Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Socio-economic determinants of physical activity across the life course: A "DEterminants of DIet and Physical ACtivity" (DEDIPAC) umbrella literature review

  • Grainne O’Donoghue ,

    Contributed equally to this work with: Grainne O’Donoghue, Aileen Kennedy, Anna Puggina, Laura Capranica, Stefania Boccia

    Roles Formal analysis, Methodology, Writing – original draft, Writing – review & editing

    grainne.odonoghue@ucd.ie

    Affiliations School of Public Health, Physiotherapy & Sports Science, University College Dublin, Dublin, Ireland, School of Health and Human Performance, Dublin City University, Dublin, Ireland

  • Aileen Kennedy ,

    Contributed equally to this work with: Grainne O’Donoghue, Aileen Kennedy, Anna Puggina, Laura Capranica, Stefania Boccia

    Roles Writing – original draft, Writing – review & editing

    Affiliation School of Health and Human Performance, Dublin City University, Dublin, Ireland

  • Anna Puggina ,

    Contributed equally to this work with: Grainne O’Donoghue, Aileen Kennedy, Anna Puggina, Laura Capranica, Stefania Boccia

    Roles Conceptualization, Data curation, Writing – review & editing

    Affiliation Section of Hygiene—Institute of Public Health; Università Cattolica del Sacro Cuore, L.go F. Vito, Rome, Italy

  • Katina Aleksovska,

    Roles Conceptualization, Data curation, Writing – review & editing

    Affiliation Section of Hygiene—Institute of Public Health; Università Cattolica del Sacro Cuore, L.go F. Vito, Rome, Italy

  • Christoph Buck,

    Roles Conceptualization, Visualization, Writing – review & editing

    Affiliation Leibniz Institute for Prevention Research and Epidemiology, BIPS, Bremen, Germany

  • Con Burns,

    Roles Conceptualization, Writing – review & editing

    Affiliation Dept of Sport, Leisure and Childhood Studies, Cork Institute of Technology, Cork, Ireland

  • Greet Cardon,

    Roles Conceptualization, Funding acquisition, Methodology, Writing – review & editing

    Affiliation Department of Movement and Sports Sciences, Ghent University, Ghent, Belgium

  • Angela Carlin,

    Roles Writing – review & editing

    Affiliation Department of Physical Education and Sports Sciences, University of Limerick, Limerick, Ireland

  • Donatella Ciarapica,

    Roles Conceptualization, Writing – review & editing

    Affiliation Council for Agricultural Research and Economics -Research Centre for Food and Nutrition, Rome, Italy

  • Marco Colotto,

    Roles Conceptualization, Data curation, Writing – review & editing

    Affiliation Section of Hygiene—Institute of Public Health; Università Cattolica del Sacro Cuore, L.go F. Vito, Rome, Italy

  • Giancarlo Condello,

    Roles Conceptualization, Data curation, Writing – review & editing

    Affiliation Department of Movement, Human and Health Sciences, University of Rome Foro Italico, P.za Lauro de Bosis, Rome, Italy

  • Tara Coppinger,

    Roles Writing – review & editing

    Affiliation Dept of Sport, Leisure and Childhood Studies, Cork Institute of Technology, Cork, Ireland

  • Cristina Cortis,

    Roles Conceptualization, Data curation, Investigation, Writing – review & editing

    Affiliation Department of Human Sciences, Society and Health, University of Cassino and Lazio Meridionale, Cassino, Italy

  • Sara D’Haese,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Movement and Sports Sciences, Ghent University, Ghent, Belgium

  • Marieke De Craemer,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Movement and Sports Sciences, Ghent University, Ghent, Belgium

  • Andrea Di Blasio,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Medicine and Aging Sciences, 'G. d'Annunzio' University of Chieti-Pescara, Chieti and Pescara, Italy

  • Sylvia Hansen,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department for Sport and Exercise Sciences, University of Stuttgart, Stuttgart, Germany

  • Licia Iacoviello,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Epidemiology and Prevention. IRCCS Instituto Neurologico Mediterraneo: NEUROMED. Pozzilli, Italy

  • Johann Issartel,

    Roles Writing – review & editing

    Affiliation School of Health and Human Performance, Dublin City University, Dublin, Ireland

  • Pascal Izzicupo,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Medicine and Aging Sciences, 'G. d'Annunzio' University of Chieti-Pescara, Chieti and Pescara, Italy

  • Lina Jaeschke,

    Roles Conceptualization, Writing – review & editing

    Affiliation Molecular Epidemiology Group, Max Delbrueck Center for Molecular Medicine in the Helmholtz Association (MDC), Berlin, Germany

  • Martina Kanning,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Sports Science, University of Konstanz, Konstanz, Germany

  • Fiona Ling,

    Roles Conceptualization, Data curation

    Affiliation Department of Physical Education and Sports Sciences, University of Limerick, Limerick, Ireland

  • Agnes Luzak,

    Roles Conceptualization, Writing – review & editing

    Affiliation Institute of Epidemiology I, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Germany

  • Giorgio Napolitano,

    Roles Conceptualization, Data curation, Writing – review & editing

    Affiliation Department of Medicine and Aging Sciences, 'G. d'Annunzio' University of Chieti-Pescara, Chieti and Pescara, Italy

  • Julie-Anne Nazare,

    Roles Conceptualization, Data curation, Writing – review & editing

    Affiliation CarMeN Laboratory, INSERM U1060, Lyon 1 University, CRNH-Rhône-Alpes, CENS, Lyon, France

  • Camille Perchoux,

    Roles Data curation, Writing – review & editing

    Affiliation CarMeN Laboratory, INSERM U1060, Lyon 1 University, CRNH-Rhône-Alpes, CENS, Lyon, France

