Kirsi E. Keskinen
JYU DISSERTATIONS 391
Features of the Physical
Environment, Walking Difficulty,
and Physical Activity in Old Age
JYU DISSERTATIONS 391
Kirsi E. Keskinen
Features of the Physical
Environment, Walking Difficulty, and Physical Activity in Old Age
Esitetään Jyväskylän yliopiston liikuntatieteellisen tiedekunnan suostumuksella julkisesti tarkastettavaksi kesäkuun 11. päivänä 2021 kello 12.
Academic dissertation to be publicly discussed, by permission of the Faculty of Sport and Health Sciences of the University of Jyväskylä,
on June 11, 2021 at 12 o’clock.
JYVÄSKYLÄ 2021
Editors Anne Viljanen
Faculty of Sport and Health Sciences, University of Jyväskylä Ville Korkiakangas
Open Science Centre, University of Jyväskylä
Copyright © 2021, by University of Jyväskylä
ISBN 978-951-39-8682-7 (PDF) URN:ISBN:978-951-39-8682-7 ISSN 2489-9003
Permanent link to this publication: http://urn.fi/URN:ISBN:978-951-39-8682-7
ABSTRACT Keskinen, Kirsi E.
Features of the physical environment, walking difficulty, and physical activity in old age
Jyväskylä: University of Jyväskylä, 2021, 101 p.
(JYU Dissertations ISSN 2489-9003; 391)
ISBN 978-951-39-8682-7 (PDF)
Mobility enables engagement in out-of-home activities and thus is important for the well-being of older adults. Walking difficulty may limit mobility, and low physical activity is one risk factor for walking difficulty. Environmental features close to home may hinder or encourage physical activity in older adults.
This study aimed to contribute knowledge on features of the physical environment that support the physical functioning and physical activity (PA) of older adults, and to investigate whether the associations between environmental features and PA may differ by perceived walking difficulty, type of residential neighbourhood, and day of the week. The data were collected as part of the Life- Space Mobility in Older People (n=848) project in structured interviews, performance tests, and accelerometer measurements (n=167) among community- dwelling 75- to 90-year-olds living in Central Finland. The data on environmental features were retrieved from open sources of geospatial data.
The results showed that a more steeply sloping road network and lower PA were risk factors for developing walking difficulty. The supportiveness of the infrastructure for outdoor mobility and diversity of large natural, green areas were positively, and slope negatively, associated with PA. Perceived infrastructure-based facilitators in dispersed areas and nature-based destinations in city centres and densely populated areas outside centres were associated with higher PA. Perceived nature was a more important PA facilitator for older adults with than without walking difficulty. Among those with walking difficulty, the presence of waterside areas increased while higher diversity in large natural and green areas decreased the odds for higher PA. Among those without walking difficulty, higher diversity in large natural and green area increased the odds for higher PA and for perceiving nature as a facilitator for outdoor mobility. Nature- based features were associated with PA consistently throughout the week but infrastructure-based features only on weekdays. Depending on the home location, environmental features appear to support older adults’ engagement in outdoor PA. At the greatest, when converted to PA volumes, the difference in supportiveness might correspond to the difference in PA accumulation between older adults with and without walking difficulty.
Keywords: outdoor mobility, mobility limitation, neighbourhood, infrastructure, nature, older adults
TIIVISTELMÄ (ABSTRACT IN FINNISH) Keskinen, Kirsi E.
Fyysisen ympäristön piirteet, kävelyvaikeudet ja fyysinen aktiivisuus iäkkäillä ihmisillä
Jyväskylä: University of Jyväskylä, 2021, 101 p.
(JYU Dissertations ISSN 2489-9003; 391)
ISBN 978-951-39-8682-7 (PDF)
Kodin ulkopuolella liikkuminen on tärkeää iäkkäiden ihmisten elämänlaadulle.
Ulkona liikkumista voivat rajoittaa kävelyvaikeudet, joille vähäinen fyysinen ak- tiivisuus altistaa. Kodin lähistön ympäristönpiirteet voivat kannustaa liikku- maan tai estää sitä.
Tämän tutkimuksen tarkoituksena oli lisätä tietoa iäkkäiden ihmisten käve- lykykyä ja reipasta liikkumista tukevista ympäristönpiirteistä ja siitä, eroavatko ne henkilön kävelykyvyn, asuinalueen tai liikkumispäivän mukaan. Tutkimuk- sessa käytettiin Iäkkäiden ihmisten liikkumiskyky ja elinpiiri (n=848) aineistoa, joka oli kerätty strukturoidulla haastattelulla, toimintakykymittauksilla ja kiihty- vyysmittarilla (n=167) kotona asuvilta 75-90 vuotiailta jyväskyläläisiltä ja muu- ramelaisilta henkilöiltä. Ympäristönpiirteiden tiedot kerättiin avoimista paikka- tietoaineistoista.
Tämä tutkimus osoitti, että runsaampi kävelyä tukeva infrastruktuuri ja ison luontoalueen monimuotoisuus kodin lähellä olivat positiivisesti ja tiever- koston suurempi mäkisyys negatiivisesti yhteydessä reippaamman liikkumisen määrään. Tieverkoston suurempi mäkisyys ja vähäisempi liikkuminen olivat it- senäisiä riskejä kävelyvaikeuksien ilmaantumiselle. Reippaampaan liikkumiseen yhdistyivät houkuttavaksi koetut infrastruktuuritekijät harvemman asutuksen alueilla ja lähiluonnon kohteet kaupungin keskustassa sekä tiheillä asuinalueilla.
Liikkumaan houkuttajana lähiluonto oli tärkeämpi kävelyvaikeuksia kokeville kuin niitä kokemattomille. Kävelyvaikeuksia kokevilla veden läheisyys lisäsi mutta suurempi monimuotoisuus laajalla luontoalueella vähensi lähiluonnon kokemista houkuttavana. Kävelyvaikeuksia kokemattomilla ison luontoalueen suurempi monimuotoisuus lisäsi todennäköisyyttä kokea luonto houkuttajana ja liikkua reippaasti. Luontopohjaiset ympäristötekijät yhdistyivät fyysiseen aktivi- suuteen samankaltaisesti viikonpäivästä riippumatta mutta infrastruktuuriteki- jät vain arkipäivisin. Ympäristön tarjoama tuki iäkkäiden ihmisten ulkona liik- kumiselle vaihtelee suuresti kodin sijainnista riippuen. Suurimmillaan ero tuessa vastasi fyysisessä aktiivisuudessa olevaa eroa kävelyvaikeuksia kokevien ja ko- kemattomien välillä.
