American Journal of Epidemiology
© The Author 2014. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of
Public Health. All rights reserved. For permissions, please e-mail: [email protected].
Vol. 180, No. 11
DOI: 10.1093/aje/kwu254
Advance Access publication:
October 4, 2014
Original Contribution
Early-Life Predictors of Leisure-Time Physical Inactivity in Midadulthood:
Findings From a Prospective British Birth Cohort
Snehal M. Pinto Pereira, Leah Li, and Chris Power*
* Correspondence to Prof. Chris Power, Population, Policy, and Practice, University College London, Institute of Child Health, 30 Guilford Street,
London WC1N 1EH, United Kingdom (e-mail: [email protected]).
Initially submitted July 2, 2014; accepted for publication August 26, 2014.
Much adult physical inactivity research ignores early-life factors from which later influences may originate. In the
1958 British birth cohort (followed from 1958 to 2008), leisure-time inactivity, defined as activity frequency of less
than once a week, was assessed at ages 33, 42, and 50 years (n = 12,776). Early-life factors (at ages 0–16 years)
were categorized into 3 domains (i.e., physical, social, and behavioral). We assessed associations of adult inactivity
1) with factors within domains, 2) with the 3 domains combined, and 3) allowing for adult factors. At each age, approximately 32% of subjects were inactive. When domains were combined, factors associated with inactivity (e.g.,
at age 50 years) were prepubertal stature (5% lower odds per 1–standard deviation higher height), hand control/
coordination problems (14% higher odds per 1-point increase on a 4-point scale), cognition (10% lower odds per
1–standard deviation greater ability), parental divorce (21% higher odds), institutional care (29% higher odds), parental social class at child’s birth (9% higher odds per 1-point reduction on a 4-point scale), minimal parental education (13% higher odds), household amenities (2% higher odds per increase (representing poorer amenities) on a
19-point scale), inactivity (8% higher odds per 1-point reduction in activity on a 4-point scale), low sports aptitude
(13% higher odds), and externalizing behaviors (i.e., conduct problems) (5% higher odds per 1–standard deviation
higher score). Adjustment for adult covariates weakened associations slightly. Factors from early life were associated with adult leisure-time inactivity, allowing for early identification of groups vulnerable to inactivity.
Britain; cohort study; leisure-time physical inactivity; life course
even low activity levels (i.e., the avoidance of inactivity) protect against death (8, 9). Inactivity is a modifiable risk factor,
and positive, sustainable behavior changes could lead to reduced health care costs and burden of disease. Inactivity prevention is, therefore, a key public health goal; understanding
the determinants of inactivity in the general population at different life stages is required to develop interventions aimed at
the prevention of inactivity-related chronic disease.
It is possible that, within the broad range of macro- to
individual-level influences on activity participation (10), earlylife factors might program later health and associated behaviors. Evidence that health behaviors tend to track from early
life to adulthood (11) provides some support for a role of factors in early life. Potentially, such early influences on later inactivity could include physical, social, and behavioral domains
(e.g., poor coordination (physical domain), adversities (social
domain), and smoking (behavioral domain)). Previous work
suggests that social and/or behavioral influences on adolescent
Physical inactivity is highly prevalent with high avoidable
health and economic costs. One-third of adults are estimated
to be inactive (1), and worldwide inactivity is associated with
approximately 5.3 million of 57 million deaths annually (2).
For those who are inactive (defined as activity frequency of
less than once per week), the risk of all-cause mortality was
found to be higher compared with those who are highly active, with higher rates of cardiovascular death of 32% and
43% in men and women, respectively (3, 4). In developed
countries, inactivity accounts for an estimated 1.5%–3.0%
of total direct health care costs (5) (e.g., in 2006–2007, inactivity cost the United Kingdom’s National Health Service
approximately £0.9 billion (6)). Given such health and economic costs, calls have been made for research on inactivity
and its determinants (7), but the literature still focuses largely
on higher activity levels. Although knowledge of influences
on higher activity levels may be desirable, it is also important
to understand the causes of inactivity, given evidence that
1098
Am J Epidemiol. 2014;180(11):1098–1108
Early-Life Predictors of Adult Leisure-Time Inactivity
activity were stronger than physical influences (12, 13), but
whether such effects are maintained into adulthood is unknown. Studies of adult (in)activity patterns have identified
several contemporary influences, including physical factors
(e.g., disabilities (14)), social factors (e.g., income (15)), and
behavioral factors (e.g., smoking (16)). However, many studies of adults tend to be of mixed-aged, nonrepresentative samples (14–17), and importantly, most pay little attention to
early-life factors that may represent the origins from which influences on later inactivity evolve. For example, adversities experienced in childhood could lead to poor emotional
adjustment (18), which in turn can lead to the adoption of hazardous behaviors, such as inactivity.
Using a nationwide general population sample, we aimed to
identify influences on adult leisure-time inactivity. Our specific objectives were to examine associations with inactivity
at ages 33, 42, and 50 years of 1) factors within each early-life
domain (i.e., physical, social, and behavioral), separately and
combined, 2) the 3 early-life domains combined, and 3) all
domains allowing for adult factors.
METHODS
The 1958 British birth cohort is an ongoing longitudinal study
of all persons born in Britain during 1 week in 1958 (n = 17,638)
(19). Information was collected during childhood (at birth and at
ages 7, 11, and 16 years) and adulthood (at ages 23, 33, 42, 45,
and 50 years). Ethical approval was given by the London Multicentre Research Ethics Committee (London, United Kingdom);
informed consent was obtained from participants at various ages
(including at age 50 years). Respondents in midadulthood are
broadly representative of the total surviving cohort (20). The
sample for this study consists of cohort members with at least
1 measurement of inactivity (at age 33, 42, or 50 years), who
were alive and living in Britain at age 50 years (n = 12,776).
Physical inactivity at ages 33, 42, and 50 years was measured by asking participants about leisure-time activity frequency with a list of examples (e.g., swimming, going for
walks). Similar to others (3, 4, 21), we identified low activity
as an activity frequency of less than once per week, hereafter
referred to as inactivity.
Early-life predictors were identified from previous studies of
adolescent activity (12, 13, 22–25) and adolescent predictors
of adult (in)activity (26, 27). Most potential predictors were assessed prospectively from birth to age 16 years and categorized
into 3 domains ( physical, social, and behavioral) (Table 1).
Adult factors at age 33 years included social class (4 categories, as for parental social class at birth; see Table 1), education
(5 categories ranging from none to university degree), measured body mass index (weight (kg)/height (m)2), number of
children in the household (range, 0–8), physically limiting illness, and psychological items of the Malaise Inventory (28).