  • Caterina Pesce,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Movement, Human and Health Sciences, University of Rome Foro Italico, P.za Lauro de Bosis, Rome, Italy

  • Tobias Pischon,

    Roles Conceptualization, Writing – review & editing

    Affiliation Molecular Epidemiology Group, Max Delbrueck Center for Molecular Medicine in the Helmholtz Association (MDC), Berlin, Germany

  • Angela Polito,

    Roles Conceptualization, Writing – review & editing

    Affiliation Council for Agricultural Research and Economics -Research Centre for Food and Nutrition, Rome, Italy

  • Alessandra Sannella,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Human Sciences, Society and Health, University of Cassino and Lazio Meridionale, Cassino, Italy

  • Holger Schulz,

    Roles Conceptualization, Writing – review & editing

    Affiliation Institute of Epidemiology I, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Germany

  • Chantal Simon,

    Roles Conceptualization, Funding acquisition, Writing – review & editing

    Affiliation CarMeN Laboratory, INSERM U1060, Lyon 1 University, CRNH-Rhône-Alpes, CENS, Lyon, France

  • Rhoda Sohun,

    Roles Writing – review & editing

    Affiliation Department of Physical Education and Sports Sciences, University of Limerick, Limerick, Ireland

  • Astrid Steinbrecher,

    Roles Conceptualization, Writing – review & editing

    Affiliation Molecular Epidemiology Group, Max Delbrueck Center for Molecular Medicine in the Helmholtz Association (MDC), Berlin, Germany

  • Wolfgang Schlicht,

    Roles Conceptualization, Funding acquisition, Writing – review & editing

    Affiliation Department for Sport and Exercise Sciences, University of Stuttgart, Stuttgart, Germany

  • Ciaran MacDonncha,

    Roles Conceptualization, Data curation, Funding acquisition, Writing – review & editing

    Affiliation Department of Physical Education and Sports Sciences, University of Limerick, Limerick, Ireland

  • Laura Capranica ,

    Contributed equally to this work with: Grainne O’Donoghue, Aileen Kennedy, Anna Puggina, Laura Capranica, Stefania Boccia

    Roles Conceptualization, Data curation, Funding acquisition, Writing – review & editing

    Affiliation Department of Movement, Human and Health Sciences, University of Rome Foro Italico, P.za Lauro de Bosis, Rome, Italy

  •  [ ... ],
  • Stefania Boccia

    Contributed equally to this work with: Grainne O’Donoghue, Aileen Kennedy, Anna Puggina, Laura Capranica, Stefania Boccia

    Roles Conceptualization, Data curation, Funding acquisition, Writing – review & editing

    Affiliation Section of Hygiene—Institute of Public Health; Università Cattolica del Sacro Cuore, L.go F. Vito, Rome, Italy

  • [ view all ]
  • [ view less ]

Socio-economic determinants of physical activity across the life course: A "DEterminants of DIet and Physical ACtivity" (DEDIPAC) umbrella literature review

  • Grainne O’Donoghue, 
  • Aileen Kennedy, 
  • Anna Puggina, 
  • Katina Aleksovska, 
  • Christoph Buck, 
  • Con Burns, 
  • Greet Cardon, 
  • Angela Carlin, 
  • Donatella Ciarapica, 
  • Marco Colotto
PLOS
x

Abstract

Background

To date, the scientific literature on socioeconomic correlates and determinants of physical activity behaviours has been dispersed throughout a number of systematic reviews, often focusing on one factor (e.g. education or parental income) in one specific age group (e.g. pre-school children or adults). The aim of this umbrella review is to provide a comprehensive and systematic overview of the scientific literature from previously conducted research by summarising and synthesising the importance and strength of the evidence related to socioeconomic correlates and determinants of PA behaviours across the life course.

Methods

Medline, Embase, ISI Web of Science, Scopus and SPORTDiscus were searched for systematic literature reviews and meta-analyses of observational studies investigating the association between socioeconomic determinants of PA and PA itself (from January 2004 to September 2017). Data extraction evaluated the importance of determinants, strength of evidence, and methodological quality of the selected papers. The full protocol is available from PROSPERO (PROSPERO2014:CRD42015010616).

Results

Nineteen reviews were included. Moderate methodological quality emerged. For adults, convincing evidence supports a relationship between PA and socioeconomic status (SES), especially in relation to leisure time (positive relationship) and occupational PA (negative relationship). Conversely, no association between PA and SES or parental SES was found for pre-school, school-aged children and adolescents.

Conclusions

Available evidence on the socioeconomic determinants of PA behaviour across the life course is probable (shows fairly consistent associations) at best. While some evidence is available for adults, less was available for youth. This is mainly due to a limited quantity of primary studies, weak research designs and lack of accuracy in the PA and SES assessment methods employed. Further PA domain specific studies using longitudinal design and clear measures of SES and PA assessment are required.

Background

The benefits of being physically active are well acknowledged in the primary and secondary prevention of many conditions such as cardiovascular disease, hypertension, type 2 diabetes, obesity, osteoporosis, anxiety and depression [1]. Compared to individuals who engage in regular moderate-intense physical activity (PA) for at least 150 minutes per week or 75 minutes per week of vigorous intensity, insufficiently physically active individuals have a 20–30% increased risk of all-cause mortality [1]. Evidence gathered from longitudinal studies undertaken in western countries show age-related changes in PA behaviours; with a steep decrease in PA levels occurring during adolescence [2,3]. Considering that PA behaviours established in youth tend to track into adulthood, PA promotion in youth has been deemed a priority in order to facilitate a carryover of active lifestyles into adulthood and to warrant a lifelong protection from other risk factors [4]. Cross-sectional global estimates of PA report 25% of adults between 18–65 years, 55% of older adults (> 65 years) and 81% of school-aged youth (11–17 years) are insufficiently active [2]. Therefore, it is crucial to identify factors having a potential effect on PA behaviours.