Asiasanat: ulkona liikkuminen, fyysinen toimintakyky, naapurusto, infrastruk- tuuri, luonto, iäkkäät ihmiset
Author Kirsi E. Keskinen, MSc
Gerontology Research Center and Faculty of Sport and Health Sciences University of Jyväskylä
Finland
kirsi.e.keskinen@jyu.fi ORCID 0000-0001-9876-5658
Supervisors Senior Researcher Erja Portegijs, PhD Gerontology Research Center and Faculty of Sport and Health Sciences University of Jyväskylä
Finland
Professor Taina Rantanen, PhD Gerontology Research Center and Faculty of Sport and Health Sciences University of Jyväskylä
Finland
Professor Emeritus Kimmo Suomi, PhD Faculty of Sport and Health Sciences University of Jyväskylä
Finland
Reviewers University Researcher Jouni Lahti, PhD, Title of Docent Center for Population, Health and Society
Faculty of Social Sciences University of Helsinki Finland
Professor Catharine Ward Thompson, PhD OPENspace research centre
Edinburgh School of Architecture and Landscape Architecture
University of Edinburgh United Kingdom
Opponent Unit Head and Program Manager Katja Borodulin, PhD, Title of Docent
Age Institute Finland
ACKNOWLEDGEMENTS
This PhD dissertation was carried out at the Gerontology Research Center and the Faculty of Sport and Health Sciences at the University of Jyväskylä, Finland.
My deepest gratitude goes to my supervisors, Senior Researcher Erja Portegijs, Professor Taina Rantanen, and Professor Emeritus Kimmo Suomi. Erja, I want to thank you for giving me the opportunity to start my academic career in your GEOage project. You have been patient, kind, and wise in teaching me how to conduct research. You always found time for me, and your advice and encouragement inspired me to find solutions, and sometimes also prompted questions that would not otherwise have occurred to me. Taina, thank you for the valuable guidance you have given me. I admire your professional skills and your drive in generating research, and I feel honoured to have conducted my PhD thesis under your supervision. Kimmo, thank you sharing your expertise and contributing viewpoints from outside the domain of gerontology. I appreciate the wider perspective your advice brought to my research. I also want to thank Senior Lecturer and Division Head of Natural Resources and Environment Anssi Lensu for introducing me to the fascinating world of geospatial data and geospatial analysis and being a member of the steering group of my PhD project.
I thank Research Director Katja Kokko at the Gerontology Research Center and Professor Ari Heinonen, the Dean of the Faculty of Sport and Health Sciences, for giving me the opportunity to belong to this unique and multidisciplinary work community. I want to express my sincere gratitude to the official reviewers of this thesis, University Researcher Jouni Lahti and Professor Catharine Ward Thompson. I also want to thank Unit Head and Program Manager Katja Borodulin for agreeing to be my opponent in the public defence of this dissertation.
I want to thank Principal Researcher Merja Rantakokko, PhD, and Ying Gao, PhD, who as co-authors have contributed to this work. I express my warmest thanks also to Michael Freeman for language editing of this thesis as well as all the original papers and to Anne Viljanen, PhD, for scientific editing of this thesis.
I have been fortunate to have my dissertation financed largely by the GEOage project, which was funded by the Ministry of Education and Culture, Finland. I am also grateful for a personal grant from the Finnish Cultural Foundation (Ester and Uuno Kokki Fund), and for personal and travel grants from the Foundation for Municipal Development. I also thank the Faculty of Sport and Health Sciences for my term as a faculty PhD student. I am also grateful to Professor Taina Rantanen for the opportunity to work for a half-year in the AGNES project. Although data and maps are captivating in themselves, the experience of interviewing older people about their activity environments and dealing with practical issues in the project provided me with knowledge on both the topic and doing research in general that I could not have gained otherwise.
It has been a pleasure to work at the Gerontology Research Center, and for this I want to thank my co-workers. Not just for your help in resolving research-
related problems but also, and above all, for your company and the good laughs we have enjoyed together. These have truly brightened up my days. Many thanks especially to Johanna Eronen, PhD, Katja Pynnönen, PhD, Taina Poranen-Clark, PhD, and Matti Munukka, PhD, who have shared an office with me, and to Tiia Kekäläinen, PhD and to all my fellow PhD students, for your friendship.
Last, I want to thank my family, relatives, and friends for their unconditional support. You never doubted whether I would make it or not.
Special thanks go to those closest to me. To my parents, Marja and Antti, I am grateful for all the support and practical help you have given me and my family during these years of my PhD project. Aku, thank you for all your support. Eero, Annu and Noora, my dearest ones, I am grateful for having you in my life.
Jyväskylä 10.5.2021 Kirsi Keskinen
FIGURES
FIGURE 1 The Ecological model of four domains of active living adapted from Sallis et al. (2006). ... 31 FIGURE 2 Theoretical framework of the study applying the social ecological
approach and consisting of the three layers of personal factors, physical activity (PA) behaviour, and external factors and three theoretical models of possible associations between the factors in the different layers. ... 34 FIGURE 3 Analytical framework of the study. Outcome measures are bolded.
PA = Physical activity. ... 36 FIGURE 4 A map example of the reclassified Corine Land Cover data and 500- m buffer with the participant’s home at the center. ... 46 FIGURE 5 Participants’ home locations and neighbourhood types in the study area at baseline. (Modified from Study III.) ... 56 FIGURE 6 Bland-Altman plots for accelerometer-measured a) number of PA
bouts on weekend days vs. weekdays; and b) MVPA minutes on weekend days vs. weekdays. Mean difference with 95% confidence intervals and regression line (n=167). PA = Physical activity; MVPA
= Moderate to vigorous physical activity. (Modified from Study IV.) ... 64
TABLES
TABLE 1 Summary of study designs and participants ... 37 TABLE 2 Summary of and references for participant baseline
measurements... 40 TABLE 3 Environmental features and datasets used in operationalising the
features into objective measures ... 44 TABLE 4 Characteristics of participants in the cross-sectional, prospective
and accelerometer sub-study analyses ... 55 TABLE 5 Summary table of physical activity levels among study participants stratified according to contextual factors at baseline ... 57 TABLE 6 Objectively assessed features of participants’ neighbourhood
environments in the different study samples at baseline ... 58 TABLE 7 Logistic regression analyses on the prevalence of walking difficulty
at baseline (n=848) and incidence of walking difficulty over the two-year follow-up (n=551) ... 59 TABLE 8 Associations of objectively assessed features of the natural
environment with reporting at least moderate physical activity (PA) (vs. only light PA) and with perceiving (vs. not perceiving) nature as a facilitator for outdoor mobility (n=848) ... 61 TABLE 9 Reports of perceived facilitators for outdoor mobility in different
neighbourhood types, and likelihood for self-reporting higher physical activity (PA) compared to only light PA when perceiving the facilitator (vs. not perceiving the facilitator) (n=848)... 63 TABLE 10 Associations of objectively assessed features of environment with
number of PA bouts and MVPA minutes on weekdays and
weekend days (n=167) ... 65 TABLE 11 Associations of objective measures of environment and walking
difficulty with overall number of PA bouts and MVPA minutes from linear regression models and estimated differences in physical activity due to high-low differences in objective measures and walking difficulty among the study participants (n=167) (Study IV and unpublished results) ... 67
LIST OF ORIGINAL PAPERS
This dissertation is based on the following original publications, which are referred to by Roman numbers. This thesis includes also unpublished data.
