–14. 4. Guthold R, Ono T, Strong KL, Chatterji S, Morabia A. Worldwide variability in physical inactivity a 51-country survey. Am J Prev Med 2008;34(6):486 –94. 5. Bellows-Riecken KH, Rhodes RE. A birth of inactivity? A review of physical activity and parenthood. Prev Med 2008;46(2):99 –110. 6. Pereira MA, Rifas-Shiman SL, Kleinman KP, Rich-Edwards JW, Peterson KE, Gillman MW. Predictors of change in physical activity during and after pregnancy: Project Viva. Am J Prev Med 2007;32(4):312–9. 7. Gennaro S, Fehder W. Health behaviors in postpartum women. Fam Community Health 2000;22(4):16 –26. 8. Nomaguchi KM, Bianchi SM. Exercise time: gender differences in the effects of marriage, parenthood, and employment. J Marriage Fam 2004;66:413–30.
www.ajpmonline.org
Adamo et al / Am J Prev Med 2012;43(2):168 –175 9. Fahrenwald NL, Atwood JR, Walker SN, Johnson DR, Berg K. A randomized pilot test of “Moms on the Move”: a physical activity intervention for WIC mothers. Ann Behav Med 2004;27(2):82–90. 10. Brown WJ, Mishra G, Lee C, Bauman A. Leisure time physical activity in Australian women: relationship with well being and symptoms. Res Q Exerc Sport 2000;71(3):206 –16. 11. Brown PR, Brown WJ, Miller YD, Hansen V. Perceived constraints and social support for active leisure among mothers with young children. Leisure Sci 2001;23:131– 44. 12. Brown WJ, Trost SG. Life transitions and changing physical activity patterns in young women. Am J Prev Med 2003;25(2):140 –3. 13. Andajani-Sutjahjo S, Ball K, Warren N, Inglis V, Crawford D. Perceived personal, social and environmental barriers to weight maintenance among young women: a community survey. Int J Behav Nutr Phys Act 2004;1(1):15. 14. Verhoef MJ, Love EJ, Rose MS. Women’s social roles and their exercise participation. Women Health 1992;19(4):15–29. 15. Verhoef MJ, Love EJ. Women and exercise participation: the mixed blessings of motherhood. Health Care Women Int 1994;15(4):297–306. 16. Verhoef MJ, Love EJ. Women’s exercise participation: the relevance of social roles compared to non-role-related determinants. Can J Public Health 1992;83(5):367–70. 17. Hamilton K, White KM. Identifying key belief-based targets for promoting regular physical activity among mothers and fathers with young children. J Sci Med Sport 2011;14(2):135– 42. 18. Jones C, Burns S, Howat P, Jancey J, McManus A, Carter O. Playgroups as a setting for nutrition and physical activity interventions for mothers with young children: exploratory qualitative fındings. Health Promot J Austr 2010;21(2):92– 8. 19. Tremblay M, Wolfson M, Gorber SC. Canadian Health Measures Survey: rationale, background and overview. Health Rep 2007;18(S):7–20. 20. Giroux S. Canadian Health Measures Survey: sampling strategy overview. Health Rep 2007;18(S):31– 6. 21. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc 2008;40(1):181– 8. 22. Bryan S, St-Denis M, Wojtas D. Canadian Health Measures Survey: clinic operations and logistics. Health Rep 2007;18(S):53–70. 23. Day B, Langlois R, Tremblay M, Knoppers BM. Canadian Health Measures Survey: ethical, legal and social issues. Health Rep 2007; 18(S):37–51. 24. Heil DP. Predicting activity energy expenditure using the Actical activity monitor. Res Q Exerc Sport 2006;77(1):64 – 80. 25. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of two objective measures of physical activity for children. J Sports Sci 2008;26(14):1557– 65. 26. Puyau MR, Adolph AL, Vohra FA, Zakeri I, Butte NF. Prediction of activity energy expenditure using accelerometers in children. Med Sci Sports Exerc 2004;36(9):1625–31.
August 2012
175
27. Esliger DW, Probert A, Gorber SC, Bryan S, Laviolette M, Tremblay MS. Validity of the Actical accelerometer step-count function. Med Sci Sports Exerc 2007;39(7):1200 – 4. 28. Colley R, Gorber SC, Tremblay MS. Quality control and data reduction procedures for accelerometry-derived measures of physical activity. Health Rep 2010;21(1):63–9. 29. Colley RC, Tremblay MS. Moderate and vigorous physical activity intensity cut-points for the Actical accelerometer. J Sports Sci 2011; 29(8):783–9. 30. National Cancer Institute. Risk factor monitoring and methods: SAS programs for analyzing NHANES 2003–2004 accelerometer data. 8-16-2007. 9-8-2010 31. Tremblay MS, Warburton ER, Janssen I, et al. New Canadian Physical Activity Guidelines. Appl Physiol Nutr Metab 2011;36:36 – 46. 32. Statistics Canada. Canadian Health Measures Survey: Cycle 1. 2007– 2009. Household Questionnaire. 2009. 33. Statistics Canada. Canadian Health Measures Survey Cycle 1 Wave 1 Derived Variable (DV) Specifıcations. 2009. 34. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard defınition for child overweight and obesity worldwide: international survey. BMJ 2000;320(7244):1240 –3. 35. Rao JNK, Wu CFJ, Yue K. Some recent work on resampling methods for complex surveys. Survey Methodol 1992;18(2):209 –17. 36. Rust KF, Rao JN. Variance estimation for complex surveys using replication techniques. Stat Methods Med Res 1996;5(3):283–310. 37. Statistics Canada. Canadian Health Measures Survey (CHMS) Data User guide: Cycle 1. 2010. 38. Bailey WT. A longitudinal study of fathers’ involvement with young children: infancy to age 5 years. J Genet Psychol 1994;155(3):331–9. 39. Craig L. Does father care mean fathers share? A comparison of how mothers and fathers in intact families spend time with children. Gender Soc 2006;20(2):259 – 81. 40. Jackson EL, Henderson KA. Gender-based analysis of leisure constraints. Leisure Sci 1995;17:31–51. 41. McBride BA, Mills G. A comparison of mother and father involvement with their preschool age children. Early Childhood Res Q 1993;8: 457–77. 42. Miller YD, Brown WJ. Determinant of active leisure for women with young children—an “Ethic of Care” prevails. Leisure Sci 2005;27: 405–20. 43. Baxter J. Patterns of change and stability in the gender division of household labour in Australia, 1996 –1997. J Sociol 2002;38(4): 399 – 424. 44. Casper L, Biachi S. Continuity & change in the American family. Thousand Oaks CA: Sage, 2002. 45. Craig L, Bittman M. The effect of children on adults’ time-use: analysis of the incremental time costs of children in Australia. Sydney: Social Policy Research Centre, 2005. SPRC discussion paper no. 143.