To establish experimental evidence related to PA behaviours’ determinants, there is a need of a clear understanding of associations or predictive relationships between variables [5]. In general, the term “determinant” is used to address causal variables, including correlates (i.e., multiple variables intervening in cause-effect relationships), mediators (i.e., variables influencing in 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 distorting the observed relationships) [6]. Furthermore, definitions of PA also present terms that lead to inconsistencies resulting in different interpretations and outcomes [5]. In fact, PA differs in term of typology (i.e. unstructured daily activities, such as occupational PA (OPA) and leisure time PA (LTPA) and structured PA such as physical exercise, grassroots sports, and competitive sports, frequency (e.g. daily, weekly, monthly), duration (e.g. activity/rest patterns), and intensity (e.g. low, moderate, moderate-vigorous, vigorous, maximal efforts). Recently, the strong links between health enhancing PA, grassroots sports, and competitive sports for the development, transfer and/or implementation of active lifestyles have been recognised [7]. Therefore, when reviewing the literature on determinants of PA, a wide perspective should be considered.

Socioeconomic status (SES) or its derivatives (e.g. income, education and occupation) has been recognised as an important determinant of health and wellbeing because it influences people's attitudes, experiences and exposure to several health risk factors across the life course [8]. In particular, children who grow up in lower SES households have a higher risk of unhealthier lifestyles, cardiovascular disease [9] and all-cause mortality [10,11] than children who live in higher SES households. Stringhini and colleagues [12] also report that the combination of potentially modifiable unhealthy behavioural factors such as physical inactivity, smoking and poor diet could explain between 12% to 54% of the SES differences in mortality; with the relationship between SES and smoking, alcohol consumption, and poor diet being more consistent than that between SES and PA [13]. This consistency is likely to be due to the multi-dimensional nature of PA and SES that pose considerable methodological measurement problems [13]. Few studies have attempted to categorise PA by domain (e.g. LTPA; occupational PA (OPA)) while at the same time ascertaining its association with SES across the life course. Researchers may need to focus on single components and use them as proxies for overall PA and SES, or use composite scores of SES (e.g., deprivation index or socioeconomic position) and PA (e.g. total PA) in order to reduce methodological inconsistencies. Furthermore, identification of barriers to PA related to SES factors, such as educational background, employment and/ or available income should play a key role in the development and implementation of future interventions and policy [9,10].

To date, the scientific literature on socioeconomic correlates and determinants of PA behaviours has been dispersed throughout a number of systematic reviews, often focusing on one factor (e.g. education or parental income) in one specific age group (e.g. pre-school children or adults). The purpose of the present study is to provide a comprehensive and systematic overview of the scientific literature from previously conducted research to assess the importance and strength of the evidence related to socioeconomic correlates and determinants of PA behaviours across the life course, through an umbrella SLR of systematic literature reviews (SLRs) and meta-analyses (MAs). The principle reason for choosing the umbrella review methodology is that it allows ready assessment of whether review authors addressing similar review questions independently observe similar results and arrive at generally similar conclusions. The aim of an umbrella review is not to repeat the searches, assessment of study eligibility, assessment of risk of bias or meta-analyses from the included reviews, but rather to provide an overall picture of findings for particular questions or phenomenon [14]. An umbrella review's most characteristic feature is that this type of evidence synthesis only considers for inclusion the highest level of evidence, namely other systematic reviews and meta-analyses. The wide picture obtainable from the conduct of an umbrella review is ideal to highlight whether the evidence base around a topic is consistent or contradictory, and to explore the reasons for the findings [15].

Methods

The current research was developed within the Thematic Area 2 of the DEterminants of Diet and Physical ACtivity Knowledge Hub (DEDIPAC-KH). To systematise and update the current evidence-base on the determinants and correlates of PA behaviours across the life course, a common protocol for the DEDIPAC-HK umbrella SLR was developed and registered in PROSPERO (PROSPERO 2014: CRD42015010616) [16]. This manuscript is drafted following the PRISMA checklist [S1 Checklist].

Search strategy and eligibility criteria

An online search was conducted using the following electronic databases; MEDLINE, ISI Web of Science, Scopus and SPORTDiscus. SLRs and MAs that focused on the association between any determinant of PA or exercise or sport as main outcomes were considered. The following exclusion was implemented: i) SLRs and MAs of intervention studies; ii) SLRs and MAs focusing on specific clinical population groups (e.g., chronic disease); and iii) umbrella SLRs on the same topic (e.g. reviews of SLRs or MAs of epidemiological studies on variables associated with PA). The search was limited to publications in English during the period from January 2004 to September 2017. Table 1 shows the MEDLINE search strategy that was also used as template for the search strategies in the other databases. Throughout this work, the term determinant will be utilised to address any variable affecting PA independently from their role; whereas the term PA will include non-structured and structured activities independently from their frequency, duration, and intensity.

Selection process

The selection process consisted of three phases. In the initial phase, relevant articles were independently screened and assessed by two reviewers belonging to the DEDIPAC KH, who screened the yielded articles based on title. In the case of doubt, the articles were included in the abstract review phase. In the second phase, all articles selected from the initial phase had their abstract reviewed and assessed by two independent reviewers of the DEDIPAC- Knowledge Hub (KH) research team (KA and AP). Any uncertainty and disagreement was resolved by consulting three further authors (SB, LC, AP). In the final phase, AK and GO’D fully reviewed the remaining articles. In this phase, any disagreement between reviewers was resolved by discussion within the DEDIPAC-KH research team. In considering the specific focus of the present umbrella SLR, studies that focused on non-socioeconomic determinants of PA behaviours were not considered.