I Keskinen, K. E., Rantakokko, M., Suomi, K., Rantanen, T. & Portegijs, E. 2020. Hilliness and the development of walking difficulties among community-dwelling older people. Journal of Aging and Health 32 (5-6), 278-284. doi: 10.1177/0898264318820448.
II Keskinen, K. E., Rantakokko, M., Suomi, K., Rantanen, T. & Portegijs, E. 2018. Nature as a facilitator for physical activity: Defining relationships between the objective and perceived environment and physical activity among community-dwelling older people. Health
& Place 49, 111-119. doi: 10.1016/j.healthplace.2017.12.003.
III Keskinen, K. E., Rantakokko, M., Suomi, K., Rantanen, T. & Portegijs, E. 2020. Environmental features associated with older adults' physical activity in different types of urban neighborhoods. Journal of Aging and Physical Activity 28 (4), 540-548. doi:
10.1123/japa.2019-0251.
IV Keskinen, K. E., Gao, Y., Rantakokko, M., Rantanen, T. & Portegijs, E. 2020. Associations of environmental features with outdoor physical activity on weekdays and weekend days: a cross-sectional study among older people. Frontiers in Public Health 8: 578275. doi:
10.3389/fpubh.2020.578275.
In the original publications, I was the first author, while also considered the comments from the co-authors. I formulated the study questions and designs for the publications, prepared the participant data for the statistical analyses, conducted all the statistical analyses, interpreted the results, and had the main responsibility for writing the manuscripts. Accelerometer-measured data on PA in Study IV was prepared and described by the second author of that paper. I was also responsible for acquiring the objective environmental data for all publications. For this purpose, I retrieved geospatial datasets from open sources, designed the operationalisation of environmental features into objective measures using geospatial data, and performed spatial operations to create objective environmental measures. In all four studies, I was privileged to use pre- existing data.
ABBREVIATIONS
CHAMPS Community Healthy Activities Model Program for Seniors CI Confidence intervals
CLC Corine Land Cover
GEOage Geographic characteristics, outdoor mobility and physical activity in old age
GIS Geographic information system GPS Global positioning system
h Hour
IPAQ International Physical Activity Questionnaire IQR Interquartile range
km Kilometre
km2 Square kilometre
LISPE Life-Space Mobility in Old Age
m Metre
MAD Mean amplitude deviation MET Metabolic equivalent min Minutes
MMSE Mini-Mental State Examination
MVPA Moderate to vigorous physical activity
n Number
OECD Organisation for Economic Co-operation and Development OR Odds ratio
pi Proportion of land area of the land type i in relation to the buffer area PA physical activity
PASE Physical Activity Scale for the Elderly
PENBOM Checklist for perceived environmental barriers to outdoor mobility PENFOM Checklist for perceived environmental facilitators for outdoor
mobility
R2 Coefficient of Determination S Number of all possible land types SD Standard deviation
SHDI Shannon's Diversity Index
SOC Selection, optimisation, and compensation SPPB Short Physical Performance Battery USA United States of America
WHO World Health Organisation
yrs Years
CONTENTS ABSTRACT
TIIVISTELMÄ (ABSTRACT IN FINNISH) ACKNOWLEDGEMENTS
FIGURES AND TABLES LIST OF ORIGINAL PAPERS ABBREVIATIONS
CONTENTS
1 INTRODUCTION ... 15
2 REVIEW OF THE LITERATURE ... 18
2.1 Walking difficulty ... 18
2.1.1 Walking difficulty as mobility limitation in old age ... 18
2.1.2 Assessing walking difficulty ... 19
2.2 Physical activity ... 20
2.2.1 Defining and assessing physical activity ... 20
2.2.2 Outdoor physical activity in old age ... 22
2.2.3 Associations between physical activity and walking difficulty… ... 23
2.3 Physical environments ... 24
2.3.1 Planning of physical environments ... 24
2.3.2 Objective and subjective features of the physical environment… ... 25
2.3.3 Neighbourhood areas in assessing environmental features ... 25
2.4 Associations of features of the physical environment with walking difficulty and physical activity among older adults ... 27
2.4.1 Associations between environmental features and walking difficulty ... 27
2.4.2 Associations between environmental features and physical activity ... 27
2.4.3 Contextual factors in environment - physical activity associations ... 29
2.5 The social ecological approach to studying physical activity ... 30
3 PURPOSE OF THE STUDY ... 35
4 MATERIALS AND METHODS ... 37
4.1 Study design and participants ... 37
4.2 Ethics ... 39
4.3 Measurements ... 40
4.3.1 Participant measures ... 40
4.3.1.1 Walking difficulty ... 40
4.3.1.2 Physical activity ... 41
4.3.1.3 Perceived features of environment ... 42
4.3.1.4 Participant characteristics ... 43
4.3.2 Objectively assessed features of the environment ... 43
4.3.2.1 Hilliness ... 44
4.3.2.2 Waterside areas ... 45
4.3.2.3 Nature and green areas ... 45
4.3.2.4 Outdoor mobility-supportive infrastructure ... 47
4.3.2.5 Urban structure ... 48
4.4 Statistical analyses ... 49
4.4.1 Descriptive statistical analyses ... 49
4.4.2 Binary logistic regression ... 50
4.4.3 Linear regression ... 51
4.4.4 Estimation of relevance of environmental features and walking difficulty for physical activity ... 51
4.4.5 Sensitivity analyses ... 52
4.4.6 Missing data ... 52
5 RESULTS ... 54
5.1 Hilliness, physical activity, and walking difficulty (Study I) ... 59
5.2 Environment-physical activity associations with contextual factors 60 5.2.1 Walking difficulty as a contextual factor (Study II) ... 60
5.2.2 Neighbourhood type as a contextual factor (Study III) ... 62
5.2.3 Day of the week as a contextual factor (Study IV) ... 64
5.3 Estimated influence of walking difficulty vs. objectively assessed features of the environment on physical activity (Study IV and previously unpublished results) ... 65
6 DISCUSSION ... 68
6.1 Interplay between physical environment, walking difficulty, and physical activity in old age ... 69
6.2 Reflections on the theoretical models used ... 74
6.3 Methodological considerations ... 76
6.4 Implications and future directions ... 79
7 MAIN FINDINGS AND CONCLUSIONS ... 82
REFERENCES ... 84 ORIGINAL PAPERS
15
Outdoor mobility is essential for older adults to be able to participate in society and access community amenities. Optimal mobility means “being able to safely and reliably go where you want to go, when you want to go, and how you want to get there” (Satariano et al. 2012). On this definition, when mobility is optimal, a person has no spatial or temporal restrictions or limitations in the choice of mode of mobility. The definition also includes environmental factors as contrib- utors to optimal mobility, as the safety and reliability of mobility may also de- pend on external factors. With mobility limitation, restrictions or constraints may arise in any of these aspects of optimal mobility. Conversely, each of the aspect may contribute positively to outdoor mobility.
