Behavioral determinants of physical activity across the life course: a “DEterminants of DIet and Physical ACtivity” (DEDIPAC) umbrella systematic literature review

Background Low levels of physical activity (PA) are a global concern and increasing PA engagement is becoming a priority in current public health policies. Despite the large number of studies and reviews available, the evidence regarding the behavioral determinants of PA is still inconclusive. Thus, the aim of this umbrella systematic literature review (SLR) was to summarize the evidence on the behavioral determinants of PA across the life course. Methods A systematic online search was conducted on MEDLINE, ISI Web of Science, Scopus, and SPORTDiscus databases. The search was limited to studies published in English from January, 2004 to April, 2016. SLRs and meta-analyses (MAs) of observational studies that investigated the behavioral determinants of PA were considered eligible. The extracted data were assessed based on the importance of the determinants, the strength of evidence, and the methodological quality. The full protocol is available from PROSPERO (PROSPERO 2014:CRD42015010616). Results Seventeen reviews on 35 behavioral determinants of PA were eligible for this umbrella SLR. Regardless of age, the most investigated determinants were those related with ‘screen use’ and ‘smoking’. For youth, probable positive evidence emerged for ‘previous PA’ and ‘independent mobility and active transport’ among children and adolescents. For the adult population, ‘transition to university’ and ‘pregnancy/having a child’ showed probable negative associations. Conclusions Although the majority of the evidence was limited and most of the determinants were not associated with PA, this umbrella SLR provided a comprehensive overview of the associations between behavioral determinants and PA. Youth should be physically active in the early years and increase active transportation to/from school, independent mobility, and ‘free-range activities’ without adult supervision, whilst adult PA behaviors are mostly influenced by the life events. Finally, more research is needed that incorporates prospective study designs, standardized definitions of PA, objective measurement methods of PA assessment, and the use of interactionist and mediational approaches for the evaluation of different behavioral determinants influencing PA behaviors.