Data extraction

A fourteen item standardised pre-piloted data extraction form was used to extract data from the included studies under the following headings: year of publication, type of review (either SLRs or MAs), number of eligible primary studies included in the represented umbrella SLR over the total number of studies included in the review; continent/s of the included studies, primary study design, overall sample size, age range or mean age, gender proportion, year range of included studies; outcome details, type of determinant/correlate, aim of the review; overall results (qualitative or quantitative), overall recommendations and limitations as provided by the review itself. Furthermore, the importance and strength of evidence of a determinant included in a particular review was evaluated by applying a modified version of the criteria adopted by the World Cancer Research Fund [17], further adapted by Sleddens et al [18].

Outcome measures

The socioeconomic factors used to classify socioeconomic status in this review were based on those used by Beenackers et al (2012) in their review on socioeconomic inequalities in physical activity among European adults. The factors considered were income, referring to an individual’s or household income. Education, referring to the highest attained level of education (e.g. university education). Occupation, such as blue or white collar workers and other social economic predictor indicators such as home ownership and the ability to pay fees and / or purchase equipment required to engage in structured physical activity [19].

In terms of PA categories, total PA, moderate-vigorous PA (MVPA), LTPA and OPA were categorised where possible. In relation to age groups, the categories included: pre-school children (2–5 years), children (6–12 years), adolescents (13–18 years), adults (19–65 years) and older adults (> 65 years). Furthermore, break-time/recess time PA and after-school PA were also considered for pre-school children and children, while uptake of and adherence to exercise referral schemes were considered for adults.

Risk of bias assessment

Assessing risk of bias of the reviews is essential because it impacts on the extent to which conclusions can be drawn from the evidence. A modified AMSTAR checklist [20] was used to perform the quality assessment of the included reviews. Two reviewers belonging to the DEDIPAC KH independently evaluated the included reviews. 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 met or with 0 when the criterion was not met.

As a consequence, the total quality score for each included SLR ranged from 0 to 11 quality scores. The quality of the SLR was labelled as weak (score range: 0–3), moderate (score range: 4–7), or strong (score range: 8–11).

Data synthesis

A narrative synthesis of the findings of this umbrella SLR is provided, structured around a modified version of the criteria for grading evidence and the data extraction employed by Sleddens and colleagues [18,21]. Results retrieved from the eligible primary studies included in the reviews were summarised combining two grading scales. The first, grades the importance of the determinants, referring to the consistency of the associations among the reviews, or the individual primary studies. The second, grades the strength of evidence, referring to the study design used among individual primary studies.

According to Sleddens [18, 21], the codes + and ++ were used if there is an association (no matter of positive or negative). This was modified for the present review to report both the association and the direction of the association (Table 2). The importance was scored (—) if all reviews, without exception, found a negative association between the determinant and the outcome. A (-) score was given if the negative association was found in 75% of the included 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 less than 75% of available reviews or of the primary studies of these reviews. The importance of a determinant was scored as (00) if all reviews, without exception, reported a null association. The importance of the determinant scored (+) if a positive association was found in 75% of the reviews or of the included primary studies and (++), if a positive association was found in all reviews, without exception.

The strength of the evidence was also summarized using the criteria adopted by Sleddens et al [18,21]. The strength of evidence was described as ‘convincing’ (Ce) if it was based on a substantial number of longitudinal observational studies, with sufficient size and duration, and showing consistent 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 fairly consistent associations between the determinant and PA. The strength of the evidence was given as ‘limited, suggestive evidence’ (Ls) if it was based mainly on findings from cross-sectional studies showing fairly consistent associations between the determinant and PA, and as ‘limited, non- conclusive evidence’ (Lnc) if study findings were suggestive, but insufficient to provide an association between the determinant and PA.

The strength of the evidence (Table 3) was described as “convincing” if based on high quality studies showing consistent associations and having longitudinal design with sufficient size and duration, whereas “probable” strength of evidence was given to determinants showing fairly consistent associations based upon at least one cohort study. In the second case, shortcomings were possible either in terms of the consistency of the results or other aspects such as limited duration of the studies, small sample sizes or inadequate follow up. Furthermore, “limited suggestive evidence” was given to determinants for which there was insufficient number of longitudinal studies and “limited, no conclusive” evidence when the evidence for the associations between a determinant and the outcome were based solely on studies of cross-sectional design.

Results

The process for undergoing the literature search and screening, including numbers of reviews excluded and reasons for exclusion is illustrated in Fig 1. In summary, the electronic search yielded 11754 records, of which 689 duplicates were removed. Of the remaining 11065 records, 10998 were excluded throughout the screening process. After the full-text reading phase, the final number of studies eligible for the umbrella review was 67. Of these, 48 did not concern socioeconomic factors. Therefore, the final number of reviews included in this umbrella SLR on socioeconomic determinants of PA was 19 SLRs. No MAs were found to be eligible.

Review characteristics

Table 4 provides detailed characteristics of the 19 systematic reviews. Of the 19 included, only two reported socioeconomic factors in all their primary studies (n = 164) [19,22]. Most of the reviews included primary studies from multiple continents, except two reviews; one that considered only cohort studies conducted in Europe [19] and the other only Chinese cohorts [23]. In general, the majority of primary studies resulted from Europe (n = 206), followed by North America, (n = 96) and Oceania (n = 14). Cross-sectional (n = 117) was the predominant study design among the 19 SLRs, though eight reported longitudinal (prospective) studies. One SLR solely focused exclusively on qualitative studies [24], whereas another one used mixed (e.g. qualitative and cross sectional) methodologies [25]. In terms of sample size, a considerable variation (from 25 to 29135 participants) emerged. Regrettably, it was not possible to retrieve the total population sample size from one SLR [26].