Walking is a basic element of mobility and at the same time a common form of physical activity (PA), which by definition is “any bodily movement produced by skeletal muscles that results in energy expenditure” (Caspersen et al. 1985).
Thereby, PA can be considered almost a prerequisite for any type of mobility. For older adults, walking is an essential component of mobility and, when adequate, walking performance enables the unassisted use of other forms of transportation (Rantanen 2013). With increasing age, walking difficulty becomes more common (Centers for Disease Control and Prevention 2009; Sainio et al. 2012). For example, in Finland, more than one third of older adults over 75 report experiencing at least some difficulty in walking 500 metres (m) (Sainio et al. 2012). Onset of walking difficulty may occur for various reasons. In older adults, a low level of PA is one of the risk factors for developing walking difficulty (Brown & Flood 2013), whereas regular aerobic PA reduces the risk for mobility limitations (Paterson & Warburton 2010). Physical activity has also several other health benefits, and higher volumes of PA may even reduce mortality risk in older adults (Bangsbo et al. 2019).
Physical activity guidelines advise older adults to engage weekly in aerobic PA from 150 to 300 minutes at moderate or 75 minutes at vigorous intensity and also be otherwise physically active whenever they can (2018 Physical Activity Guidelines Advisory Committee 2018; UKK-institute 2019; WHO 2020). In Finland, for example, 12 % of older adults over age 75 (Bennie et al. 2017), meet
1 INTRODUCTION
16
this recommendation. Although walking is a popular form of leisure-time PA (Official Statistics of Finland 2009; Dai et al. 2015), a large proportion of older adults’ total PA is related to the performance of daily activities (Colley et al. 2019) such as shopping and social activities (Tsai et al. 2016; Cerin et al. 2017). Moreover, moving outside home is necessary for an older adult to achieve a higher volume of PA (Portegijs et al. 2015).
Most of older adults’ PA takes place close to home (Chaudhury et al 2016).
In the home neighbourhood, specific features of the physical environment, including both the built and natural environment, may serve as motivators or barriers to PA and outdoor mobility. Recent systematic reviews and meta- analyses have shown that presence of diverse destinations and walk-friendly infrastructure enables higher PA among older people (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018). In more urban areas, living in a neighbourhood with higher residential density and a denser road network is associated with being more active physically (Barnett et al. 2017; Cerin et al. 2017;
Van Cauwenberg et al. 2018). However, in reality, great differences exist in the supportiveness of neighbourhoods for outdoor mobility. A large international study estimated that differences in the extent of environmental features supporting outdoor mobility between the least and the most activity-supporting urban environments could mean a difference of more than 60 minutes in the amount of moderate to vigorous PA (MVPA) performed weekly by adults (Sallis et al. 2016). To the authors’ best knowledge, corresponding estimates have not thus far been reported for older adults.
However, being physically active is not solely a result of walking capability and a favourable physical environment. Social ecological models, such the Ecological Model of Four Domains of Active Living (Sallis et al. 2006), posit that various personal and perceived environmental factors and social, cultural, and policy factors may affect PA behaviour. At the same time, it is possible that the associations between separate factors and PA are modified by other factors in the social ecological framework. Thus, depending on the context of among whom, where, and when the associations are investigated, different features of the environment may be identified as beneficial for PA. However, less is known about the factors modifying the associations between the environment and PA in old age (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018).
With the current high proportion of older people in the general population and its estimated increase in the coming decades (United Nations, Department of Economic and Social Affairs, Population Division 2017), enhancing PA in older adults is a global issue of high societal importance. The importance of environments enabling older people to engage in PA has been highlighted by the WHO (WHO 2016) and the European Union (European Union 2020), both advocating their member states and partners to plan age-friendly environments.
Environments that enable and attract older adults to engage in PA can contribute positively to their physical capability and possibilities for outdoor mobility at the population level.
17
This study was conducted to contribute knowledge on the environmental features that support older adults’ outdoor mobility by contributing positively to physical functioning and PA. The associations between several features of the physical environment, walking difficulty, and PA were investigated. By considering contextual factors in these associations, this research aimed to clarify how walking difficulty, type of neighbourhood, and day of the week shape the associations between the environment and PA. In addition, the relevance of considering environmental features as PA correlates was examined by estimating differences in PA accumulation due to differences in local environmental features in the study area.
18 2.1 Walking difficulty
2.1.1 Walking difficulty as mobility limitation in old age
Optimal mobility means “being able to safely and reliably go where you want to go, when you want to go, and how you want to get there” (Satariano et al. 2012).
Mobility limitation refers to limitation in covering spatial areas or distances, timing or frequency of moving, or in ways of preferred travel. Mobility limitation thus threatens possibilities for active participation in society and independence in many areas of life. Mobility also refers to physical and physiological ability:
walking, for example, sets requirements for the functioning of the musculoskeletal, cardio-respiratory, sensory, and neural systems (Rantanen 2013). Decline in physiological and sensory capacities are part of the aging process (Manini & Pahor 2009; Viljanen et al. 2012), and hence aging-related decline in mobility capability relates to impairments in body systems (Stenholm et al. 2015). The gradual and rather slow functional decline with increasing age may be accelerated by diseases affecting the physiological systems essential for walking (Grimmer et al. 2019) and lead to increased risk for functional limitations and disabilities (Verbrugge & Jette 1994). In general, the most common risk factors for walking difficulty are higher age, low level of PA, obesity, strength or balance deficits, and chronic diseases (Brown & Flood 2013). Additionally, among older adults, cognitive decline may also contribute to decline in physical functioning (Anton et al. 2015).
Difficulty in walking, especially over longer distances is a typical early sign of mobility decline (Rantanen 2013). The concept of walking difficulty may be clarified by applying the indicators of mobility limitation to walking. In general, walking difficulty refers to not being able to walk where or how far one wishes or needing help in walking from others. Perceived difficulty in walking 500 m is already an indicator of mobility limitation and of being at the threshold of losing independent community mobility (Mänty et al. 2007). Walking difficulty
2 REVIEW OF THE LITERATURE
19
becomes more common with ageing (Centers for Disease Control and Prevention 2009; Sainio et al. 2012). According to the Health 2011 study, 35% of men and 49%
of women aged 75 years or older in Finland perceived at least some difficulty in walking 500 m (Sainio et al. 2012). According to the 2005 Survey of Income and Program Participation, conducted in the United States, 32% of noninstitutionalised persons age 65 and over reported difficulty in walking three city blocks (approximately 300 m; Centers for Disease Control and Prevention 2009).