To facilitate comparisons, we converted all continuous
early-life factors (i.e., parental body mass index, maternal
age, prepubertal stature, cognitive ability, internalizing behaviors (i.e., emotional problems), and externalizing behaviors (i.e.,
conduct problems or antisocial problems) to internally standardized z scores (mean = 0, standard deviation, 1). To estimate
associations of early-life factors with adult inactivity, we conducted analyses in 3 steps. First, we examined associations of
Am J Epidemiol. 2014;180(11):1098–1108
1099
factors within each domain (separately and then combined)
with inactivity at ages 33, 42, and 50 years. We considered
whether the within-person inactivity correlation at these 3
adult ages was an important analytical consideration by modeling the repeated binary outcome using generalized estimating
equations, assuming an unstructured covariance matrix. To establish whether associations with early-life factors differed
by age, we included an interaction term (age × factor). Estimated associations from models of generalized estimating
equations were consistent with those from separate analyses
of inactivity at ages 33, 42, and 50 years (Web Table 1,
available at http://aje.oxfordjournals.org/). For ease of interpretation, all subsequent analyses were performed using
separate analyses for inactivity at ages 33, 42, and 50
years. Hence, within each early-life domain, we examined
univariate associations between each factor and inactivity
at each age using separate logistic regression models. We
then built 3 domain-specific multivariable models by including only factors specific to each domain. Second, we
combined factors from the 3 domains in analyses of inactivity; any factor associated with inactivity (at age 33, 42, or 50
years) in domain-specific models was selected for inclusion
in a model of the 3 domains combined. Third, we examined
whether associations were independent of adult factors by
adjusting for the adult factors described above.
Differences in associations by sex between early-life factors and inactivity were examined using an interaction term
(sex × factor). There was little evidence of effect modification; hence, data from both sexes were combined in analyses.
Missing data ranged from 1% (on internalizing and externalizing behaviors) to 38% (on physical handicap/disabling
condition). To minimize data loss, we imputed missing data
by using multiple imputation chained equations; imputation
models included all model variables, including previously
identified key predictors of missingness (20). Regression
analyses were run across 10 imputed data sets. Imputed results
were broadly similar to those produced when using observed
values; the former are presented.
RESULTS
At each adult age, approximately 32% of participants were
inactive (Table 2); however, there was variation within persons
as they aged. A total of 52% of the participants who were inactive at age 33 years were also inactive at age 42 years, and
48% of the participants who were inactive at age 42 years were
also inactive at age 50 years (data not shown).
Univariate and multivariable associations
In initial analysis, we found no associations with inactivity at
any of the 3 adult ages for low birth weight, birth order, paternal
body mass index, pubertal timing, overweight or obesity, maternal employment, childhood abuse, or sleeping problems.
Thereafter, these factors were excluded from analyses.
Early-life physical factors. Most physical factors were related to inactivity at all 3 adult ages in univariate analyses.
Associations were maintained in multivariable analyses, except for maternal smoking in pregnancy and physical handicap/disabling condition. Shorter prepubertal stature, poorer
1100 Pinto Pereira et al.
Table 1. Potential Early-Life (Birth to Age 16 Years) Predictors of Adult Inactivity Among Participants in the 1958 British Birth Cohort, Followed
From 1958 to 2008
Early-Life Factor
Participant Age at
Data Ascertainment,
years
Data
Ascertainment
Method
Description
Categories/Units
Physical domain
Low birth weight
0
Measured
<2.5 kg
Yes or no
Maternal smoking
during pregnancya
0
Parental report
≥1 Cigarette/day after fourth month of
pregnancy
Yes or no
Maternal
prepregnancy BMI
0
Parental report
Weight (kg)/height (m)2
Continuous (z score)
Age (in years) in 1958
Maternal age
0
Parental report
Birth order
7
Parental report
Continuous (z score)
Prepubertal stature
7
Measured
Height, measured using standardized
protocolsb
Paternal BMI
11
Parental report
Weight (kg)/height (m)2
Continuous (z score)
Pubertal timing
16
Medical
examination
For boys, rating of axillary hair stage
Absent, sparse, intermediate, or
adult
≤11, 12, 13, 14, or ≥15 Years
Cognitive ability
16
Reading and
mathematics
tests (56)
Derived age-standardized score for
each test, converted to a scale of
0–100; average of the 2 tests was
used (if missing, corresponding
average from age 11 or 7 years
was used).
Continuous (z score)
Physical handicap/
disabling condition
7, 16
Parental report
Physical handicap/disabling condition
at age 7 or 16 years
Yes or no
Overweight or obese
7, 11, 16
Measured
International age-appropriate BMI
cutoffs (57), summed across ages
Number of ages (0, 1, 2, or 3) at
which the participant was
overweight or obese
Hand control/
coordination
7, 11, 16
Teacher rating
At each age, measured as no problems Number of ages (0, 1, 2, or 3) at
(score = 0) versus somewhat or
which the participant displayed
such problems
certainly applies (score = 1),
summed across ages.
Parental social class
0
Parental report
Father’s occupational class at child’s
birth (or if missing, at age 7 years),
categorized according to the 1951
Registrar General’s Classification
(59)
4 Categories (professional/
managerial, skilled
nonmanual, skilled manual, or
semiskilled/unskilled and
single-parent households)
Maternal
employment
7
Parental report
Full-time or part-time work since
participant attended school
Yes or no
First, second to fourth, or fifth or
higher
For girls, age at menarche
Continuous (z score)
Social domain
Parental divorce
33
Self-report
Childhood abuse
45
Self-report
Abuse (physical, psychological, or
sexual) by age 16 years (58)
Yes or no
Yes or no
Minimal parental
education
0, 7
Parental report
Both parents had minimal schooling
Yes or no
Childhood neglect
7, 11
Parental and
teacher reports
≥2 of 5 Items at age 7 and/or 11 years Yes or no
(teacher report of child’s physical
appearance, parent report of
involvement with child)
Household
amenities
7, 11, 16
Parental report
Score of 0–18 (higher scores
Availability of bathroom, indoor
indicate more limited access)
lavatory, and hot water; reported as
sole use (score = 0), shared use
(score = 1), or not available (score =
2); a composite score for all 3 ages
was derived.
Institutional care
7, 11, 16
Parental/guardian Ever in institutional care by age
report
16 years
Yes or no
Table continues
Am J Epidemiol. 2014;180(11):1098–1108
Early-Life Predictors of Adult Leisure-Time Inactivity
1101
Table 1. Continued
Early-Life Factor
Participant Age at
Data Ascertainment,
years
Data
Ascertainment
Method
Description
Categories/Units
Behavioral domain
Participants were asked how often they Most active, very active, active, or
least active
played 1) outdoor and 2) indoor
games and sports; 3) went
swimming; and 4) went dancing.
Responses were often (score = 2),
sometimes (score = 1), or never
(score = 0). Scores were summed
across the 4 variables and collapsed
to 4 categories (55).
Physical activity level
16
Self-report
Average or below
average sports
aptitude
16
Self-report
Yes or no
Dissatisfaction with
nearby sporting
facilities
16
Self-report
Yes or no
Internalizing
behaviors
(emotional
problems)
16
Teacher rating
Using 5 items from the 26-item Rutter Continuous (z score)
Behavior Scale, categorized as
applies (score = 2), somewhat
applies (score = 1), or does not apply
(score = 0); scores were summed
across the 5 items (if missing,
measurements from age 11 or 7
years were used) (54).
Externalizing
behaviors
(conduct
problems)
16
Teacher rating
Using 9 items from the Rutter Behavior Continuous (z score)
Scale; scores were summed (if
missing, measurements from age 11
or 7 years were used) (54).