Eleven SLRs considered primary studies that included only the young population, with pre-school children being the focus of two SLRs [27,28], children of two [29,30], adolescents of two [31,32], and a combination of children and adolescents of five [23,26,33,34,35]. Eight SLRs focused only on adults [19,22,24,25,36,37,38,39], with two also considering older adults [24,37].

Quality assessment

Table 5 summarises the quality assessment of the19 SLRs. The majority (n = 15) of the SLRs showed a moderate quality (range: 4–7 points), three [19,31,35] were evaluated as weak (range: 0–3 points) and only one [37] as strong (8-11pts).

thumbnail
Table 5. Quality assessment of the included reviews using the AMSTAR checklist.

https://doi.org/10.1371/journal.pone.0190737.t005

Major findings

Table 6 summarises the findings of the SLRs on the associations between the socioeconomic determinants and PA. The most frequently studied correlates were SES (n = 12) [19,22,23,26,27,28,30,31,32,33,34,38], payment of fees or equipment (n = 5) [24,25,29,35,36], education level (n = 6) [22,23,25,28,33,39], and individual or household income level (n = 6) [22,25,33,36,37,39]. In addition, neighbourhood income (n = 2) [36,37], employment levels (n = 2) [36,37] and number of working hours of an individual or parent (n = 2) [25,33] were also examined. Finally, two SLRs [33,33] also considered parental occupational status and home ownership determinants of PA in children. In combining the resulting importance of the correlates or determinants (e.g., the number of included primary studies or the number of SLRs that showed an association between socioeconomic determinant and PA and the strength of evidence (e.g., the study design of the primary studies) [17]. Table 7 summarises the final judgments on the associations between the investigated socioeconomic determinants and PA. The following paragraphs refer to age-related findings for individuals <18 years (e.g., children and adolescents) and >18 years of age (e.g., adults).

thumbnail
Table 7. Summary of the results of the included reviews: The importance of a determinant and its strength of evidence.

https://doi.org/10.1371/journal.pone.0190737.t007

Children and adolescents (3–18 years)

In preschool children, overall SES was consistently found to be unrelated to overall PA levels [23,27,28], with a probable level of evidence (00, Pe). Similarly, it was consistently found to be unrelated to moderate and vigorous activity levels (MVPA) [27] in more than 75% of the primary studies assessing pre-school children and this category of determinants, resulting in a probable level of evidence (0, Pe).

Among reviews that combined children and adolescents, SES was also found not to be significantly related to overall PA levels [33,34], with 75% of the primary studies reporting no relationship to SES, thus the evidence was rated as probable (0, Pe). However, in SLRs that addressed children and adolescents separately, between 25–75% of SLRs found some evidence (mixed findings) that SES influences overall PA levels. Considering that some reviews found SES to be a correlate of overall PA, while others did not [26,31,34], the evidence is (0, Pe).

In terms of the specific domains of PA and SES considered in this age group, results were mixed. Limited and inconclusive evidence (+, Lns) showed school break time PA to be associated with SES in children and adolescents [32]. In contrast, SES was not associated with after school PA in children and adolescent girls, but again the level of evidence was limited and not conclusive (0, Lns) [32].

In terms of the relationship between parental education level and PA in children and adolescents, <25% of the primary studies of the SLRs showed any significant association [34], with a probable level of evidence (0, Pe). Similarly, among SLRs that included only children or pre-schoolers, no association was found between parental education level and overall PA; though the level of evidence was limited (00, Lns) due to a limited number of primary studies (n = 10), most of which were cross-sectional (n = 10) in design.

Parental income, another measure of SES [33] can be added to the list of the factors associated with overall PA for children and adolescents, however the evidence available is limited (+, Ls) while parental home ownership [33] was found to have no correlation with overall PA in the same group (0, Ls). There were mixed results in relation to parental occupation and overall PA, with parental occupation found to be a determinant in some reviews (25% to 75% of available reviews or of the studies reviewed in these reviews), but not in others (+, Pe) [33,34]. However, number of parental working hours consistently showed no effect on overall PA levels with a probable level of evidence (00, Pe) [31]. Payment of fees or equipment required for PA [32] showed divergent results with limited suggestive evidence. A consistent association was found between payment of fees or equipment and overall PA in reviews combining children and adolescents (+, Ls), while another SLR on children only [29] consistently found no association with payment of fees or equipment and overall PA; though evidence was limited and not conclusive [29].

Adults (>18 years)

Among adults aged over 18 years, it emerged that SES [19,22,39] was the sole correlate, with convincing evidence in the majority of the available original studies (+, Ce). In addition to total PA, LTPA (+, Ce) and OPA (-, Ce) [19] were also found to be associated with SES, with a convincing level of evidence (>75% of the available original studies). Similarly, among older adults (> 65 years old), SES [38] was associated with overall PA and LTPA (+, Lns), though the evidence was limited.

There were mixed results in relation to individual income and overall PA, with income found to be a determinant in some reviews (25% to 75% of available reviews or of the studies reviewed in these reviews), but not in others (0, Pe) [22,36]. Furthermore, there was limited non-conclusive evidence (+, Lns) that income was associated with moderate and vigorous PA, based mainly on findings from cross-sectional studies and mixed non-conclusive evidence (0, Lns). There was limited non-conclusive evidence that neighbourhood income was not a correlate of leisure time PA (00, Lns). Similarly, neighbourhood income was unrelated to adherence to exercise referral schemes in adults over 40 years of age (00, Lns) [37]. Yet, there was some evidence that neighbourhood income was a factor when it came to the uptake of exercise referral schemes (+, Lns) [37] but the level of evidence was limited due to the low sample size and cross-sectional design of studies.