2.1.2 Assessing walking difficulty
While perceived difficulty in walking may be assessed by self-reports, mobility performance can also be assessed by direct standardised performance-based assessments or by observation. While several approaches exist, some of these have become more widely used during recent years. In large studies, self-reports of perceived difficulty in walking certain distances (e.g., a quarter of a mile, a mile, 500 m, 2 kilometres (km)) or climbing a flight of stairs or 10 steps have commonly been used in assessing mobility among community-dwelling older adults (Chung et al. 2015). Commonly used performance-based instruments of mobility limitation, include measurement of gait speed and performance-based test batteries such as the Short Physical Performance Battery (SPPB; Guralnik et al. 1994; Chung et al. 2015). The SPPB includes tests of balance, walking speed over 2.44 m, and total time to perform five chair rises, with higher summary score (range 0-12) indicating better performance (Guralnik et al. 1994). Summary scores of the SPPB test are indicative of strength, and especially power, of the leg muscles (Bean et al. 2003; Reid et al. 2008); this in turn is associated with better functional status, including better performance in walking tasks (Foldvari et al.
2000).
Performance-based measures and self-reported measures of mobility capture somewhat different, but nevertheless closely related, aspects of mobility.
For example, less than 10% of those scoring at least 10 points in the SPPB reported needing help in performing a half-mile walk, whereas more than 50% of those scoring less than five points needed help in walking the same distance (Guralnik et al. 1994). Moreover, SPPB scores of 10 or lower predict loss of ability to walk 400 m after three years (Vasunilashorn et al. 2009). However, despite of being closely associated, performance-based test measures have not explained all the variance in real-life mobility capability whether assessed as self-reported walking capability (Guralnik et al. 1994) or as objectively defined activity measures and gait characteristics (Giannouli et al. 2016). The different types of measures complement each other, as performance-based test measures provide information on the upper limit of performance in specific functions in standardised conditions, while self-reports of mobility reflect individuals’
subjective estimates of how well they could do the task in their own environment in real-life conditions (Guralnik et al. 1994). Hence, conceptually, self-report or self-assessment methods might be better suited to assess optimal mobility among
20
older adults as they may better capture the person and environment interaction in different dimensions of mobility.
2.2 Physical activity
2.2.1 Defining and assessing physical activity
Physical activity is part of everyday life for nearly everyone. It is accumulated from any type of behaviour and daily chore in which one moves oneself, and has been defined as “any bodily movement produced by skeletal muscles that results in energy expenditure” (Caspersen et al. 1985). Being physically active contributes to preventing mobility loss. Going outdoors and walking even short distances regularly helped older people to maintain walking performance at a threshold level that enabled independent community mobility (Simonsick et al.
2005), thus facilitating possibilities for independent living and participation in society. Higher PA is associated with lower risk for several diseases and with reduced mortality risk among older adults (Bangsbo et al. 2019).
Physical activity domains, classified by purpose, include occupational, household/domestic, transport/utilitarian, and leisure/recreational PA (Sallis et al. 2006; Strath et al. 2013). When the way of performing the activity is aimed specifically at maintaining or improving physical fitness (Caspersen et al. 1985) and is repetitive, planned, and structured (Wong et al. 2003), physical activity is considered physical exercise. Based on this definition, physical activity can broadly be divided into exercise and habitual physical activity (Wong et al. 2003).
Further, physical activity may also be categorised according to its strenuousness, e.g., as of light, moderate, or vigorous intensity, based on its immediate physiological effects on energy expenditure (Strath et al. 2013).
In assessing physical activity, regardless of how it is categorised, information on the mode or type, frequency, duration, and/or intensity of physical activity is needed (Strath et al. 2013). Several methods of acquiring this data are available. Questionnaires and activity diaries are considered as subjective methods for assessing physical activity, as they rely on self-reports of physical activities. Measurements and direct observations are more objective methods, as they record direct indicators of physical activity (Strath et al. 2013).
Which of these methods is chosen will depend on the primary outcome variable of interest, number of participants, schedule, available resources, and participant burden (Strath et al. 2013; Sylvia et al. 2014). The doubly labelled water method is the gold standard for assessing total energy expenditure, but owing to cost, participant burden, and the time needed, it is not commonly used in PA research studies (Sylvia et al. 2014). Among older adults, the vast majority of studies have used self-reports, and a considerably smaller proportion of studies has utilized accelerometer measurements (Sun et al. 2013; Barnett et al. 2017).
In physical activity self-reports, a person is asked, in a questionnaire, about his or her physical activities performed over a past period. Among older adults,
21
commonly used questionnaires are the International Physical Activity Questionnaire (IPAQ) (IPAQ 2018), Community Healthy Activities Model Program for Seniors (CHAMPS) (Stewart et al. 2001) and Physical Activity Scale for the Elderly (PASE) (Washburn et al. 1993). A single-item question with response options on the duration and intensity of the physical activity of interest, based on work by Grimby (1986), has been used in assessing PA levels in older adults (Portegijs et al. 2016). While questionnaires are relatively cost-efficient, able to capture information on PA performed in different intensity categories, and especially suited for measuring PA at the group level, they have the disadvantages of reliance on recall of past PA, especially light or moderate intensity PA, and the potential for participant misunderstanding of items (Sylvia et al. 2014).
Physical activity measurements are performed using wearable devices, such as an accelerometer, a Global Positioning System (GPS) tracker, or some other type of activity recorder. Accelerometers record the magnitude of acceleration resulting from bodily movements and are thus suitable for producing time-stamped data on the frequency, duration, and intensity of PA (Strath et al. 2013). Depending on the accelerometer type, their output may be either accelerometer raw data, commonly using gravity derived ‘g’ as the unit of measurement, or readily classified physical activity data. In the case of readily classified data, the accelerometer applies inbuilt calculation methods and intensity threshold values to process the raw data and create output data, which typically expresses duration at each physical activity intensity level in activity counts, whereas in the case of accelerometers which provide raw data, self- defined algorithms for processing the data can be used (Strath et al. 2013). In accelerometers providing output data classified by intensity level, the threshold values are commonly based on expected performance levels among adults and are the same for all users of the same accelerometer type (Karavirta et al. 2020).
However, the threshold values applied may vary widely between different accelerometer types, resulting in considerable differences in measurement results (Gorman et al. 2014). When absolute values are used as thresholds for PA intensity categories, it has been shown that accumulating higher amounts of higher intensity PA is easier for younger than for older persons (Kujala et al. 2017).
Among older people, age-related decline in physical function manifests in the increasing effort required to complete a task. Thus, a physical activity task requiring a lower level of intensity among younger people requires a higher intensity among older people (Fleg et al. 2005). To categorise physical activity intensity among older adults, new methods of accelerometer-measured physical activity which also include measures of relative intensity have been proposed (Karavirta et al. 2020). The popularity of accelerometers is based on their ability to record large amounts of PA data, including intensity categorised data, in free- living conditions. Their weaknesses include high cost, the need for data processing expertise, lack of standardised threshold values, and inability to provide any contextual information (Sylvia et al. 2014).