Sociability
16
Teacher rating
5 Categories (sociable to
withdrawn)
Smoking
16
Self-report
Yes or no
Alcohol consumption
in last week
16
Self-report
Sleeping problems
16
Parental report
≥7 Units, 3–6 units, 0–2 units, or
Number and type of drinks (pints of
never drinkers
beer, glasses of wine, or measures
of spirits) consumed in the past
week; amount of alcohol was coded
into standard units and categorized
into 4 groups (53).
None, mild, or severe
Abbreviation: BMI, body mass index.
a
Maternal smoking in pregnancy was included as a physical factor because of its association with offspring’s physical characteristics (such as
neurological deficits) (60).
b
Height was measured with no shoes to the nearest centimeter, half-inch, or inch (1 inch = 2.54 cm), converted to meters, and standardized.
cognition, and poor hand control/coordination were related to
inactivity, with similar magnitudes of association at ages 33,
42, and 50 years (Table 3).
Early-life social factors. Most social factors were related
to inactivity at all 3 adult ages in univariate analyses. Associations were maintained in multivariable analyses, except for
parental divorce, poor household amenities, and institutional
care at some, though not all, ages. Lower parental social class
at child’s birth, minimal parental education, and childhood
neglect were associated with higher odds of inactivity at
ages 33, 42, and 50 years (Table 3).
Early-life behavioral factors. In univariate analyses, most
behavioral factors were related to inactivity at all 3 adult ages.
Associations were maintained in multivariable analyses except for dissatisfaction with nearby sporting facilities, and
at some ages, internalizing behavior, unsociability, and lower
Am J Epidemiol. 2014;180(11):1098–1108
alcohol consumption. Lower activity levels, average or below
average sports aptitude, externalizing behavior, and smoking
were related to inactivity at ages 33, 42, and 50 years, with
varying magnitudes of association (Table 3).
Associations for combined domains and with adjustment
for adult factors
In models that adjusted for factors from all domains simultaneously (Table 4), the odds ratios for inactivity at age
33 years ranged from 0.86 (95% confidence interval: 0.82,
0.91) for 1–standard deviation higher cognition to 1.41
(95% confidence interval: 1.23, 1.61) for average or below
average sports aptitude. Lower activity levels, average or
below average sports aptitude, and poorer cognition were
related to inactivity at all 3 adult ages. Associations with
1102 Pinto Pereira et al.
Table 2. Characteristics of Participants in the 1958 British Birth
Cohort Study, Followed From 1958 to 2008a
Characteristic
No.
%
Mean (SD)
Physical inactivityb
At age 33 years
3,426 31.30
At age 42 years
3,822 34.42
At age 50 years
2,955 30.38
Physical Domain
Prepubertal stature (at age
7 years), cm
Physical handicap/disabling
condition at age 7 or
16 years
122.41 (5.90)
934 11.75
No. of ages at which hand
control/coordination
problems were present
(assessed at ages 7, 11,
and 16 years)
0
4,077 56.81
1
1,970 27.45
2
841 11.72
3
288
4.01
Social Domain
Parental social class at child’s
birth
Professional/managerial
2,210 17.87
Skilled nonmanual
1,206
6,053 48.93
Semiskilled, unskilled, or
single-parent household
2,901 23.45
Minimal parental educationc
6,609 60.32
Childhood neglectc
2,249 19.05
Parental divorce
1,673 15.38
Institutional care (at ages 7, 11,
or 16 years)c
495
DISCUSSION
9.75
Skilled manual
6.38
Household amenities scorec
1.10 (2.71)
Behavioral Domain
Physical activity level at age
16 years
Low
4,475 46.84
Active
1,844 19.30
Very
1,405 14.71
Most
Average or below average sports
aptitude at age 16 yearsc
1,830 19.15
7,004 73.89
Sociability at age 16 yearsc
1 (Sociable)
2,279 23.33
2
3,313 33.91
3
2,664 27.27
4
1,356 13.88
5 (Withdrawn)
158
cognition were similar at all ages, whereas the associations
for lower activity levels, average or below average sports aptitude, and smoking attenuated with age (e.g., the odds ratio
for smoking was 1.23 (95% confidence interval: 1.10, 1.38)
at age 33 years but nonsignificant at age 50 years). Shorter
prepubertal stature, poor hand control/coordination, lower
parental social class at child’s birth, minimal parental education, parental divorce, poorer household amenities, institutional care, externalizing behavior, and unsociable behaviors
were related to inactivity at ages 42 and/or 50 years but not at
age 33 years. Maternal body mass index and age, childhood
neglect, internalizing behaviors, and alcohol consumption
were not associated with adult inactivity when domains were
analyzed simultaneously.
After adjustment for adult covariates, parental education
and household amenities were no longer associated with
adult inactivity (Table 5). For poorer cognition, smoking,
and average or below average sports aptitude, associations
were seen at age 33 years and/or 42 years but not at age 50
years; whereas for shorter prepubertal stature, poor hand control/coordination, lower parental social class at child’s birth,
parental divorce, institutional care, and externalizing and unsociable behaviors, associations remained, although slightly
attenuated, at ages 42 and/or 50 years but not at 33 years. Associations between activity level at age 16 years and inactivity at all 3 adult ages were maintained.
1.62
Abbreviation: SD, standard deviation.
a
Numbers vary because of missing data.
b
Leisure-time physical activity of less than once per week.
c
See Table 1 for a description of this factor.
In a nationwide general population sample followed from
birth, we identified 3 important findings. First, approximately
one-third of the population was inactive during leisure time in
adulthood, and the prevalence of inactivity was similar across
3 ages (i.e., 33, 42, and 50 years). Low activity levels during
adolescence were associated with inactivity decades later in
adulthood, suggesting that inactivity tracks over long periods
of life. Second, by examining an extensive array of early-life
factors, we showed that factors from childhood, spanning exposures and developmental markers (e.g., shorter prepubertal
stature, low parental social class, and externalizing behavior)
were related to adult inactivity. Associations ranged from 14%
lower odds of inactivity for 1–standard deviation higher cognition to 41% higher odds of inactivity for average or below
average sports aptitude. Third, our findings suggest that pathways through which early-life factors influence adult inactivity
vary. For example, adolescent cognition may operate partly
through subsequent adult characteristics, whereas prepubertal stature may influence inactivity independently of other
early-life or adult factors, possibly indicating “programming”
of adult behaviors.