Background
Physical activity (PA) is widely recognized, together with appropriate nutritional and mental status, to be an important component of healthy lifestyle [1]. Since the benefits of being physically active for the prevention and the management of the non-communicable diseases are well known [2], the European Union (EU) is strongly engaged in promoting health-enhancing physical activity [3] for all individuals independently from age and social status. However, in Europe, a large proportion of children, adolescents, adults, and older adults still adopts sedentary lifestyles [4] and, consequently, are more at risk of non-communicable diseases.
Several ecological models for the exploration of the lifestyle choice have been proposed [5][6][7][8], including individual (e.g., biological, psychological, and behavioral aspects), interpersonal (e.g., relationships with parents, relatives, peers, and socio-cultural networks), environmental (e.g., access/availability of tools/services, and proximal/distal built/natural surroundings), and policy (e.g., organizational and governmental aspects) dimensions. Within those dimensions, positive, negative, inconclusive, or no associations might exist between several determinants and PA. However, both determinants and PA present a great diversity in research designs, measurement approaches, populations studied, types of measurement, terminologies, which still make difficult to draw a comprehensive understanding. In general, the term 'determinant' is used to address causal variables also including correlates (i.e., multiple variables intervening in causeeffect relationships), whilst mediators (i.e., variables influencing a cause-effect relationship between variables), moderators (i.e., variables effecting the strength of a relationship between variables), and/or confounders (i.e., variables associated with the outcome that distort the observed relationships) are considered different variables [9,10]. Furthermore, a lack of commonality exists in the PA terminology applied in the studies and different forms of PA are considered, ranging from unstructured daily activities, occupational PA, leisure time PA to structured PA (e.g., exercise, grassroots sports, and competitive sports) and considering the most relevant parameters of PA, such as frequency (e.g., daily, weekly, monthly), duration (e.g., total time of activity, rest intervals), and intensity (e.g., low, moderate, moderate-vigorous, vigorous, maximal efforts).
Within the individual dimension, the determinants related to behaviour are widely investigated. Commonly, behavioral determinants have been already defined as "those that can be eliminated or reduces through lifestyle or behavioral changes" [11]. Thus, among all ages, investigations into the behavioral determinants considered those related with major life changing events (e.g., pregnancy and retirement) [12], habits (e.g., smoking and alcohol consumption) [13], sedentary behaviors (e.g., TV viewing and video/computer game use) [14], and available mobility (e.g., active school transport and independent mobility) [15,16]. They have been studied in a number of primary epidemiological studies, systematic literature reviews (SLRs), and meta-analyses (MAs). However, the diversity in research designs, theoretical and measurement approaches, population groups, determinants investigated, and PA outcomes, across the literature, makes it difficult to understand the evidence and to draw appropriate conclusions on the importance of these determinants in influencing PA behaviors.
Recently, the European Commission endorsed a Joint Programming Initiative to increase research capacity across Member States to engage in a common research agenda on a Healthy Diet for a Healthy Life [17] and the DEterminants of DIet and Physical ACtivity-Knowledge Hub (DEDIPAC-KH) project was created [18]. In order to expand knowledge and to develop new insights and solutions in PA, the Thematic Area 2 of the DEDIPAC-KH project organized and carried out an umbrella systematic literature review (SLR) [19] on all the possible determinants of PA. Overall, seven categories of determinants of PA have been identified: biological, psychological, behavioral, physical, socio-cultural, socio-economic, and policy determinants. Due to the amount of researches available, the DEDIPAC-KH Management Team decided to organize the findings in seven separated umbrella SLRs, each focused on a single category. The DEDIPAC-KH Management Team is conscious that splitting the categories may cause a lost in the analysis of the interactions between those categories of determinants which may share commonalities. However, this strategy was considered necessary for a clear dissemination of insights on the determinants of PA.
The aim of this umbrella SLR was to provide a systematic overview of studies investigating behavioral determinants of PA across the life course by evaluating existing SLRs and MAs. The summary also captures the different study designs and definitions applied for behavioral determinants and PA. Additionally, overall results of the retrieved SLRs and MAs have been evaluated in terms of the importance of the determinants, the strength of the evidence, and the methodological quality.

Materials
The manuscript has been drafted following the PRISMA checklist [20]. A common protocol of the seven umbrella SLRs (biological, psychological, behavioral, physical, socio-cultural, socio-economic, and policy) was registered and is available on PROSPERO (Record ID: CRD42015010616), the international prospective register of systematic reviews [21]. Review title, timescale, team details, methods, and general information were all recorded in the PROSPERO register prior completing data extraction.

Search strategy and eligibility criteria
The present study applied the same search strategy as was used for the other umbrella SLRs of the DEDIPAC-KH. SLRs and MAs investigating the determinants of PA across the life course were systematically searched on MEDLINE, ISI Web of Science, Scopus, and SPORTDiscus. The search was limited to SLRs and MAs published in English, between January 2004 and April 2016. To avoid duplications of the earliest individual studies, reviews published before 2004 were not included. Table 1 shows the  MEDLINE search strategy, and Fig. 1 summarizes the   process of the literature research, common to the subsequent 7 umbrella SLRs. Thus, Table 1 provides the overall list of searched terms, whilst Fig. 1 shows the count of the overall excluded/included reviews, related and not-related with behavioral determinants. SLRs or MAs of observational primary studies on the association between any determinant and PA, exercise (Ex) or sport as the main outcome, were included in the seven umbrella SLRs. The following studies were excluded: i) SLRs and MAs of intervention studies; ii) SLRs and MAs that focused on specific population groups (e.g., chronic diseases); and iii) umbrella SLR's on the same topic (e.g., reviews of SLRs or MAs of epidemiological studies on variables in association with PA). In line with Peel and colleagues [10], behavioral determinants were defined as "those that can be eliminated or reduced through lifestyle or behavioral changes".