22
Physical activity is commonly quantified as energy expenditure expressed as kilocalories or metabolic equivalents (MET) hours, or as minutes of PA at different intensity levels, or as number of bouts of continuous PA of minimum duration, e.g., bouts of 10 minutes (Sylvia et al. 2014). Of these, PA duration at different intensity levels is especially well suited for use in studies in which physical activity is related to measures of health and physical functioning, as physical activity guidelines, designed to promote PA for disease prevention and health, are given in these units. The national guidelines for physical activity in the United States (2018 Physical Activity Guidelines Advisory Committee 2018), which has also served as the basis for the Finnish national guidelines (UKK- institute 2019) and the WHO guidelines on physical activity and sedentary behaviour (WHO 2020), advise older adults to do aerobic physical activity for 150 to 300 minutes at moderate or 75 minutes at vigorous intensity weekly and, whenever possible, engage in light physical activity and avoid longer sedentary periods. The recommendations also advise strength and balance training at least two times a week and exercises aimed at maintaining and improving mobility.
However, even if the recommended amounts of physical activity are not reached, more modest increases in activity times are beneficial for health, especially for those who have basically led sedentary lives (2018 Physical Activity Guidelines Advisory Committee 2018). The main difference between the present and previous guidelines, dating from the year 2008, stemming from new evidence on the health benefits of bouts of physical activity of any duration, is that bouts shorter than 10 minutes are now also counted in the weekly total of PA minutes (2018 Physical Activity Guidelines Advisory Committee 2018). Based on datasets collected between the years 2012 and 2014, in the age group of 65 years and above, one person out of three in the USA (Keadle et al. 2016), one out of two in Europe and in Israel (Marques et al. 2015), and one out of five in Finland (WHO 2015) met the previous recommendations for aerobic PA. In all these regions, older people aged 65 years and above were less physically active than younger people (Marques et al. 2015; WHO 2015; Keadle et al. 2016). In Finland, the aging-related decreasing trend in PA has also accelerated among older people (Sun et al. 2013), as 27% of those aged 65 to 74 years but only 12 % of those aged 75 or above met the recommendations for aerobic physical activity (Bennie et al. 2017).
2.2.2 Outdoor physical activity in old age
Walking is the most commonly reported leisure-time physical activity and the most common form of outdoor activity among people aged 65 years and older (Official Statistics of Finland 2009; Dai et al. 2015). Evaluated in the year 2009, walking accounted for 48% of total time spent in sports and outdoor activities among Finnish people in this age group (Official Statistics of Finland 2009). Daily activities form a large share of older adults’ PA (Colley et al. 2019). Reviewing data from Australia, Hong Kong, and the United Kingdom, Cerin et al. (2017) found that shopping, running errands, and engaging in recreational and social activities were the most frequently reported reasons for trips outside of home.
Among Finnish older adults, shopping, walking for exercise, social visits, and
23
running errands have also been reported as the most common reasons for going outdoors (Tsai et al. 2016). Among older adults, walking to a daily destination typically takes slightly longer than 10 minutes (Perchoux et al. 2019). Thus, higher PA may be accumulated from several activity patterns. High PA volumes were observed among older adults who spent a lot of time running daily errands and among those who frequently took recreational walks (Liu et al. 2020). Walking to a destination or for recreation at least initially, takes places in the home neigh- bourhood. Hence, most older adults’ outdoor PA occurs in the vicinity of their homes (Chaudhury et al. 2016). For the accumulation of higher intensity PA, moving in and even beyond the home neighbourhood is especially important (Portegijs et al. 2015).
One study reported that PA seems to vary more between weekdays and weekend days than between weekend days, and that both walking time and the number of walking episodes were higher on weekdays than weekend days among older adults (Abel et al. 2019). Another study reported that older adults were most active on weekdays, then on Saturdays, and least active on Sundays, the difference in the volume of PA on weekdays being statistically significantly higher than the volume on Sundays (Davis et al. 2011). Having different routines on different days of the week has been proposed as the reason for the differing accumulation of PA on different days of the week (Abel et al. 2019).
2.2.3 Associations between physical activity and walking difficulty
In older adults, better physical performance in the lower extremities associates positively with more daily PA (Rapp et al. 2012; Spartano et al. 2019). The prevalence of walking difficulty often coincides with low levels of physical activity, although some older people with mobility limitations may be physically active while others with intact mobility may be sedentary (Hirvensalo et al. 2000).
In the early phases of mobility decline, walking modifications, e.g., using an aid or resting during a walk, may provide a means to continue being physically active outdoors (Skantz et al. 2020a). Nevertheless, older people with mobility disabilities make fewer trips outside home (Shumway-Cook et al. 2002) and are less physically active (Rosenberg et al. 2011) than those without mobility disabilities.
One might conclude that being physically active is related to physical functional capability (Anton et al. 2015). Although older people can only be physically active within the limits of their physical capacity, there is evidence that habitual physical activity and systematic exercise training predict favourable trajectories in physical functioning and mobility. Regular aerobic activity and physical exercise training programs reduce risk for functional limitations and disability in older people (Paterson & Warburton 2010). Walking one mile on 4 to 7 days a week compared to not having a habit of walking such a distance was shown to reduce the risk for lower body disability by at least one third over a six- year period among persons over 70 years of age (Clark 1996). Additionally, through participation in physical training programmes, older adults have been able to improve walking speed (Anton et al. 2015). Regular walks outdoors help
24
those who already have experienced some decline in their physical functioning maintain their walking capability (Simonsick et al. 2005).
2.3 Physical environments
2.3.1 Planning of physical environments
On the societal level, the most important tool for planning physical environments is land use planning, in which decisions on land use are made in the presence of different interests and needs, and which takes place at several hierarchical spatial levels, each of which serves specific aims while being connected to one other (OECD 2017). In the OECD countries, either national governments or federal states lay down the planning systems (OECD 2017). In Finland, the Land Use and Building Act (1999) describes the land use planning system and the planning process. In addition, national land use guidelines stipulate specific aspects to be considered in land use planning. Currently in Finland, the Land Use and Building Act (1999) and the national land use guidelines aim at achieving a good living environment and sustainable development through land use, and ensuring open communication and possibilities for citizen participation during land use planning processes. In Finland, the land use planning is done both on the regional and local levels (Land Use and Building Act 1999). Regional plans are drafted and approved by regional councils and local master and detailed town plans by municipal councils. Individual citizens and non-governmental organizations can participate in the planning processes at several stages. At the regional level, land use is planned typically for one or a few specific themes at the same time, e.g., for transport and energy, covering the whole regional area. Regional plans provide a general framework, which is followed by the municipalities in the region when they create local master plans. A master plan typically presents a general plan for all land uses in one municipality and thus determines, e.g., the urban structure by defining locations for residential areas and traffic routes. A detailed local plan or a town plan is in the smallest spatial scale. In such plans, detailed characteristics of physical living environments are defined at the local level, e.g., the locations, sizes, and uses of buildings and recreation areas. Additionally, a municipality may also have one or several detailed lake shore or coastal plans.
In land use planning, geographic information systems (GIS) are widely used to process and visualise map-based data. In GIS, features of the environment with their coordinates are available in geospatial datasets, thus enabling the running of numerical analyses with the data. In addition to spatial references, geospatial data are time-stamped, and thus describe the environment at a specific location at specific point in time. However, geospatial datasets may differ in their level of detail on environmental data as well as in spatial resolution.