A strength of this study is the repeat, identical measurements that allowed comparison of inactivity over a substantial
period in adulthood. The data enabled examination of a wide
range of prospectively measured predictors from different
early-life domains. To our knowledge, no other study has investigated such an extensive array of early-life factors simultaneously. We were also able to examine whether early-life
factors were associated with adult inactivity independent of
adult factors. However, the study has limitations. Our activity
Am J Epidemiol. 2014;180(11):1098–1108
Am J Epidemiol. 2014;180(11):1098–1108
Table 3. Odds Ratios of Physical Inactivity at 3 Adult Ages for Early-Life Factors in 3 Domains, Estimated From Domain-Specific Univariate and Multivariable Modelsa in the 1958 British Birth
Cohort, Followed From 1958 to 2008
Univariate Models
Early-Life Factor
Age 33 Years
OR
95% CI
Age 42 Years
OR
95% CI
Multivariable Models
Age 50 Years
OR
95% CI
Age 33 Years
OR
95% CI
Age 42 Years
OR
Age 50 Years
95% CI
OR
95% CI
Physical Domain
Prepubertal statureb
0.90
0.86, 0.94 0.89 0.85, 0.93 0.89
0.85, 0.93 0.94
0.90, 0.99 0.93
0.89, 0.98
0.93
0.89, 0.98
Hand control/coordination problemsc,d
1.20
1.14, 1.27 1.20 1.13, 1.28 1.28
1.21, 1.36 1.12
1.05, 1.20 1.11
1.05, 1.17
1.19
1.11, 1.27
Cognition at age 16 yearsb,d
0.80
0.77, 0.83 0.79 0.76, 0.82 0.77
0.74, 0.81 0.84
0.81, 0.88 0.84
0.80, 0.87
0.83
0.79, 0.88
d,e
Physical handicap/disabling condition
1.18 0.997, 1.40 1.28 1.10, 1.49 1.18 1.002, 1.38 1.02
0.84, 1.24 1.11
0.96, 1.28
0.98
0.83, 1.15
Maternal smoking in pregnancyd,e
1.16
1.06, 1.28 1.12 1.03, 1.22 1.12
1.02, 1.23 1.09
0.99, 1.20 1.05
0.96, 1.14
1.05
0.95, 1.15
Maternal BMI in 1958b,f
1.04
0.99, 1.08 1.05 1.01, 1.10 1.06
1.02, 1.11 1.03
0.99, 1.08 1.05
1.01, 1.10
1.06
1.01, 1.11
Maternal age at child’s birthb
0.98
0.94, 1.02 0.95 0.92, 0.99 0.97
0.93, 1.01 0.99
0.94, 1.03 0.96
0.92, 0.997 0.97
0.93, 1.02
Parental social class at child’s birthc,d,g
1.12
1.07, 1.16
1.16 1.11, 1.20 1.23
1.17, 1.29 1.06
1.01, 1.11 1.10
1.05, 1.15
1.13
1.07, 1.19
Minimal parental educationd,e
1.26
1.15, 1.37
1.28 1.18, 1.38 1.42
1.29, 1.57 1.14
1.04, 1.26 1.13
1.03, 1.24
1.22
1.10, 1.35
Childhood neglectd,e
1.32
1.16, 1.50
1.33 1.20, 1.49 1.43
1.28, 1.60 1.20
1.04, 1.38 1.21
1.08, 1.36
1.21
1.08, 1.35
Parental divorce
1.17
1.04, 1.32
1.11 0.99, 1.25 1.34
1.18, 1.51 1.12
0.99, 1.27 1.06
0.94, 1.20
1.25
1.10, 1.41
Institutional cared,e
1.24
1.02, 1.52
1.14 0.91, 1.41 1.73
1.43, 2.08 1.10
0.90, 1.35 1.00
0.80, 1.25
1.44
1.19, 1.74
Household amenitiesc,d
1.03
1.01, 1.04
1.03 1.01, 1.04 1.04
1.03, 1.05 1.02 1.001, 1.03 1.01 0.999, 1.03
1.02
1.01, 1.04
Social Domain
e
Activity level at age 16 yearsc,d,h
1.20
1.15, 1.26 1.19 1.13, 1.24 1.11
1.06, 1.15 1.14
1.08, 1.20 1.15
1.10, 1.21
1.08
1.04, 1.13
Average or below average sports aptitude at age 16 yearsd,e
1.64
1.45, 1.86 1.39 1.24, 1.56 1.28
1.15, 1.43 1.41
1.23, 1.62 1.20
1.06, 1.35
1.17
1.04, 1.32
Dissatisfaction with nearby sporting facilities at age 16 yearsd,e
1.15
1.05, 1.26 1.12 1.01, 1.24 1.00
0.89, 1.12 1.08
0.98, 1.19 1.06
0.96, 1.17
0.96
0.85, 1.08
Internalizing behavior at age 16 yearsb,d
1.14
1.10, 1.19 1.13 1.09, 1.18 1.15
1.10, 1.20 1.07
1.03, 1.12 1.04
0.99, 1.09
1.07
1.03, 1.12
Externalizing behavior at age 16 years
1.09
1.04, 1.14 1.16 1.12, 1.21 1.17
1.12, 1.21 1.05
1.01, 1.10 1.12
1.07, 1.18
1.13
1.08, 1.18
Sociability at age 16 yearsc,i
1.11
1.06, 1.16 1.15 1.10, 1.20 1.11
1.06, 1.17 1.03
0.98, 1.08 1.07
1.02, 1.13
1.04
0.98, 1.10
Smoking at age 16 yearse
1.28
1.15, 1.44 1.34 1.21, 1.49 1.25
1.10, 1.41 1.31
1.17, 1.46 1.28
1.14, 1.45
1.18
1.04, 1.34
Alcohol consumption in last week at age 16 yearsc,d,j
1.08
1.02, 1.13 1.02 0.98, 1.06 1.02
0.96, 1.09 1.08
1.02, 1.14 1.03
0.98, 1.08
1.03
0.97, 1.10
b,d
1103
Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio.
a
Analysis adjusted for sex. Initial analysis found no associations with inactivity at ages 33, 42, and 50 years for several factors (birth weight, birth order, paternal BMI, pubertal timing,
overweight or obesity, maternal employment, childhood abuse, and sleeping problems).
b
Per 1–standard deviation increase.
c
Per increase in scale.
d
See Table 1 for a description of this factor.
e
Those with versus those without the attribute.
f
BMI is weight (kg)/height (m)2.
g
Measured on a scale ranging from high (professional/managerial) to low (semiskilled/unskilled or single-parent household).
h
Measured on a scale ranging from most active to least active.
i
Based on teacher ratings on a scale ranging from sociable to withdrawn.
j
Measured on a scale ranging from 7 or more units/week to never drinkers.