Selection process
Relevant articles were independently screened and assessed by two reviewers belonging to the DEDIPAC-KH, who screened the titles, the abstracts, and the full texts. Before the final study inclusion or exclusion, a common decision was reached for each study. Any uncertainty and disagreement was resolved by consulting three further authors of the DEDIPAC-KH (BS, CL, PA).

Data extraction
For each included review, data was extracted on a predefined data extraction form, developed by the DEDIPAC-KH and checked by two authors (AK, PA). In reporting data, authors agreed to use the terms "reviews" as those SLRs and MAs found eligible for the umbrella SLR, and 'primary studies' as those studies included in the eligible SLRs and MAs. Moreover, authors agreed to consider all of the terminologies and forms of PA, including unstructured (i.e., PA linked with daily life) and structured (i.e., Table 1 Search strategy: key words used for the literature research Set Search terms #1 "physical activit*" OR "physical exercise*" OR sport OR "motor activit*" OR "locomotor activit*" OR athletic* OR fitness OR "physical movement*" OR "physical performance*" OR "aerobic exercise*" OR "physical effort*" OR "physical exertion*" #2 determinant OR determinants OR correlate OR correlates OR mediator OR mediators OR moderator OR moderators OR contributor OR contributors OR factor OR factors OR association OR modifier OR modifiers OR confounder OR confounders OR pattern OR patterns OR predictor* #3 demographic* OR motivation OR cognition OR emotion* OR attitude* OR "self-perception" OR "self-confidence" OR "self-efficacy" OR competence OR reward* OR success* OR challenge* OR knowledge OR belief* OR "personal trait*" OR "body image" OR satisfaction OR "time availability" OR "perceived environment" OR family OR peer* OR school* OR leader* OR coach* OR group* OR "climate" OR network* OR employment OR retirement OR "educational level" OR SES OR "socioeconomic status" OR "local identity" OR "national identity" OR value* OR tradition* OR "social expectation*" OR "social trend*" OR "social barriere*" OR "availability of tool*" OR "availability of service*" OR "access to tool*" OR "access to service*" OR neighborhood OR "community route*" OR "school environment" OR "work environment" OR architecture OR urbanization OR transport OR traffic OR "facilit* in public space*" OR advertisement OR "availability of sport club*" OR "availability of fitness center*" OR advocacy OR lobbying OR "corporate social responsibility" OR "physical activity promotion initiative*" OR legislation OR health OR education OR tourism OR environment OR "urban planning" OR transport* OR sport OR sports OR culture OR dance OR theater OR "gender mainstreaming" OR "social inclusion" OR "fiscal measure*" OR program* OR plan OR plans OR communication OR media OR guideline* #4 "systematic literature review" OR "meta-analysis" exercise and sports) independently from their frequency, duration, and intensity. The following information was extracted from each included review: year of publication, type of review (either SLR or MA), number of eligible primary studies included in the behavioral umbrella SLR over the total number of primary studies included in the review, continent/s of the included primary studies, primary study design, overall sample size, age range or mean age, gender proportion, year of publication range of included primary studies, outcome details, type of determinant/s, aim of the review, overall results (qualitative or quantitative), overall recommendations, and limitations as provided by the review itself.