Single SLRs considered sub-categories adults; with one particular category being rural women [25]. In this sub-group, education level, income, payment of fees and equipment costs showed a limited non-conclusive association (++, Lns). Furthermore, limited suggestive evidence (++, Ls) emerged for the number of hours spent working. Another specific group investigated was Native Americans [39]. In this sub-group, employment was not correlated to overall PA (00, Ls), whereas findings related to educational levels and income were mixed (0, Ls). Finally, a qualitative SLR [24] considered the sub-group African-Americans. In this sub-group payment of fees and/or equipment costs were perceived as impediments to PA, mostly among women, though the level of evidence was limited and non-conclusive (++, Lns).

Discussion

This is the first umbrella SLR that provides a detailed overview of reviewed research regarding economic factors that influence PA across the life course. Factors studied most frequently among all age groups that demonstrated evidence of some association with PA, particularly in adults, were overall SES, income, payment of fees, and /equipment costs for PA. However, because of the general use of cross sectional designs in the studies covered in the available reviews, the evidence for true determinants is suggestive at best.

The included SLRs suggest that for adults, overall SES is the sole factor identified, with convincing evidence to be significantly related to overall PA, OPA and LTPA. This finding was also evident for older adults (>65 years), though the strength of the evidence was less convincing. In the reviews that examined specific components associated with SES such as education and income, the evidence was less consistent, resulting in mixed findings for overall PA and LTPA, with a probable level of evidence. The reasons behind these mixed results, as reported by the reviews are: small sample sizes, high diversity of the population included between studies and the diversity of the measurement methods of PA used among the primary studies.

While most associations were not statistically significant, both positive and negative associations were identified between educational level and PA with the direction of the relationship between education and PA being domain dependent. For example, one particular study found educational level to be positively associated with LTPA and negatively associated with OPA [19]. Similarly, in fact, the most recent data on PA from the European Commission [40] indicates that high socio-professional categories (e.g. managers, white collars, and self-employed) tend to engage in PA more frequently compared to the unemployed, retired, and those that work in the home. Furthermore, the majority (68%) of European citizens with a limited educational level (≤ 10 grade; 15yrs of age) report never exercising or playing sport, whereas those who ended education at 16–19 years and ≥20 years was 47% and 27%, respectively. These findings highlight the importance of domain specific PA research to accurately assess its association with different socioeconomic determinants. Distinguishing between PA behaviours by purpose (e.g. work, leisure), environment (e.g. location, type of community, physical environment), type (e.g. exercise) and time (e.g. time of the day, month or year) might assist researchers to further identify the determinants of specific PA behaviour and mitigate inconsistencies from previous studies.

Despite convincing evidence for the impact of SES measures on adult PA levels, the findings were less convincing for children and adolescents (< 18 years). There was no association with SES and overall PA in preschool children (3–6 years). In school-aged children and adolescents, (6–17 years), the findings were inconsistent. Socioeconomic indicators were not related to PA in children and adolescents when the groups were considered as one. When separating these groups, the findings varied, with some studies showing a relationship with SES and overall PA, while others showed no relationship. This may be due to the vastly different age ranges of children and adolescents (4–18 years) and a dramatic decline in PA from childhood to adolescence [41]. It is perhaps understandable that when youth is considered as one it is difficult to accurately identify associations.

Many changes with respect to physical development and social interactions are taking place during the transition from childhood to adolescence that directly influences PA levels. For example, when children are young, parents are highly responsible for their access to PA. However, parental influence decreases with advancing age, as the child gains more independence and is increasingly exposed to other environments (e.g. school environment, peer influences). Cost of PA has also been shown to be somewhat influenced dependent on the child’s age [41]. Among younger children, PA is most frequently informal in nature and scarcely gives rise to additional costs. In older children and adolescents, it often has more costs associated with it through membership of sports clubs and purchase of equipment, which can result in socially disadvantaged adolescents being less likely to remain, or become, active.

Another reason for the absence of a consistent direction in the SES inequalities and total PA in children and adolescents might be caused by the fact that PA is a complex variable with many components. One SLR looked at two specific activity domains for children and adolescents (e.g. break-time PA and after school PA) [32]. Stanley et al [32] showed that there was a consistent relationship between SES and break-time PA in both children and adolescents, while no consistent relationship between SES and after school PA was identified; albeit the level of evidence was weak and for females only. During the school day, physical education and playtime enables children to engage in regular PA, although it has shown these increases have only made a small contribution to total daily PA [30].

Based on the current literature, one may conclude that the importance of SES measures throughout the youth may increase with age but the sparse evidence needs further clarification. As with adults, investigating PA from a domain specific perspective in children and adolescents appears to be essential to achieve a detailed understanding of the determinants of SES measures on PA.

Limitations of the study

The main limitation that we have identified with regards to this umbrella review is the lack of SLRs available for inclusion. Only two SLRs [19,22] looked at SES or its derivatives (income, education, employment) as their primary outcome. The remainder of the included SLRs focused primarily on social, biological and environmental determinants and included on average only two studies per SLR that investigated potential socioeconomic correlates. Furthermore, most of the individual studies looked at only one component of SES such as income, education, or payment of fees in relation to PA; making both quality and strength of evidence assessment difficult.

Certain general limitations and assumptions of SES studies should be considered. Firstly, SES is a theoretical construct involving various measures (e.g., income, occupation, education) that tap into different components of this construct [42]. However, there is no overarching agreement in the literature on the use of specific SES measures and SES definitions applied in the studies often differ. Thus, a reported association between a given SES measure and PA may not always be consistent with findings observed in other studies due to inconsistent SES definitions employed. Although SES is the measure most commonly used in the reviews examined, educational level, income and occupation were also looked at in some reviews.

