Table 3 Results of the included reviews

Barnett et al.

(MA) [34]

Object control movement skill, competency, locomotor skill competency, stability, motor coordination, skill composite

Age, sex, BMI

To identify factors correlated with motor competence

Age is positively correlated with physical activity, while adiposity is negatively. Boys are more skilled than girls for object control and motor coordination

Correlation coefficient for age: 0.37, 95% CI 0.29–0.45. 0.45, 95% CI 0.36–0.53. 0.34, 95% CI 0.29–0.39. Correlation coefficient for sex: 0.28, 95% CI 0.20–0.36

Few studies focused on the same correlate and the same motor skill outcome. Few studies provided correlation coefficients

Additional research that investigates the role of many correlates of motor competence

Oglund et al.

(MA) [31]

Overall PA

Birth weight, motor development, early growth

To explore whether birth weight, early growth and motor development act as determinants of physical activity in children and youth

Birth weight is not an important determinant of physical activity in youth. Available data do not allow firm conclusions whether early growth and motor development act as determinants of physical activity in youth

b = −3.08, 95% CI − 10.20, 4.04

Several of the studies had limitations impacting the quality of the results, but these were not necessarily captured in the standardized quality assessment

More data from high quality birth-cohort studies are warranted before firm conclusions can be made

Olsen et al.

(SLR) [22]

PA behaviour

Health, age, lack of energy, weight

Health, age, lack of energy, weight

To identify factors that influence PA in rural women

Rural women were found to be less active and experience more barriers to PA than urban women; PA determinants among rural women can be categorized according to personal, socioeconomic, and physical environment factors

N.A.

Evaluation of data and analysis was done by one reviewer only; the terms “rural” and “PA” were inconsistently defined among studies; exclusion of articles studying women outside USA

Additional research that clearly defines and consistently applies the terms rural and PA is needed to strengthen knowledge in this area

Babakus et al.

(SLR) [24]

Mixture of PA (total, leisure time, home, work, active commuting, energy expenditure, occupational, intensity, steps or physical inactivity) and sitting time

Sex, ethnicity

To assess what is known about the levels of PA and sedentary time and to contextualize these behaviors among South Asian women with an immigrant background

South Asian women were less active than the other ethnic groups as well as compared to South Asian males

N.A.

No standardized method for quality evaluation; lack of details from some of the included papers; publication and researcher bias possibility; significant heterogeneity among studies

More research should be dedicated to standardize objective PA measurement and to understand how to utilize the resources of the individuals and communities to increase PA levels and overall health of South Asian women

Barnett et al.

(SLR) [33]

PA change across transition to retirement (secondary: leisure time PA, structured exercise, total PA)

Sex, ethnicity

To gain a deeper understanding of qualitative evidence on PA around the transition to retirement

Overall, exercise and leisure time PA increased after the transition to retirement, whereas the findings regarding changes in total PA were inconclusive; men tend to be more active than women

N.A.

Limited number of studies with population from limited socioeconomic diversity; different approaches to assess PA between studies

Future research should address predictors of maintenance of recreational PA after the transition to retirement, the broader benefits of PA, and barriers to PA among retirees from lower occupational groups

De Craemer et al.

(SLR) [18]

Overall PA, MVPA, active transport, during recess

Sex, family risk, preterm birth, birth weight, age, ethnicity, waist circumference, movement skills

To systematically review the correlates of PA, sedentary and eating behaviour in pre-school children 4–6 years old

Little support for biological correlates and PA in general; strong correlation with sex and age and PA; negative association with family obesity risk and positive correlation with gestational age

N.A.

Some limitations regarding the coding of the association of the variables; several studies included wider age range

Strategies should target both boys and girls, all ethnic groups, and parents of both low and high SES; especially on weekdays, should be a focus on maintaining the level of PA and decreasing the level of sedentary behaviour; on weekends, the focus should be on increasing the level of PA

Ridgers et al.