Evaluation of importance of determinants and strength of the evidence
The results retrieved from the eligible primary studies included in the reviews were summarized combining two slightly modified grading scales, previously used by Sleddens et al. [22]. The first scale, grades the importance of the determinants, referring only to the consistency and direction of the associations among the reviews, or the individual primary studies. The second scale, grades the strength of evidence, referring to the study design used among individual primary studies. For its importance, a determinant scored a "-" if all reviews, without exception, reported a negative association between the determinant and the outcome and a "-" if the negative association was found in more than 75% of the reviews or of the original primary studies. The importance of the determinant was scored a "0" if the results were mixed, or more specifically, if the variable was found to be a determinant and/or reported an association (either positive or negative) in 25 to 75% of available reviews or of the primary studies of these reviews, but not in others. Furthermore, the importance of the determinant scored a "+" if a positive association was found in more than 75% of the reviews or of the included primary studies and a "++" if a positive association was found in all reviews, without exception. Despite in the literature the codes "+" and "++" were used in presence of an association, independently from its direction (e.g., positive or negative) [22], in the present SLR these codes denote both the strength and positive direction of the association.
The strength of the evidence was described as 'convincing' (Ce) if it was based on a substantial (n > 10) number of longitudinal observational studies showing associations between the determinant and PA. The strength of the evidence was defined as 'probable' (Pe) if it was based on at least two cohort studies or five cross-control studies showing associations between the determinant and PA. Furthermore, the strength of the evidence was given as 'limited suggestive evidence' (Ls) if it was based mainly on findings from cross-sectional studies showing associations between the determinant and PA. Evidence was labelled as 'limited, no conclusive evidence' (Lns) if the study findings

Quality assessment
The methodological quality of the included reviews was assessed using a modified version of the AMSTAR Checklist [23]. After a consensus between the DEDIPAC-KH partners, the question (number 11) referring to the presence of any conflict of interest was modified to allow for any conflict of interest to be evaluated in any of the included reviews but not in the primary studies included in each review. Two authors belonging to the DEDIPAC-KH independently evaluated the included reviews, using the same methodology of Sleddens et al. [22]. Any uncertainty and disagreement was resolved by consulting three further authors (SB, LC, AP). The eleven criteria were evaluated and scored with 1 when the criterion was applicable to the analyzed review or with 0 when the criterion was not fulfilled, not applicable to the analyzed review, or could not be answered based on the information provided by the review. As a consequence, the total quality score for each included review ranged from 0 to 11. The quality of the review was labelled as weak (score ranging from 0 to 3), moderate (score ranging from 4 to 7), or strong (score ranging from 8 to 11).

SLRs and MAs selection process
As summarized in Fig. 1, the systematic literature search identified 17,941 reviews that were potentially relevant for inclusion in our umbrella SLR. After the removal of duplicates, 15,147 reviews remained for screening. After reading title and abstract, 12,414 and 2,198 reviews, respectively, were excluded because they did not meet the inclusion criteria. Thus, a total of 535 full-text reviews were assessed for eligibility. From these, another 472 reviews were removed because they did not meet the inclusion criteria. Thus, after the full-text reading phase, the number of reviews eligible for the umbrella review was 63. Of these, 46 reviews did not include behavioral determinants of PA. Therefore, the final number of reviews included in the present umbrella SLR on behavioral determinants of PA was 17 (15 SLRs and 2 MAs).

Characteristics of the reviews and quality assessment
The characteristics of the 17 included reviews are summarized in Table 2. All the primary studies included in the individual SLRs or MAs referring to non-behavioral determinants and PA were not considered. Most of the reviews included primary studies from multiple continents, mostly from Europe, North America, and Australia. The predominant study design used among the primary studies was cross-sectional [12,13,15,16,[24][25][26][27][28][29][30][31][32][33]. Nine reviews included prospective and cohort studies, either as the only eligible study design [34,35] or as part of the included studies [12,15,16,25,26,29,32]. In six reviews, it was not possible to retrieve the total population sample size of the included primary studies because it was either not reported or not complete [26,27,29,30,32,33]. In the remaining studies, the total population sample size ranged from 26 [12] to 82,918 [13].