Early-Life Predictors of Adult Leisure-Time Inactivity
Behavioral Domain
1104 Pinto Pereira et al.
Table 4. Odds Ratio of Physical Inactivity at 3 Adult Ages for Early-Life Factors in All Domains (Physical, Social, and
Behavioral) Combined, Estimated From Multivariable Models in the 1958 British Birth Cohort, Followed From 1958 to
2008a
Age at Which Participants Were Inactive
Early-Life Factor
33 Years
OR
95% CI
42 Years
OR
95% CI
50 Years
OR
95% CI
Physical Domain
Prepubertal statureb
0.96
0.91, 1.00
0.94
0.90, 0.99
0.95
0.90, 0.996
Hand control/coordination problemsc,d
1.05
0.98, 1.12
1.06
0.99, 1.12
1.14
1.06, 1.23
Cognition at age 16 yearsb,d
0.86
0.82, 0.91
0.87
0.83, 0.92
0.90
0.85, 0.95
Maternal BMI in 1958b,e
1.02
0.97, 1.07
1.04
0.99993, 1.09
1.05
0.997, 1.10
Maternal age at child’s birthb
0.99
0.95, 1.04
0.96
0.92, 1.01
1.00
0.95, 1.04
Social Domain
Parental social class at child’s birthc,d,f
1.01
0.97, 1.06
1.05
1.002, 1.10
1.09
1.03, 1.15
Minimal parental educationd,g
1.05
0.95, 1.16
1.03
0.94, 1.13
1.13
1.01, 1.25
Childhood neglectd,g
1.03
0.89, 1.19
1.02
0.89, 1.16
1.03
0.91, 1.16
Parental divorce
1.08
0.95, 1.24
0.99
0.88, 1.13
1.21
1.07, 1.37
Institutional cared,g
0.99
0.80, 1.22
0.87
0.69, 1.10
1.29
1.06, 1.57
Household amenitiesc,d
1.01
0.995, 1.03
1.01
0.99, 1.02
1.02
1.002, 1.03
g
Behavioral Domain
c,d,h
Activity level at age 16 years
1.15
1.09, 1.21
1.16
1.11, 1.22
1.08
1.04, 1.13
Average or below average sports aptitude at
age 16 yearsd,g
1.41
1.23, 1.61
1.19
1.05, 1.34
1.13
1.01, 1.28
Internalizing behavior at age 16 yearsb,d
1.03
0.99, 1.08
1.00
0.95, 1.05
1.02
0.97, 1.06
Externalizing behavior at age 16 years
1.00
0.96, 1.06
1.07
1.02, 1.13
1.05
1.01, 1.10
Sociability at age 16 yearsc,i
1.02
0.97, 1.07
1.07
1.02, 1.12
1.03
0.97, 1.08
Smoking at age 16 yearsg
1.23
1.10, 1.38
1.21
1.06, 1.37
1.10
0.97, 1.25
Alcohol consumption in last week at age
16 yearsc,d,j
1.05
0.99, 1.11
0.99
0.94, 1.05
0.99
0.93, 1.05
b,d
Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio.
a
Analysis adjusted for sex and all factors in the table.
b
Per 1–standard deviation increase.
c
Per increase in scale.
d
See Table 1 for a description of this factor.
e
BMI is weight (kg)/height (m)2.
f
Measured on a scale ranging from high (professional/managerial) to low (semiskilled/unskilled or single-parent
household).
g
Those with versus those without the attribute.
h
Measured on a scale ranging from most active to least active.
i
Based on teacher ratings on a scale ranging from sociable to withdrawn.
j
Measured on a scale ranging from 7 or more units/week to never drinkers.
measures are limited to leisure-time activity, rather than total
activity, which would include other domains (e.g., occupation). There is no consistent definition of inactivity that we
could apply here; some studies use failure to meet recommended activity levels (2, 16), whereas others identify the
least active persons in a population (26, 27, 29). Our measure
is based on self-reported infrequent activity (<1 time/week),
which has been used previously (3, 4, 21) and found to be
associated with mortality risk (3, 4). Organization of earlylife factors into 3 domains was subjective, but it afforded a
structured and pragmatic approach; moreover, the domains
are related to theoretical constructs of health and social
capital (30). Because the study is observational, uncontrolled
covariates could account for some of the observed associations. Finally, as with any long-term study, sample attrition
occurred over follow-up, although respondents in midadulthood were broadly representative of the surviving cohort
(20). Maximizing available data, our models included participants with at least 1 measurement of inactivity at age 33, 42,
or 50 years, who were living in Britain at age 50 years. We
avoided sample reductions due to missing information by
using multiple imputation (31).
Despite calls for better understanding of inactivity (7),
most research focuses on higher levels of activity or sedentary
Am J Epidemiol. 2014;180(11):1098–1108
Early-Life Predictors of Adult Leisure-Time Inactivity
1105
Table 5. Odds Ratios of Physical Inactivity at 3 Adult Ages for Early-Life Factors in All Domains (Physical, Social, and
Behavioral) Combined, Estimated From Multivariable Models, With Adjustment for Adult Factorsa in the 1958 British
Birth Cohort, Followed From 1958 to 2008
Age at Which Participants Were Inactive
Early-Life Factor
33 Years
OR
95% CI
42 Years
OR
95% CI
50 Years
OR
95% CI
Physical Domain
Prepubertal statureb
0.96
0.91, 1.01
0.94
0.90, 0.99
0.94
0.89, 0.99
Hand control/coordination problemsc,d
1.03
0.96, 1.10
1.03
0.97, 1.11
1.09
1.01, 1.18
Cognition at age 16 yearsb,d
0.95
0.90, 1.01
0.94
0.88, 0.999
1.01
0.94, 1.08
Maternal BMI in 1958b,e
1.01
0.96, 1.06
1.03
0.99, 1.08
1.02
0.97, 1.07
Maternal age at child’s birthb
1.00
0.95, 1.04
0.96
0.92, 1.01
1.00
0.95, 1.05
Social Domain
Parental social class at child’s birthc,d,f
1.00
0.95, 1.05
1.04
0.99, 1.09
1.07
1.01, 1.13
Minimal parental educationd,g
1.02
0.92, 1.13
1.00
0.91, 1.10
1.07
0.96, 1.19
Childhood neglectd,g
0.99
0.85, 1.14
0.99
0.87, 1.13
0.99
0.88, 1.12
Parental divorce
1.04
0.91, 1.19
0.97
0.85, 1.10
1.19
1.04, 1.35
Institutional cared,g
0.96
0.78, 1.19
0.86
0.68, 1.08
1.30
1.07, 1.59
Household amenitiesc,d
1.01
0.99, 1.02
1.01
0.99, 1.02
1.01
0.999, 1.03
g
Behavioral Domain
c,d,h
Activity level at age 16 years
1.15
1.09, 1.21
1.16
1.11, 1.21
1.09
1.04, 1.14
Average or below average sports aptitude at age
16 yearsd,g
1.41
1.23, 1.62
1.18
1.05, 1.33
1.11
0.98, 1.26
Internalizing behavior at age 16 yearsb,d
1.03
0.99, 1.07
0.99
0.95, 1.05
1.01
0.97, 1.06
Externalizing behavior at age 16 years
0.98
0.93, 1.03
1.06
1.01, 1.11
1.03
0.99, 1.09
Sociability at age 16 yearsc,i
1.02
0.97, 1.07
1.06
1.01, 1.12
1.02
0.97, 1.08
Smoking at age 16 yearsg
1.18
1.06, 1.32
1.17
1.03, 1.33
1.06
0.94, 1.20
Alcohol consumption in last week at age
16 yearsc,d,j
1.05
0.995, 1.12
1.00
0.95, 1.05
1.00
0.93, 1.06
b,d
Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio.
a
All models were adjusted for sex, all factors in the table, and the following variables assessed at age 33 years:
educational level, BMI, number of children in the household, mental health, physically limiting illness, and social class.
b
Per 1–standard deviation increase.
c
Per increase in scale.
d
See Table 1 for a description of this factor.
e
BMI is weight (kg)/height (m)2.
f
Measured on a scale ranging from high (professional/managerial) to low (semiskilled/unskilled or single-parent
household).
g
Those with versus those without the attribute.
h
Measured on a scale ranging from most active to least active.
i
Based on teacher ratings on a scale ranging from sociable to withdrawn.
j
Measured on a scale ranging from 7 or more units/week to never drinkers.
behavior. However, sedentary behavior and physical (in)activity may represent separate constructs, contributing to disease
risk via independent pathways (32). The various definitions
of inactivity make comparison with other studies difficult,
but some consistent themes can be identified. As expected,
we found that lower activity levels and average or below average sports aptitude at age 16 years were associated with higher
odds of inactivity in adulthood, which agrees with the results
of other studies of (in)activity from adolescence to adulthood
(33, 34). Moreover, the attenuating association with age, from
52% higher odds at age 33 years (most vs. least active at age 16
Am J Epidemiol. 2014;180(11):1098–1108
years) to 26% higher odds at age 50 years, is consistent with
observations that tracking is stronger over short periods and
attenuates through the life course (7). Even so, continuities
in inactivity over long periods suggest that efforts to reduce
childhood inactivity may have benefits in later life.