(SLR) [17]

Recess PA

Age, sex, BMI/overweight, body mass, maturation, ethnicity, fitness, special educational needs

To examine the correlates of children’s and adolescent’s PA during school recess periods

Boys are more physically active during recess, no association was found for BMI/central adiposity and grade level

N.A.

The majority are small-sized and cross-sectional studies; MA is difficult to obtain given the limited number of studies and the lack of consistency between them; lack of objective measures

More research is needed concerning correlates of PA in recess period, particularly in adolescents; schools to increase overall facility provision, unfıxed equipment and methods to increase social support, particularly by peers

Stanley et al.

(SLR) [19]

School break time PA cross-sectional studies

Age, sex, motor skills, BMI, ethnicity

To identify the correlates of children’s PA (8–14 years) occurring during the school break time and after school periods

Boys and younger children tend to be more active during break time and after school; BMI in females negatively associated with after school PA, age was negatively associated in school break and after school

N.A.

Small number of studies that vary in methodological aspects; possibility that some studies are missed during the search process; majority of cross-sectional studies

Need for high quality evidence upon which PA promotion in young people can be tailored to specific settings and contexts

Uijtdewillingen et al.

(SLR) [28]

Overall PA

Age, sex, ethnicity, BMI/skinfolds

To summarize and update the existing literature on determinants of PA and sedentary behavior in young people

Moderate evidence of positive relationship between age PA and negative relationship between ethnicity and PA among adolescents

N.A.

Included studies assessed overall PA only; used two databases only; the selected language of publication was English only

To develop long- term interventions more prospective studies with high quality are needed

Craggs et al.

(SLR) [29]

Overall PA

Sex, anthropometry, ethnicity, age, developmental state,

To systematically review the published evidence regarding determinants of change in PA in children and adolescents

Inconclusive associations were reported for large proportion of the determinants examined

N.A.

Possibility of publication bias (included published studies only); heterogeneity in study samples, exposure and outcome measures included in this review; some studies draw data from the same cohorts; semi-quantitative reporting used in the review that limits the classification of the associations

Further research should include objective measures of PA and use previously validated questionnaires to assess the investigated determinants; more high quality research is needed in all age groups, especially in younger children

Dumith et al.

(MA) [30]

Overall PA

Age

To systematically review the international literature regarding PA change in adolescence and quantify that change

The decline in PA during adolescence is consistent finding among studies. In the later studies the decline is more prominent among girls than in boys, although these differences are not significant.

Mean decline (95% CI) 7.0 (5.2–8.8)

Lack of methodological evaluation of the studies included; some studies may be missed in the search process; the original estimate of PA change variability (e.g. standard error) of each study should be preferable to the meta-regression analyses, rather than the used estimate based on the sample size

Interventions that attempt to attenuate the PA decline could be effective; evidence from developing and undeveloped countries is warranted; to improve the validity and comparability of instruments across studies and standardize PA definition in terms of light, moderate and vigorous intensity; aggregation of self-reported and objective PA measures

Koeneman et al.

(SLR) [32]

Overall PA, overall ex, overall PA/ex

Sex, age, ethnicity, chronic conditions/ disease, general physical health, BMI

To systematically review determinants of PA and exercise among healthy older adults

The heterogeneity of the studies allowed only moderate conclusions

N.A.

There may be possibility of publication bias; a wide age range is applied that might have masked some of the differences between subsamples inside that population; they excluded some specific subsamples of the older population; overall low quality of the studies included

The determinants of PA need further study that include the use of objective measures of PA and exercise and valid and reliable measures of determinants

Siddiqi et al.

(SLR) [27]

Overall PA

Physical disability/disease, fatigue, body shape/physical appearance

To systematically review the qualitative literature pertaining to impediments and enablers of PA participation among African Americans

Primary biological determinants influencing PA were fatigue and preexisting chronic diseases

N.A.

Possibility of publication bias; many included studies included only women

To effectively promote PA among African Americans, targeted interventions will need to addressimpediments at multiple levels

Andersen et al.