Umbrella reviews can be prone to bias in various ways. The SLRs examined in this umbrella review had mostly primary studies of a cross-sectional design and their findings were limited, in that only association could be established. Therefore, it was impossible to truly identify determinants. Significant differences in reviewing methodology and reporting were apparent. Most of the included reviews were of moderate methodological quality; with only one review having strong methodological quality. The majority of the SLRs did not include grey literature and the probability of publication bias was rarely assessed. Additionally, 18 out of the 19 SLRs did not provide lists of excluded studies and most did not assess and document the scientific quality of the included studies. Finally, there were no existing approved criteria for grading the evidence of the individual SLRs included in the umbrella review. In order to try and increase the relevance and comparability of our results, we therefore used grading methods that were applied in previous reviews of this kind [17,18,19].

By their nature, umbrella reviews lead to loss of detail, with some individual studies included in multiple SLRs. This may have led to an overrepresentation of single studies in our results. The majority of the studies included in the SLRs were also conducted in the developed world (Europe, North America and Australia). As a consequence, some socioeconomic determinants that maybe more relevant in less developed countries, or countries with greater inequalities, may not have been identified.

Conclusions

This is the first umbrella review providing an overview of socioeconomic correlates of PA across the life course. While some evidence is available for adults, less was available for youth. Specific SES measures identified such as overall SES, educational level, income and payment of fees/equipment for PA, should be used to develop and steer interventions and programmes in accordance with the target group. Individual and population-related interventions should primarily be targeted on correlates that can be strongly influenced and are most likely to bring about behaviour changes.

For future studies, it would be advisable not only to ensure an appropriate study design but also to use consistent, reliable and validated measurement methods in the assessment of PA and socioeconomic correlates. Using multiple domain measures of PA would also provide a more complete description of their associated correlates. An attempt to further understand the impact SES and its individual components has on adults, as well as on the parent-child PA relationship, is essential. Finally, in order to reach the most disadvantaged groups, a better understanding of the interrelationships between the dimensions of socio-economic circumstances and their effects on PA is paramount.

Supporting information

Acknowledgments

The authors thank N Lien, J Lakerveld, M Mazzocchi, D O’Gorman, P Monsivais, M Nicolaou, B Renner, D Volkert, and the DEDIPAC-KH Management team for their helpful support.

The results of the present study, presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation, do not constitute endorsement by ACSM. The authors declare no conflict of interest.