Categorization of the included determinants
During the preliminary phase, 52 behavioral determinants of PA were identified. Among those, similar determinants were identified to form a single determinant. For example, the determinants 'TV viewing' , 'TV viewing/ playing video games' , 'television viewing/sedentary' , 'enjoy television viewing' , 'video/computer game use' , 'computer' , 'screen time' , and 'video games' were merged into the In-depth interviews (n = 2) Focus group (n = 1) Semi-structured interviews            determinant 'screen use'. Differently, other similar determinants were identified to form a sub-group with a specific label. For example, the determinants 'being retired' , 'change in employment status' , and 'moving into an institution' were part of the sub-group 'job-related'. Thus, a final consensus between authors had been achieved for this categorization and the final number of behavioral determinants was 35 (Table 4). Any uncertainty and disagreement was resolved by consulting three further authors of the DEDIPAC-KH (BS, CL, PA).

Findings of the reviews
The findings of the included reviews on the associations between the behavioral determinants and PA, considering different age groups and different types of PA, are summarized in Table 4.

Preschool children
One review specifically assessed the behavioral correlates of PA in preschool children that were aged between 4 and 6 years old [26]. Among the correlates investigated, only those related with 'screen use' were negatively associated with MVPA in all the studies included in the review with a limited suggestive level of evidence (−−, Ls). Inconclusive findings were found regarding 'participation in organized sports' (0, Lns) and 'screen use' (0, Lns) with respect to overall PA and no association was found between 'children have no energy to use active transport' and active transport (0, Lns).

Children
Few determinants were investigated among children aged between 10 and 13 years old in respect to overall PA [34]. 'Previous PA' was the only determinant positively associated with overall PA in more than 75% of the studies included in the review with a probable level of evidence (+, Pe). Conversely, 'participation in sport teams outside school' (0, Ls) and 'alcohol consumption' (0, Ls) were not associated with overall PA in children.
Considering after-school PA, one review investigated the behavioral correlates of PA in children between 8 and 14 years old [30]. Among the correlates investigated, only those related with 'screen use' were negative associated with after-school PA in more than 75% of the studies included in the review with a limited suggestive level of evidence (−, Ls). 'Member of organized activities' (0, Lns) was not associated with after-school PA, whilst for 'use of facilities' the evidence was limited and inconclusive (0, Lns).

Adolescents
Two reviews investigated the behavioral determinants [34] and correlates [32] of PA in adolescents aged between 14 and 18 years [32,34]. 'PE/school sports' [32] and 'previous PA' [34] were both positively associated with overall PA in all the studies included in the review, without exception, both with a limited suggestive level of evidence (++, Ls). 'Vigorous PA' was negatively associated with overall PA in all the studies included in the review, without exception, with a limited suggestive level of evidence (−−, Ls) [34]. No association was found between 'dietary habits' and overall PA (0, Ls) [34].

Adults
Three reviews assessed the behavioral determinants of PA in adults older than 40 years [24,25,27]. 'Baseline activity level' was positively associated with overall PA/ Ex in all the studies included in the review, without exception, with a limited suggestive level of evidence (++, Ls) [27]. Conversely, a negatively association was found between 'language difficulties' [24] and overall PA/Ex in all the studies included in the review, without exception, with a limited suggestive level of evidence (−−, Ls). For 'being retired' (0, Lns) and 'lifelong participation in PA' (0, Ls), the evidence was limited and inconclusive [25].

Older adults
For older adults (≥70 years), only one review [27] investigated the association between 'smoking' and overall PA/Ex suggesting negative associations in all the studies included in the review, without exception, with a limited suggestive level of evidence (−−, Ls).

All ages
Five reviews examined the behavioral determinants of PA across all ages [12,13,24,25,31]. A negative association was found between 'moving into an institution' and overall PA/Ex in all the studies included in the review, without exception, with a limited suggestive level of evidence (−−, Ls) [12]. 'Pregnancy/having a child' and 'transition to university' were negatively associated with overall PA/Ex in more than 75% of the studies included in the review with a probable level of evidence (−, Pe) [12], while 'time limitation' was negatively associated with overall PA/Ex in more than 75% of the studies included in the reviews with a limited suggestive level of evidence (−, Ls) [24,31]. Finally, for 'marital status' (0, Lns) [12,25,31], 'smoking' (0, Lns) [13], 'change in employment status' (0, Ls), and 'experiencing harassment at work, violence, and disaster' (0, Ls) [12] the evidence was limited and inconclusive.