Matching the specificity of geospatial data, including the environmental, spatial, and temporal aspects with the specific use purpose of the data is nevertheless essential.
25
2.3.2 Objective and subjective features of the physical environment
Neighbourhood and street audits, produced using observational techniques with predefined evaluation criteria, and geospatial datasets, based on measurements made with different technologies, are referred to as objective environmental data (Orstad et al. 2017). In other words, objectively assessed environmental data should model the actual environment unambiguously and independently of the data producer. Subjective environmental data is based on individual’s self- reports of his or her perceived environment. It is a cognitive representation of the environment based on information captured through all the senses, and it is affected by personal and social factors as well as past experiences, which explains why different persons may perceive the same environment and its features differently (Orstad et al. 2017).
It is quite common that low to moderate agreement will be observed between comparable objective and subjective measures of the environment (Orstad et al. 2017). Among older adults, the likelihood for a mismatch between measures of perceived and objective environment is higher among people with lower socioeconomic status, low self-efficacy, a low level of PA, those using neighbourhood facilities less often, and those who have lived in the neighbourhood for less than two years (Ball et al. 2008). Additionally, individuals with different physical capabilities and PA behaviours may perceive the features in their environment differently (Sakari et al. 2017; Herbolsheimer et al. 2020).
The features of physical environment most commonly addressed in PA research concern the built environment. The availability of services and destinations of various types, features of the pedestrian infrastructure, traffic safety, walkability, residential density, and street connectivity are commonly studied PA correlates among older adults (Barnett et al. 2017; Cerin et al. 2017;
van Cauwenberg et al. 2018). For most of these, both objective and subjective assessments have been widely applied. Walkability has predominantly been studied objectively while safety has more commonly been subjectively assessed (Barnett et al. 2017; Cerin et al. 2017; van Cauwenberg et al. 2018). Walkability is a composite index calculated from residential density, intersection density, and the evenness of land used for residential, commercial, and office purposes (Frank et al 2005). Parks and public open spaces are features of the natural environment that are often studied as one type of destination (Barnett et al. 2017; Cerin et al.
2017; van Cauwenberg et al. 2018). However, while it has been suggested that instead of proximity measures, qualitative aspects of natural and green areas should be considered together with PA (Ekkel & de Vries 2017), established practises to capture qualitative characteristics with objective or subjective assessment remain lacking.
2.3.3 Neighbourhood areas in assessing environmental features
Environmental features are typically evaluated based on their presence, magnitude, or qualities in the neighbourhood area. A person’s neighbourhood area is commonly defined based on his or her home location in a specific
26
administrative and census area, as a buffer area with a radius around the home of 400 to 500 m, or as the distance walked in 10 to 20 minutes (Barnett et al. 2017;
Cerin et al. 2017). Road network buffers, i.e., the area including roads and pavements up to a specific distance from the home, have been used to define neighbourhoods (Frank et al. 2017). However, in practice, the sizes and shapes of areas perceived as a neighbourhood will differ between individuals. A study among adults showed that the size of the self-defined neighbourhood area was smaller among the older participants and women, and among those with less education, shorter duration of residency, lower perceived attachment to place, and living in an area of higher residential density (Charreire et al. 2016). In turn, a study conducted in Australia suggested not using different sized buffers for different age groups. The authors found that the median distance older adults walked to shops or services was 0.76 km and to natural features 0.97 km (Sugiyama et al. 2019). Thus, the size of a person’s neighbourhood area may differ owing to types of environmental features and PA domain. For example, the features of a pedestrian infrastructure important for transport PA may be identified by focusing on the vicinity of the home, whereas investigating attractive destinations for leisure-time PA might mean examining a more extended area (Van Cauwenberg et al. 2018; Sugiyama et al. 2019). Similarly, to detect significant environment-PA associations, depending on the nature of the environmental feature(s) in question, it is important to capture environmental barriers close to home while for environmental facilitators the neighbourhood needs to be extended (Portegijs et al. 2020).
The use of static neighbourhood definitions, such as a circular or road network buffer or census tract, has been criticised for not capturing the actual environments used by a person, i.e., not providing information on the person’s actual exposure to different environmental features (Laatikainen et al. 2018).
Evaluating environmental features within activity spaces, that is, the actual area within which individual moves, has gained in popularity (Laatikainen et al. 2018).
However, only considering the environmental features that are most likely to be encountered, tells only one part of the story of the associations between neighbourhood environmental features and PA. Within their radius, circular buffers capture environmental features more distant from roads and areas equally around the home, thus most likely also covering areas outside the individual’s activity space, roads, and roadsides. Including these areas is important, as it is possible that environmental features present or missing in them may also be important when individuals make mobility decisions and thereby affect their PA behaviour.
Neighbourhood types within urban structures, meaning types of residential areas that are separable from one another based on their characteristics, are commonly categorised into urban and rural (Lee & Park 2015) or urban, suburban, and rural areas (Hanibuchi et al. 2011; Maisel 2016). Typically, differences between neighbourhood types are based on a specific objective characteristic or on a combination of these. Neighbourhoods have been categorised, for example, based on their residential or household density (Li et al. 2005; Troped et al. 2014),
27
intersection density (Li et al. 2005), walkability (Bracy et al. 2014; Orstad et al.
2018), and areas of green and open space (Li et al. 2005). Subjectively assessed walkability (Merom et al. 2015), distance to services (Van Cauwenberg et al. 2013), and pedestrian safety (Bracy et al. 2014) have also been used in categorising neighbourhoods.
2.4 Associations of features of the physical environment with walking difficulty and physical activity among older adults
2.4.1 Associations between environmental features and walking difficulty Studies investigating the associations between the environment and walking difficulty have most commonly been based on perceived features of the environment (Balfour & Kaplan 2002; Rantakokko et al. 2012; Eronen et al. 2013).
Subjective assessments of the environment reflect, besides the actual environment the individual’s interpretations, past experiences, and functional capabilities. Associations between objectively defined features of the environment and physical functioning have been studied less often. In the few existing studies, higher objectively defined greenness in the neighbourhood area and proximity to natural environments with green or blue spaces has been associated with slower decline in physical functioning in old age (de Keijzer et al.
2019).
Older people with signs of decline in their walking perceive a higher number of nature- and infrastructure-related outdoor mobility barriers and fewer facilitators in their home neighbourhood compared to those with no signs of walking decline (Skantz et al. 2020b). Among older adults with intact walking ability, a higher number of environmental facilitators decreases the risk of developing walking difficulty (Eronen et al. 2013). More specifically, lack of resting places and long distances to destinations are associated with increased risk for developing difficulty in walking 500 m (Rantakokko et al. 2012). However, it seems that walking modifications, e.g., resting in the middle of a walk or walking more slowly, may help to reduce the risks for incident mobility difficulty presented by environmental barriers (Rantakokko et al. 2016).