(MA) [16]

Active vs. inactive (various definitions according studies)

Birth weight

To assess the association between birth weight and LTPA

The association between birth weight and undertaking LTPA is very weak within the normal birth weight range, but both low and high birth weights are associated with a lower probability of undertaking LTPA

OR (95%) CI for: 1.26–1.75 (kg); 1.76–2.25 (kg); 2.26–2.75 (kg); 4.26–4.75 (kg); 4.76–5.25 (kg) respectively

0.67 [0.47, 0.94]; 0.72 [0.59, 0.88]; 0.89 [0.79, 0.99]; 0.92 [0.81, 1.03]; 0.65 [0.50, 0.86]

OR (95%) CI for: 1.26–1.75 (kg); 1.76–2.25 (kg); 2.26–2.75 (kg); 4.26–4.75 (kg); 4.76–5.25 (kg) respectively

0.67 [0.47, 0.94]; 0.72 [0.59, 0.88]; 0.89 [0.79, 0.99]; 0.92 [0.81, 1.03]; 0.65 [0.50, 0.86]

Some information on birth weight and all information on physical activity was self-reported

If PA constitutes a link between birth weight and morbidity and mortality, promotion of PA may be of special importance among subjects of low and high birth weights

Hinkley et al.

(SLR) [21]

Overall PA

Age, sex, family risk (CVD), preterm birth, wheezing/asthma, ethnicity, BMI/weight, movement skills

To study the correlates of pre-school children’s PA behaviors

Boys were significantly more active than girls; age and BMI showed no association with PA

N.A.

Small sample sizes in included studies, as well as small variability in PA; measurement methods may not be sensitive enough; MA is impossible given the variety of effect-sizes

Simultaneous investigation of multiple variables across multiple domains may assist in the identification of potential mediating, moderating, or confounding influences on pre-school children’s PA; the use of larger samples may allow for the detection of small yet significant associations

Tzormpatzakis et al.

(SLR) [23]

Total PA, leisure time PA, occupational PA, exercise and sports, exercise

Sex, age

To evaluate the evidence from research relevant to participation in PA and exercise in Greece

Men exercise more vigorously and more actively than women

N.A.

None of the studies used objective measurements and also they used different self-reported estimates; lack of appropriate use of the terms “exercise” and “PA”

PA promotion should be organized in a systematic way; intervention studies and longitudinal designs to evaluate the long-term effects are suggested; a clear definition of variables is needed; studies should concentrate on the total PA profile of the participants

Van der Horst et al.

(SLR) [20]

Overall PA

Age, sex, ethnicity BMI/skinfolds

To summarize and update the literature on correlates of PA, insufficient PA, and sedentary behavior in young people

Of all potential biological determinants, sex (being male) was positively associated with PA; in children, ethnicity was found to have no effect, for adolescents some of the studies concluded that it was negatively associated with the ethnic minorities, but a final conclusion cannot be made; BMI was found to have no association in both groups

N.A.

Publication bias may be present; possibility of missed studies as a result of the search strategy; the main outcome was overall PA without other classifications; mostly cross-sectional studies included; because of the variability, it was not possible to assess the overall strength of the associations

More prospective studies are needed and more research including children

Coble et al.

(SLR) [25]

Overall PA

Age, sex, body weight, health status

To understand PA behavior of Native Americans

PA levels tend to decrease with age; Native American men are more active than their female counterparts; overall, PA levels of Native Americans tend to be lower than in nonminorities; body weight showed inconsistent results

N.A.

Not all measurement methods used in the studies have been validated; only published papers were included

More studies, especially with longitudinal design are required; there is a need for application of psychological models to understand the PA motivations, as well as culturally appropriate and validated measurement tools

Rhodes et al.

(MA) [26]

Overall PA

Sex, age

To understand the association between major personality traits and PA

The data for the age and sex were inconclusive, given the small number of studies, still the results of those studies suggest that these factors don’t influence personality and PA relationships

N.A.

Research is too limited to draw definitive conclusions about sex, age and culture interactions with personality and physical activity, but preliminary research suggests relative invariance

Future research using multivariate analyses, personality-channeled PA interventions, longitudinal designs and objective PA measurement is recommended