Discussion
This umbrella SLR aimed to provide a comprehensive and systematic overview of behavioral determinants of PA across the life course, evaluating importance, strength of the evidence, and methodological quality. Through a systematic evaluation of the existing SLRs and MAs, an overview of the associations between the considered behavioral determinants and PA is presented to promote effective health enhancing policies and identify gaps for future research strategies. To our knowledge, this is the first umbrella SLR that examined various potential behavioral determinants of PA across the life course.
Whether or not individuals choose to engage in PA behaviors is affected by a number of inter-dependent and multilevel factors. A relevant contribution of the DEDIPAC-KH project was the development of a dynamic framework of determinants of PA, which identified two main themes related to the person and to the society, respectively [36]. Behavioral determinants are part of the individual theme, together with the biological and psychological determinants, whereas the other theme comprises physical (e.g., environmental), socio-cultural, Table 4 Summary of the results of the included reviews: the importance of a determinant and its strength of evidence socio-economic, and policy determinants. Through the systematic analysis of scientific evidence, the present findings expand the knowledge and the understanding of behavioral determinants of PA to prepare the ground for a coherent approach towards PA behaviors research and interventions to favor active lifestyle choices.
Overall, for the majority of determinants, the evidence was limited and inconclusive or there was not an association with the different forms of PA. Furthermore, for several potential determinants it was not possible to ascertain definite conclusions on their association with PA. These findings urge researchers to consider new and innovative approaches when investigating factors influencing PA behaviors.
In relation to the investigated populations, more studies (n = 11) focused on youth (e.g., ≤18 years) with respect to those on individuals older than 18 years (n = 5). The reason for a greater focus on young populations might be due to the fact that being physically active at the youngest stages of age is crucial to maintain high levels of PA at the adulthood [32,37], decreasing the risk factors for noncommunicable diseases [2] and having a favorable effect on public health [37]. Nonetheless, the understanding of the relationships between behavioral determinants and PA is also crucial among older adults as they are more vulnerable and more likely to have decreasing level of PA and increasing levels of sedentary behaviors.
Among all the young population groups, 7 determinants were positively associated and 6 determinants were negatively associated with PA. A probable positive association was evident between 'previous PA' and overall PA in children, with the review showing a moderate quality [34], confirming the importance of PA in childhood. Moreover, a probable positive association was found in three reviews [15,16,28] between 'independent mobility and active transport' and overall PA in children and adolescents, with the reviews showing a moderate quality. Larouche and colleagues [15] argued that it is likely that an active traveler is also more inclined to be active throughout the day. However, other determinants (e.g., low socioeconomic status and lower income households), which were not considered in their review, may influence the relationship between 'independent mobility and active transport' and PA. Furthermore, potential biases due to classification, such as active school transport only once per week and direction of active transport (e.g., only to school and not from school) could also play a role. Other relevant aspects proposed by Schoeppe and colleagues [16] concern the positive associations of active travel to destination other than school, the health benefits for children of being involved in 'free-range' activities, like active travel or play outdoor without adult supervision, and the higher frequency of active travel from school to home rather than from home to school. Other determinants related to 'sport/PA in and out of school' [32,33] and to 'other daily activities' [29] were positively associated with PA but with a limited suggestive level of evidence. The present findings are in line with the determinants related to school physical activity and youth sport participation (e.g., "PA Programs in School", "Availability/Access/Proximity of PA Organized Sport Facilities/Tools", "PA Education (at School/Work)/Knowledge of Effects of PA", "Group Activities (Outdoor/Indoor)", "Involvement in Organized Sport", "Time Spent Outdoor/Playing Spaces") highlighted by the European PA determinants (EU-PAD) framework as having a high priority for research [36]. To improve the strength of evidence, future research adopting longitudinal research designs is strongly needed. Indeed, motor skills in early childhood, school physical education, and youth sports are important aspects for a holistic development of children and adolescents, so an effective interaction between researchers and policy makers should be established to promote of health-enhancing lifestyles in early years, in line with the recommendations of the Expert Group on Health-Enhancing Physical Activity of the European Commission [38].