The particularly novel finding of our study is that many
factors across early-life domains were related to adult inactivity. Because most previous studies addressed adult rather than
early-life predictors (14–17), they cannot shed light on how
influences on adult inactivity have evolved. In some instances, our results underline the early-life or developmental
1106 Pinto Pereira et al.
origins of particular predictors of adult inactivity. For example,
educational attainment, which largely reflects cognitive development across earlier life (i.e., before adulthood) has been associated with adult activity (15, 16, 26, 27, 29); whereas our
study highlights earlier abilities specifically, with approximately 12% lower odds of adult inactivity per 1–standard deviation higher cognition. Such findings, in agreement with
those from a Finnish birth cohort (27), imply that early cognitive ability has a long-term influence on adult behavior, partly
because it influences educational level and related social destinations in adulthood. However, in addition to operating in
part via adult characteristics, this long-term association for
cognition and other findings from our study may point to “programming” of behaviors such as inactivity (an idea raised by
others (12)) or at least early-life vulnerabilities.
Notably, in this regard, we found shorter prepubertal stature to be related to adult inactivity. Shorter stature in childhood and adulthood is associated with adult morbidity and
mortality risk (35), but to our knowledge, there is only 1 previous study linking short stature with subsequent activity levels (23). Thus, our finding is novel: a 1–standard deviation
increase in height (6.5 cm) at age 7 years was related to 6%
lower odds of inactivity at age 50 years, independent of other
early-life and adult factors. This finding may reflect growth
tempo, yet no association was seen for pubertal timing, and
maturation status has been shown to have inconsistent associations with adolescent activity (22). It is possible that this
is a chance finding, although associations seen for other
early-life physical factors suggest that some developmental
markers may affect later risk of inactivity. Specifically, hand
control/coordination problems at age 7, 11, or 16 years were
associated with 14% higher odds of inactivity at age 50 years,
whereas problems identified at all 3 ages in childhood were
associated with 48% higher odds of inactivity at age 50
years. This finding is consistent with reported associations
between fewer health problems in childhood and higher
odds of being a very active adult (26). Interestingly, shorter
prepubertal stature and poor hand control/coordination were
related to inactivity at later ages (42 and 50 years) but not at
an earlier age (33 years). A possible explanation could be that
prepubertal stature and hand control/coordination problems
augur earlier onset of poor health in adulthood. We explored
this possibility in analyses of inactivity at age 50 years by including additional adjustment for biomarkers at age 45 years
(i.e., lung function, blood pressure, lipids, and glycated hemoglobin) and self-reported poor health at age 50 years,
and we found that associations were little affected (data not
shown). Thus, our findings for prepubertal stature and hand
control/coordination problems are consistent with early programming of adult inactivity. There are several processes that
may be operating. There may be a lack of suitable activity options in childhood for those with poor hand control/coordination or short stature; were these options available, they might
subsequently mitigate against leisure-time inactivity in adulthood. Alternatively, there may be neurological deficits in these
groups (36) that impede activity participation, as suggested
by our finding that poorer early cognitive ability predicts
adult inactivity. Further investigation of the possible processes involved is warranted. However, if early programming
of adult inactivity is operating, it did not appear to extend to
other early factors. Specifically, we found no association for
low birth weight or birth order, which agrees with results of
studies of adolescent activity (37) though not for birth order
(12, 13). Whereas others have reported that adiposity predicts
subsequent lower activity levels (38–40) and sedentary behavior
(41), we, like others (12), found no such association. Discrepancies could be caused by differences in definitions of inactivity or
age of outcome with studies examining children (38, 40) or
mixed-age adult samples (39), even though the adiposity–
physical inactivity association could change with age (42).
Externalizing and unsociable behaviors in adolescence
were related to inactivity at age 42 years and/or 50 years,
but not at age 33 years. Corresponding findings for greater
adolescent sociability and higher levels of activity in adolescence (24) and adulthood (26) have been observed, and 1
review identified extraversion (of which sociability is a component) as a correlate of activity (43). However, internalizing
behavior in adolescence was shown here to be unrelated to
inactivity in adulthood after allowance for physical and social
factors. Smoking in adolescence has been shown to predict
lower activity levels in white adolescents (25), and here we
found an association with later inactivity that attenuates
with age from 33, to 42, to 50 years. It is possible that, because
risk behaviors such as inactivity and smoking tend to cooccur
(44), the association for adolescent smoking reflects or anticipates future cooccurring behaviors.
Several social factors (i.e., low parental social class, minimal
parental education, parental divorce, poor household amenities, and institutional care) were related to inactivity later in
adulthood, but not at age 33 years. As would be expected for
such “downstream” influences, associations appear to operate
mostly via early-life physical and behavioral development, as
suggested by a comparison of Tables 3 and 4. We might also
expect that associations for early-life factors operate via adult
factors, given associations of low adult social class with inactivity in cross-sectional studies (45, 46). Yet upon adjustment,
attenuation of associations for social factors was modest, suggesting that their influence is independent of adult factors. Per
1-point lower parental social class at child’s birth (on a 4-point
scale), the odds of inactivity at age 50 years were higher by 9%
(i.e., those from unskilled manual backgrounds had 30% higher
odds of inactivity at age 50 years compared with those from
professional/managerial backgrounds). These associations
may be caused by poorer adult health of those from unskilled
manual groups, as previously documented (e.g., low childhood
social class was related to chronic widespread pain in midadulthood, independently of adult social position (47)). In turn,
chronic widespread pain may lead to inactivity. Yet, associations for inactivity at age 50 years were little affected by additional adjustment for adult biomarkers and poor health (data not
shown). Alternatively, a manual occupation may not be conducive to physical activity during leisure time among parents
(48), with transmission of behavior to offspring (i.e., children
were more active if their parents were active early in the children’s lives (13)). Intergenerational transfer of behaviors is known
to occur, with specific parental health-risk behaviors (e.g., not
exercising) predicting similar behavior in their children (49)
and, with tracking, inactivity could continue into adulthood.