References

  1. 1. Warburton DER, Nicol CW, Bredin SSD. Health benefits of physical activity: the evidence Canadian Med Ass J. 2006; 174(6): 801–9
  2. 2. World Health Organization: Prevalence of insufficient physical activity. Global Health Observatory (GHO) data 2010; http://www.who.int/gho/ncd/risk_factors/physical_activity_text/en/
  3. 3. Telema R. Tracking of physical activity from childhood to adulthood: a review. Obes Facts. 2009; 2(3): 187–95 pmid:20054224
  4. 4. Huurre T, Aro H, Rahkonen O. Well-being and health behaviour by parental socioeconomic status: a follow-up study of adolescents aged 16 until age 32 years. Soc Psychiatry Psychiatr Epidemiol. 2003; 38(5): 249–255 pmid:12719840
  5. 5. Bauman AE, Sallis JF, Dzewaltowski DA, Owen N. Toward a better understanding of the influences on physical activity: the role of determinants, correlates, causal variables, mediators, moderators, and confounders. Am J Prev Med 2002; 23(2 Suppl): 5–14
  6. 6. Hayes AF. Introduction to mediation, moderation and conditional process analysis: A Regression-based approach. 2013: New York, NY: The Guilford Press pp. 3–22
  7. 7. European Commission. Expert Group on Health-enhancing physical activity: Recommendations to encourage physical education in schools, including motor skills in early childhood, and to create valuable interactions with the sport sector, local authorities and the private sector. 2015: http://ec.europa.eu/transparency/regexpert/index.cfm?do=groupDetail.groupDetailDoc&id=19860&no=1)
  8. 8. Marmot M. Social determinants of health inequalities. Lancet 2005; 365(9464):1099–104 pmid:15781105
  9. 9. Laaksonen M, Talala K, Martelin T, Rahkonen O, Roos E, Helakorpi S et al. Health behaviours as explanations for educational level differences in cardiovascular and all-cause mortality. Eur J Public Health. 2008;18(1): 38–43 pmid:17569702
  10. 10. Bukman AJ, Teuscher D, Feskens EJM, van Baak MA, Meershoek A, Renes RJ. Perceptions on healthy eating, physical activity and lifestyle advice: opportunities for adapting lifestyle interventions to individuals with low socioeconomic status. BMC Public Health 2014; 14:1036–63 pmid:25280579
  11. 11. Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults. JAMA. 1998; 279(21): 1703–8 pmid:9624022
  12. 12. Stringhini S, Sabia S, Shipley M, Brunner E, Nabi H PhD; Kivimaki M, Singh-Manoux A: Association of Socioeconomic Position With Health Behaviors and Mortality. JAMA 2010; 303(12): 1159–66 pmid:20332401
  13. 13. Macintyre S, Mutrie N. Socio-economic differences in cardiovascular Disease and physical activity: stereotypes and reality. J Royal Soc Prom Health 2004; 124: 66–9
  14. 14. Hartling L, Chisholm A, Thomson D, Dryden DM. A descriptive analysis of overviews of reviews published between 2000 and 2011. PLoS One 2012; 7: e49667 pmid:23166744
  15. 15. Aromataris E, Fernandez R, Godfrey C, Holly C, Khalil H, and Tungpunkom P. Summarizing systematic reviews: methodological development, conduct and reporting of an umbrella review approach. Int J Evid-Bas Healthcare 2015; 13: 132–40
  16. 16. Capranica L, MacDonncha C, Puggina A. PROSPERO International prospective register of systematic reviews Determinants of physical activity: an umbrella systematic literature review. 2015;1–5.
  17. 17. World Cancer Research Fund / American Institute for Cancer Research. Body fatness. In: Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective (2007). doi:978-0-9722522-2-5
  18. 18. Sleddens EF, Kroeze W, Kohl LF, Bolten LM, Velema E, Kaspers P et al. Correlates of dietary behavior in adults: an umbrella review. Nutr Rev 2015; 73(8): 477–99 pmid:26106126
  19. 19. Beenackers MA, Kamphius CBM, Giskes K, Brug J, Kunst AE, Burdorf A et al. Socioeconomic inequalities in occupational, leisure-time, and transport related physical activity among European adults: A systematic review. Int J Behav Nut Phys Act 2012; 9:116–138
  20. 20. Shea BJ, Grimshaw JM, Wells G, Boers M, Andersson N, Hamel C et al. Development of AMSTAR: a measurement tool to assess the methodological quality of systematic reviews. BMC Med. Res. Methodol. 2007;7:10–23 pmid:17302989
  21. 21. Sleddens EF, Kroeze W, Kohl L, Bolten LM, Velema E, Kaspers PJ et al. Determinants of dietary behavior among youth: an umbrella review. Int. J. Behav. Nutr. Phys. Act. 2015;12: 7–29 pmid:25638322
  22. 22. Gidlow C, Johnston LH, Crone D, Ellis N, James D. A systematic review of the relationship between socio-economic position and physical activity. Health Educ Jour 2006; 65(4): 338–67
  23. 23. Lu C, Stolk RP, Sauer PJ, Sijtsma A, Wiersma R, Huang G, Corpeleij E. Factors of physical activity among Chinese children and adolescents. Int J Behav Nut Phys Act 2017; 14: 36–46
  24. 24. Siddiq Z., Tiro JA, Shuval K. Understanding impediments and enablers to physical activity among African American adults: A systematic review of qualitative studies. Health Educ. Res. 2011; 26(6):1010–24 pmid:21873458
  25. 25. Olsen JM. An Integrative Review of Literature on the Determinants of Physical Activity Among Rural Women. Public Health Nurs. 2013; 30(4), 288–311 pmid:23808855
  26. 26. Craggs C, Corder K, van Sluijs EM, Griffin SJ. Determinants of Change in Physical Activity in Children and Adolescents. Am. J. Prev. Med. 2011; 40(6): 645–58 pmid:21565658
  27. 27. De Craemer M, De Decker E, Bourdeaughuij I, Vereecken C, Deforche B, Manios Y et al. ToyBox-study group. Correlates of energy balance-related behaviours in preschool children: a systematic review. Obes. Rev. 2012; 13(Suppl 1): 13–28
  28. 28. Hinkley T, Crawford D, Salmon J, Okely AD, Hesketh K. Preschool Children and Physical Activity. Am J Prev Med. 2008; 349(5): 435–441
  29. 29. Edwardson CL, Gorely T. Parental influences on different types and intensities of physical activity in youth: A systematic review. Psych Sport Exer 2010; 11: 522e
  30. 30. Ridgers ND, Salmon J, Parrish AM, Stanley RM, Okely AD. Physical Activity During School Recess. Am J Prev Med 2012; 43(3): 320–8 pmid:22898126
  31. 31. Van Der Horst K, Paw MJ, Twisk Jw, van Mechelen W. A Brief Review on Correlates of Physical Activity and Sedentariness in Youth. Med Sci Sport Exerc 2007; 39(8): 1241–50
  32. 32. Stanley RM, Ridley K, Dollman J. Correlates of children’s time-specific physical activity: a review of the literature. Int. J. Behav. Nutr. Phys. Act. 2012; 9: 40–56
  33. 33. Ferreria I, van der Horst K, Wendel-Vos W, Kremers S, van Lenthe FJ, Brug J. Environmental correlates of physical activity in youth–a review and update. Obes. Rev. 2006; 8(2): 129–54
  34. 34. Gustafson SL, Rhodes RE. Parental Correlates of Physical Activity in Children and Early Adolescents. Sports Med. 2006; 36(1): 79–97 pmid:16445312
  35. 35. Beets MW, Cardinal BJ, Alderman BL. Parental Social Support and the Physical Activity–Related Behaviors of Youth: A Review. Health Educ Behav. 2010; 37(5): 621–44 pmid:20729347
  36. 36. Wendel-Vos W, Droomers M, Kremers S, Brug J, van Lenthe F. Potential environmental determinants of physical activity in adults: a systematic review. Obes. Rev; 2007; 8(5): 425–40 pmid:17716300
  37. 37. Pavey T, Taylor A, Hillsdon M, Fox K, Campbell J, Foster C et al. Levels and predictors of exercise referral scheme uptake and adherence: a systematic review. J. Epidemiol. Community Heal. 2012; 66(8): 737–44
  38. 38. Barnett I, Guell C, Ogilvie D. The experience of physical activity and the transition to retirement: a systematic review and integrative synthesis of qualitative and quantitative evidence. Int. J. Behav. Nutr. Phys. Act. 2012; 9: 97–111 pmid:22897911
  39. 39. Coble JD, Rhodes RE. Physical Activity and Native Americans. A Review. Am. J. Prev. Med. 2006; 31(1): 36–46 pmid:16777541
  40. 40. European Commission. Sport and Physical Activity Report: Special Eurobarometer 412 2014: http://ec.europa.eu/public_opinion/index_en.htm
  41. 41. Chung AE, Skinner AC, Steiner MJ, Perrin EM. Physical Activity and BMI in a Nationally Representative Sample of Children and Adolescents. Clin Pediatr 2012; 51(12): 122–9
  42. 42. Geyer S, Hemstrom O, Peter R, Vagero D. Theory and methods: Education, Income and occupational class cannot be used interchangeably in social epidemiology. Empirical evidence against a common practice. J Epidemiol Comm Health 2006; 60(9): 804–810