The determinants 'composite sedentary behavior' , 'internet use' , 'reading' , 'sedentary activities' , and 'vigorous PA' [26,29,33,34] were negatively associated with PA with a limited suggestive level of evidence. Even though mainly based on cross-sectional studies, the negative associations of 'composite sedentary behavior' , 'internet use' , and 'reading' are corroborated by the results of the MA, showing small to moderate negative effect sizes [29]. Regarding 'vigorous PA' , Bruner and colleagues [39] suggested that a decline in PA over the course of the school year, in adolescents, is mainly due to a decrease in level of vigorous intensity activity rather than moderate intensity activity. Despite a clear dose-response specificity of low, moderate, and vigorous PA intensities on cardiovascular risk, metabolic health, osteoporosis, immune function, and mental health [40], the relationship between intensity (but also frequency, duration, and mode of activity) and adherence of PA is controversial [35,41]. To deepen our understanding of the quality of the PA and to provide sound guidelines for PA prescriptions, further research on this area is recommended to consider the intensity of PA on both absolute (e.g., oxygen uptake, oxygen uptake relative to body mass, kcal or kJ per minute, and METs) and relative (e.g., percentages of maximal oxygen uptake, oxygen uptake reserve, heart rate reserve, maximal heart rate, Borg's Rating of Perceived Exertion) terms [42].
The investigation of determinants in adults showed a higher proportion of negative associations (n = 6) with respect to positive associations (n = 1). Regarding the negative associations, a probable level of evidence was found only for 'transition to university' and 'pregnancy/having a child' with respect to overall PA/Ex in all ages >18 years, with the review showing a moderate quality [12]. Furthermore, 'moving into an institution' [12], 'time limitation' [24,31], 'language difficulties' [24], and 'smoking' [27] were negatively associated but with a limited suggestive level of evidence. Conversely, 'baseline activity level' was positively associated with overall PA/Ex in adults ≥40 years but with a limited suggestive level of evidence and the quality of the review was moderate [27]. Although without a consistent convincing evidence among all the determinants, these findings highlight the possible impact of specific life events (i.e., 'transition to university' , 'pregnancy/having a child' , and 'moving into an institution') on the relationship between behavioral determinants and PA and the need for effective interventions and planning to increase healthenhancing PA.
An interactionist approach may help further our understanding of the role played by specific life events. Different life events with shorter-term (e.g., pregnancy) or longer-term effects (e.g., retirement) might exert their effects on PA behaviors separately, in different life phases. Differently, life events, whose shorter-and longer-lasting effects partially overlap in time (e.g., pregnancy and marital status) might generate interactive effects. Moreover, the impact of life events on PA behaviors might be moderated by other behavioral factors. For instance, the extent to which 'retirement' impacts actual PA behavior may be moderated by past PA habits. Moreover, the relationship linking the influence of different behavioral determinants of PA may be mediational in nature, with 'time limitation' probably mediating most of the life event effects on PA behaviors.
However, not only people older than 18 years should represent a target group population for the investigation of those life events and for considering health promotion interventions. Adequate attention should be placed also to those related to 'job' , such as temporary or long-term unemployment, retirement, change in employment status, as well as to those concerning the 'marital status' , such as starting a new close personal relationship, starting to live with someone, marriage, separation, divorce, widowhood, interpersonal loss, which are frequent during adulthood and may interfere and influence PA behaviors.