Adult leisure-time inactivity appears to be influenced by a
range of factors over the formative early-life period and not
Am J Epidemiol. 2014;180(11):1098–1108
Early-Life Predictors of Adult Leisure-Time Inactivity
solely in adulthood. Our study highlights specific early-life
capabilities, attributes, and environments that were associated with adult inactivity. These findings suggest potential
pathways through which early-life factors could influence
different aspects of adult health. For example, prepubertal
stature, poor physical control/coordination in childhood, and
low social class all have well-known associations with adult
outcomes such as obesity, cardiovascular risk, and death (30,
35, 36). Our results suggest that adult inactivity could be on
the pathway between these early-life factors and adult outcomes, whereas null findings for low birth weight suggest that
adult inactivity is less likely to be on the pathway to adult
health outcomes.
Although we acknowledge broader macro-level influences
on inactivity (10, 50), our findings imply the existence of
groups who are vulnerable in early life for whom intervention
is needed to avoid future leisure-time inactivity in adulthood.
Health 2020, the World Health Organization’s European health
policy framework, indicates that social gradients of health are a
key challenge; the World Health Organization’s website has
identified the importance of promoting activity in socially disadvantaged groups (51). These policies and our findings, taken
together, raise the question of whether there might be strategies
or interventions to broaden activity participation to such groups
(e.g., those with shorter prepubertal stature, lower social class,
or unsociable behaviors) who might otherwise become inactive
adults. For example, strategies could include broadening the
range of activities available. More generally, our findings support arguments for early-life investments to maximize health
and social capital for long-term health gains (52).
ACKNOWLEDGMENTS
Author affiliations: Population, Policy, and Practice, Institute
of Child Health, University College London, London, United
Kingdom (Snehal M. Pinto Pereira, Leah Li, Chris Power).
This work was supported by the Department of Health Policy Research Programme through the Public Health Research
Consortium (PHRC). The views expressed in the publication
are those of the authors and not necessarily those of the Department of Health. Information about the wider program of
the PHRC is available at http://phrc.lshtm.ac.uk. The Great
Ormond Street Hospital/University College London Institute
of Child Health was supported in part by the Department of
Health’s National Institute for Health Research Biomedical
Research Centre.
The funders had no input into the study design; data collection, analysis, and interpretation; the writing of the report; or
the decision to submit the article for publication. Researchers
were independent of influence from study funders.
Conflict of interest: none declared.
REFERENCES
1. Hallal PC, Andersen LB, Bull FC, et al. Global physical activity
levels: surveillance progress, pitfalls, and prospects. Lancet.
2012;380(9838):247–257.
Am J Epidemiol. 2014;180(11):1098–1108
1107
2. Lee IM, Shiroma EJ, Lobelo F, et al. Effect of physical
inactivity on major non-communicable diseases worldwide: an
analysis of burden of disease and life expectancy. Lancet. 2012;
380(9838):219–229.
3. Vatten LJ, Nilsen TI, Romundstad PR, et al. Adiposity and
physical activity as predictors of cardiovascular mortality.
Eur J Cardiovasc Prev Rehabil. 2006;13(6):909–915.
4. Brown RE, Riddell MC, Macpherson AK, et al. The association
between frequency of physical activity and mortality risk across
the adult age span. J Aging Health. 2013;25(5):803–814.
5. Oldridge NB. Economic burden of physical inactivity:
healthcare costs associated with cardiovascular disease.
Eur J Cardiovasc Prev Rehabil. 2008;15(2):130–139.
6. Scarborough P, Bhatnagar P, Wickramasinghe KK, et al. The
economic burden of ill health due to diet, physical inactivity,
smoking, alcohol and obesity in the UK: an update to 2006–07
NHS costs. J Public Health (Oxf ). 2011;33(4):527–535.
7. Telama R. Tracking of physical activity from childhood to
adulthood: a review. Obes Facts. 2009;2(3):187–195.
8. Wen CP, Wai JP, Tsai MK, et al. Minimum amount of physical
activity for reduced mortality and extended life expectancy: a
prospective cohort study. Lancet. 2011;378(9798):1244–1253.
9. Lee IM, Skerrett PJ. Physical activity and all-cause mortality:
what is the dose-response relation? Med Sci Sports Exerc. 2001;
33(6 Suppl):S459–S471.
10. Bauman AE, Reis RS, Sallis JF, et al. Correlates of physical
activity: Why are some people physically active and others not?
Lancet. 2012;380(9838):258–271.
11. Due P, Krolner R, Rasmussen M, et al. Pathways and
mechanisms in adolescence contribute to adult health
inequalities. Scand J Public Health. 2011;39(6 Suppl):62–78.
12. Hallal PC, Wells JC, Reichert FF, et al. Early determinants of
physical activity in adolescence: prospective birth cohort study.
BMJ. 2006;332(7548):1002–1007.
13. Mattocks C, Ness A, Deere K, et al. Early life determinants of
physical activity in 11 to 12 year olds: cohort study. BMJ. 2008;
336(7634):26–29.
14. Xue QL, Bandeen-Roche K, Mielenz TJ, et al. Patterns of
12-year change in physical activity levels in
community-dwelling older women: Can modest levels of
physical activity help older women live longer? Am J
Epidemiol. 2012;176(6):534–543.
15. Barnett TA, Gauvin L, Craig CL, et al. Distinct trajectories of
leisure time physical activity and predictors of trajectory class
membership: a 22 year cohort study. Int J Behav Nutr Phys Act.
2008;5:57.
16. Picavet HS, Wendel-vos GC, Vreeken HL, et al. How stable are
physical activity habits among adults? The Doetinchem Cohort
Study. Med Sci Sports Exerc. 2011;43(1):74–79.
17. Seiluri T, Lahti J, Rahkonen O, et al. Changes in occupational
class differences in leisure-time physical activity: a follow-up
study. Int J Behav Nutr Phys Act. 2011;8:14.
18. Clark C, Caldwell T, Power C, et al. Does the influence of
childhood adversity on psychopathology persist across the
lifecourse? A 45-year prospective epidemiologic study. Ann
Epidemiol. 2010;20(5):385–394.
19. Power C, Elliott J. Cohort profile: 1958 British birth cohort
(National Child Development Study). Int J Epidemiol. 2006;
35(1):34–41.
20. Atherton K, Fuller E, Shepherd P, et al. Loss and representativeness
in a biomedical survey at age 45 years: 1958 British birth cohort.
J Epidemiol Community Health. 2008;62(3):216–223.
21. Aarnio M, Winter T, Kujala U, et al. Associations of health
related behaviour, social relationships, and health status with
persistent physical activity and inactivity: a study of Finnish
adolescent twins. Br J Sports Med. 2002;36(5):360–364.
1108 Pinto Pereira et al.
22. Dumith SC, Gigante DP, Domingues MR, et al. Predictors of
physical activity change during adolescence: a 3.5-year
follow-up. Public Health Nutr. 2012;15(12):2237–2245.
23. Hallal PC, Dumith SC, Ekelund U, et al. Infancy and childhood
growth and physical activity in adolescence: prospective birth
cohort study from Brazil. Int J Behav Nutr Phys Act. 2012;9:82.
24. Vilhjalmsson R, Thorlindsson T. Factors related to physical
activity: a study of adolescents. Soc Sci Med. 1998;47(5):
665–675.