This umbrella SLR demonstrated that some potential determinants were found to be specific for each age group, whilst other determinants can be considered across all ages. For the latter, 'screen use' and 'smoking' were the most often investigated determinants, with six and four reviews examining the associations with several forms of PA, respectively. Actually, 'screen use' comprises several determinants, such as 'TV viewing' , 'TV viewing/ playing video games' , 'television viewing/sedentary' , 'enjoy television viewing' , 'video/computer game use' , 'computer' , 'screen time' , 'video games' , which have been investigated for their potential associations with PA from preschool children to adolescents. The negative association between 'enjoy television viewing' and MVPA in preschool children emphasized the need for intervention programs to decrease screen related behaviors in the early years age group [26]. A similar negative association was also found in children with respect to after-school time periods (i.e., after-school PA) [30]. Particularly due to the fact that 'TV viewing' may negatively influence PA during time spent at home with high TV accessibility, intervention programs should be encouraged for increasing after-school PA [30]. Future research should adopt longitudinal study designs to investigate the effectiveness of long-term PA promoting strategies in children after school for several components, like fitness, health, cognitive functioning, engagement, motivation, psychological well-being, and social inclusion [43].
No consistent and conclusive association was found between 'screen use' and overall PA when children and adolescents were considered together [14,26,29,32,35]. However, a MA reported small to medium effect sizes between 'TV viewing' and PA and between 'video/computer game use' and PA and also considering possible moderators affecting the strength of their relationship [14]. In particular, the moderator analysis showed that only vigorous PA, but not combined, MVPA, and sports, is negatively associated with 'TV viewing'. These and the present findings corroborate the need of further research on the effects of different variables of PA (e.g., intensity, duration, frequency) on PA behaviors. Moreover, the potential impact of the new technologies on PA behaviors of young people should be considered. Although TV viewing remains the most prominent leisure time sedentary behavior amongst youth, new techniques providing unlimited access to social and internet platforms (e.g., mobile phones and tablet computers) may have both a social and educational role for children and adolescents [29]. Recently, technologically augmented realities have been created to combine real and virtual objects in a real environment [44], also linked to PA (e.g., the smart phone game Pokémon Go). Such advancements in technology-driven PA ("exergaming") [45,46] are attracting the attention of decision makers involved in PA promotion as a potential innovative strategy for reducing physically inactive time, increasing adherence to exercise programs, and promoting enjoyment of PA [47]. However, this still very young line of research does not allow to draw definitive conclusions on effectiveness and long-term effects of exergaming and its role as a determinant of PA behaviors for different populations [48,49].
Smoking is considered one of the biggest public health threats, killing around 6 million people globally every year [50]. It is associated with several diseases, including various types of cancer, coronary heart diseases, stroke,  [25] mostly influenced by the life events, which represent a time constraint for possible involvement in PA. The outcomes of the present umbrella SLR are not only limited to the identification of the behavioral PA determinants, which have a sufficient evidence-base to inform policy development, and of those determinants, which received little or no attention and are therefore in need for further research. The present findings also allow to suggest what type of research designs may be most useful for future research to yield stronger conclusions and generate substantial recommendations to multiple experts converging on the promotion of active lifestyles. Firstly, long-term prospective and longitudinal experimental designs should be employed to tackle the change in relevance of different determinants in critical transition phases from childhood to adulthood and from adulthood to older years. Secondly, an interactionist approach may help understand the joint contribution of different behavioral factors in determining PA behaviors and therefore the need to consider them jointly for effective PA promotion actions. Also, a mediational approach may allow to understand whether different behavioral determinants are directly liked to PA, or indirectly through a mediational chain, in order to identify which is the factor to be primarily targeted to obtain a positive PA behavior change. Finally, future research should address potential clustering and interactions of behavioral and non-behavioral (e.g., psychological, social, environmental) determinants, which may be crucial to inform multisectoral strategies of PA promotion.