25. Kimm SY, Glynn NW, Kriska AM, et al. Decline in physical
activity in black girls and white girls during adolescence.
N Engl J Med. 2002;347(10):709–715.
26. Kuh DJ, Cooper C. Physical activity at 36 years: patterns and
childhood predictors in a longitudinal study. J Epidemiol
Community Health. 1992;46(2):114–119.
27. Tammelin T, Näyhä S, Laitinen J, et al. Physical activity and
social status in adolescence as predictors of physical inactivity
in adulthood. Prev Med. 2003;37(4):375–381.
28. Rodgers B, Pickles A, Power C, et al. Validity of the Malaise
Inventory in general population samples. Soc Psychiatry
Psychiatr Epidemiol. 1999;34(6):333–341.
29. Kaleta D, Jegier A. Predictors of inactivity in the working-age
population. Int J Occup Med Environ Health. 2007;20(2):
175–182.
30. Power C, Kuh D, Morton S. From developmental origins of adult
disease to life course research on adult disease and aging: insights
from birth cohort studies. Annu Rev Public Health. 2013;34:7–28.
31. Sterne JA, White IR, Carlin JB, et al. Multiple imputation for
missing data in epidemiological and clinical research: potential
and pitfalls. BMJ. 2009;338:b2393.
32. Dunstan DW, Howard B, Healy GN, et al. Too much sitting—a
health hazard. Diabetes Res Clin Pract. 2012;97(3):368–376.
33. Twisk JWR, Kemper HCG, van Mechelen W, et al. Tracking of
risk factors for coronary heart disease over a 14-year period: a
comparison between lifestyle and biologic risk factors with data
from the Amsterdam Growth and Health Study. Am J
Epidemiol. 1997;145(10):888–898.
34. Lefevre J, Philippaerts RM, Delvaux K, et al. Daily physical
activity and physical fitness from adolescence to adulthood: a
longitudinal study. Am J Hum Biol. 2000;12(4):487–497.
35. Gunnell D, Whitley E, Upton MN, et al. Associations of height,
leg length, and lung function with cardiovascular risk factors in
the Midspan Family Study. J Epidemiol Community Health.
2003;57(2):141–146.
36. Osika W, Montgomery SM. Physical control and coordination
in childhood and adult obesity: Longitudinal Birth Cohort
Study. BMJ. 2008;337:a699.
37. Ridgway CL, Brage S, Sharp SJ, et al. Does birth weight
influence physical activity in youth? A combined analysis of
four studies using objectively measured physical activity. PLoS
One. 2011;6(1):e16125.
38. Metcalf BS, Hosking J, Jeffery AN, et al. Fatness leads to inactivity,
but inactivity does not lead to fatness: a longitudinal study in
children (EarlyBird 45). Arch Dis Child. 2011;96(10):942–947.
39. Golubic R, Ekelund U, Wijndaele K, et al. Rate of weight gain
predicts change in physical activity levels: a longitudinal
analysis of the EPIC-Norfolk cohort. Int J Obes (Lond). 2013;
37(3):404–409.
40. Richmond RC, Davey Smith G, Ness AR, et al. Assessing
causality in the association between child adiposity and
physical activity levels: a Mendelian randomization analysis.
PLoS Med. 2014;11(3):e1001618.
41. Ekelund U, Brage S, Besson H, et al. Time spent being
sedentary and weight gain in healthy adults: reverse or
bidirectional causality? Am J Clin Nutr. 2008;88(3):612–617.
42. Parsons TJ, Power C, Manor O. Physical activity, television
viewing and body mass index: a cross-sectional analysis from
childhood to adulthood in the 1958 British cohort. Int J Obes
(Lond). 2005;29(10):1212–1221.
43. Rhodes RE, Smith NE. Personality correlates of physical
activity: a review and meta-analysis. Br J Sports Med. 2006;
40(12):958–965.
44. Spring B, Moller AC, Coons MJ. Multiple health behaviours:
overview and implications. J Public Health (Oxf ). 2012;
34(Suppl 1):i3–i10.
45. Crespo CJ, Ainsworth BE, Keteyian SJ, et al. Prevalence of
physical inactivity and its relation to social class in U.S. adults:
results from the Third National Health and Nutrition
Examination Survey, 1988–1994. Med Sci Sports Exerc. 1999;
31(12):1821–1827.
46. Holme I, Helgeland A, Hjermann I, et al. Physical activity at
work and at leisure in relation to coronary risk factors and social
class. A 4-year mortality follow-up. The Oslo Study. Acta Med
Scand. 1981;209(4):277–283.
47. Power C, Atherton K, Strachan DP, et al. Life-course influences
on health in British adults: effects of socio-economic position in
childhood and adulthood. Int J Epidemiol. 2007;36(3):532–539.
48. Kirk MA, Rhodes RE. Occupation correlates of adults’
participation in leisure-time physical activity: a systematic
review. Am J Prev Med. 2011;40(4):476–485.
49. Wickrama KAS, Conger RD, Wallace LE, et al. The
intergenerational transmission of health-risk behaviors:
adolescent lifestyles and gender moderating effects. J Health
Soc Behav. 1999;40(3):258–272.
50. Bosdriesz JR, Witvliet MI, Visscher TL, et al. The influence of
the macro-environment on physical activity: a multilevel
analysis of 38 countries worldwide. Int J Behav Nutr Phys Act.
2012;9:110.
51. World Health Organization. Physical Activity Promotion in
Socially Disadvantaged Groups: Principles for Action.
Copenhagen, Denmark: World Health Organization Regional
Office for Europe; 2013.
52. Hertzman C, Siddiqi A, Hertzman E, et al. Bucking the
inequality gradient through early child development. BMJ.
2010;340:c468.
53. Jefferis BJMH, Power C, Manor O. Adolescent drinking level
and adult binge drinking in a national birth cohort. Addiction.
2005;100(4):543–549.
54. Power C, Li L, Atherton K, et al. Psychological health
throughout life and adult cortisol patterns at age 45 y.
Psychoneuroendocrinology. 2011;36(1):87–97.
55. Parsons TJ, Power C, Manor O. Longitudinal physical activity
and diet patterns in the 1958 British Birth Cohort. Med Sci
Sports Exerc. 2006;38(3):547–554.
56. Jefferis BJMH, Power C, Hertzman C. Birth weight, childhood
socioeconomic environment, and cognitive development in the
1958 British birth cohort study. BMJ. 2002;325(7359):305.
57. Cole TJ, Bellizzi MC, Flegal KM, et al. Establishing a standard
definition for child overweight and obesity worldwide:
international survey. BMJ. 2000;320(7244):1240.
58. Li L, Denholm R, Power C. Child maltreatment and household
dysfunction: associations with pubertal development in a
British birth cohort. Int J Epidemiol. 2014;43(4):1163–1173.
59. Office of Population Censuses and Surveys. Classification of
Occupations. 1990. London, United Kingdom: Her Majesty’s
Stationery Office.
60. Larsson M, Montgomery SM. Maternal smoking during
pregnancy and physical control and coordination among
offspring. J Epidemiol Community Health. 2011;65(12):
1151–1158.
Am J Epidemiol. 2014;180(11):1098–1108