2.4.2 Associations between environmental features and physical activity Recent systematic reviews and meta-analyses have provided evidence that several features of the physical environment correlate with total PA (Barnett et al. 2017), walking for transport (Cerin et al. 2017), and leisure-time PA in older adults (Van Cauwenberg et al. 2018). Environmental features that correlate with higher PA include access to various types of destinations, especially shops and other commercial destinations (Barnett et al. 2017; Cerin et al. 2017), and park/
open space/recreational destinations (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018), public transport (Barnett et al. 2017; Van Cauwenberg
28
et al. 2018), and a pedestrian-friendly infrastructure (Barnett et al. 2017; Cerin et al. 2017). Moreover, higher PA is more common in neighbourhoods with higher residential density (Cerin et al. 2017) and walkability among older people (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018).
Associations of PA with less frequently studied features of the physical environment, such as hilliness, or more detailed measures of PA, such as MVPA or active bouts of 10 minutes, have not yet been reported in reviews. In separate studies, hilly terrain in the neighbourhood was associated with less time spent in walking (Gómez et al. 2010), total PA (Tanaka et al. 2016), and recreational PA (Wu et al. 2019) among older adults. However, some studies have reported for hilliness no associations with leisure-time PA (Ribeiro et al. 2013) and a positive association with walking activity (Hanibuchi et al. 2011; Abe et al. 2020) and MVPA (Abe et al. 2020). Positive associations with MVPA have been shown for street connectivity, residential density, and mixed land use, either as separate components or combined to form a walkability index, (Carlson et al. 2012; Ding et al. 2014; Van Cauwenberg et al. 2016; Van Holle et al. 2016; Colley et al. 2019), although not consistently (Timmermans et al. 2016; Amagasa et al. 2019; Chen et al. 2019; Pettigrew et al. 2020). In addition, the closeness of parks (Thornton et al.
2017) and density of recreation facilities (Cerin et al. 2016) have shown positive yet also non-significant (Carlson et al. 2012; Ding et al. 2014) associations with MVPA. A positive association of neighbourhood walkability with MVPA bouts of at least 10 minutes duration has also been reported (Amagasa et al. 2019).
Both subjective and objective assessments of environmental features have been used in PA studies (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018). Generally, subjectively assessed environmental features have been slightly more often associated with physical activity than objectively assessed environmental features (Barnett et al. 2017; Orstad et al. 2017). When the same environmental feature has been operationalised as both a subjective and an objective environmental variable, the variables have only seldom been associated with the same PA outcome (Orstad et al. 2017). It seems that presence of a feature in the environment alone is not sufficient to increase PA. The feature needs to be perceived as useful in order to correlate with a higher level of outdoor PA (Carlson et al. 2016). Therefore, instead of considering interchangeability between objective and subjective environmental measures, these two constructs might complement each other when examining the associations of environmental factors with behaviour (Nyunt et al. 2015), as they may contribute differently to PA (Orstad et al. 2017).
Environment-PA associations are sensitive to the distance used to define a given neighbourhood area. Reporting environmental mobility barriers close to home increases the risk for low volumes of PA (Portegijs et al. 2020), whereas greenness (Browning & Lee 2017; Ekkel & de Vries 2017) and attractive destinations (Portegijs et al. 2020) at longer distances from home correlate with higher volumes of PA. Higher walkability in the immediate vicinity to home as well as in the larger neighbourhood area contribute to more walking activity among older adults (Villanueva et al. 2014). Different ways of operationalising a
29
neighbourhood area, varying use of assessment methods for environmental features, and possible modifying effects, which have thus far not been studied extensively enough, may underlie the current inconsistencies in study findings for many environmental features (Cerin et al. 2017).
2.4.3 Contextual factors in environment - physical activity associations According to the social ecological approach, PA behaviour results from interaction between several internal and perceived factors and factors external to a person (Sallis et al. 2006). Recent reviews of the environmental correlates of domain-specific PA investigated this interplay by analysing modifying factors in environment-PA associations but found no conclusive evidence of any individual or environmental characteristics or neighbourhood definition that consistently modified these associations (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018). However, the reviews presented a few suggestions on potential modifying factors. Although reported in only a small number of studies, the presence of a pedestrian-friendly infrastructure and safety in outdoor mobility were reported to have appeared as significant modifying factors revealing positive associations between recreational facilities and leisure-time PA (Van Cauwenberg et al. 2018). Additionally, mobility capability and relevant destinations were suggested as potential modifying factors in the associations between walking infrastructure and active travel (Cerin et al. 2017).
Several studies on personal contextual factors have shown that physical functioning plays a significant role in environment-PA associations. These studies indicate that a person’s functional capacity affects how features in the environment are perceived (Moura et al. 2017; Sakari et al. 2017; Herbolsheimer et al. 2020) and how perceived (Satariano et al. 2010; Gallagher et al. 2012;
Haselwandter et al. 2015; Levasseur et al. 2015) and objective environmental features (Satariano et al. 2010; King et al. 2011; Gong et al. 2014) are associated with PA. Perceptions of environmental features as supporting outdoor mobility are especially important for mobility choices among older adults with functional or cognitive impairments (Brookfield et al. 2017). Mobility limitations may also affect the use of environment for outdoor mobility. Older adults with mobility limitations may plan their routes so that they encounter fewer challenging physical features of the environment during trips into the community compared to those without mobility limitations (Shumway-Cook et al. 2003). Further, features of the physical environment that facilitate walking seem to be more strongly associated with PA among older adults with chronic conditions, such as hearing and vision impairments and musculoskeletal and genitourinary diseases, than among those without these conditions (Barnett et al. 2016).
In the few studies that have reported on external contextual factors, positive associations between recreational facilities and leisure-time PA have been more apparent in environments with a good pedestrian infrastructure and in environments categorised as safe than in environments where these features were absent (Van Cauwenberg et al. 2018). Environment-PA associations have also been reported to differ between neighbourhood types categorised as urban
30
vs. rural (Lee & Park 2015), urban vs. suburban vs. rural (Maisel 2016) or according to their residential density (Troped et al. 2014) or walkability (Bracy et al. 2014). It has been pointed out that geographical location should also be considered when interpreting and generalizing results on environment-PA associations. The range within which environmental features are present in the study area in general as well as cultural issues related to the environment and PA behaviour may vary greatly depending on the geographical location (Cerin et al. 2017). Additionally, European cities typically differ from North American or Australian cities in residential density and urban structure (Shipperijn et al.
2013).
It seems that age does not modify environment-PA associations, as they appear rather similar among older adults and among adults under 65. Among adult populations, residential density, intersection density, public transport density, and number of parks in the neighbourhood have shown positive associations with MVPA minutes (Sallis et al. 2016). These environmental features also correlate with higher volumes of PA among older adults (Barnett et al. 2017; Cerin et al. 2017; Van Cauwenberg et al. 2018). A study comparing different age groups found no effect of age on the likelihood of being physically active in the nearest urban green space (Schipperijn et al. 2013). Walkability has shown positive associations with MVPA among adults aged 18 to 39, 40 to 59, and 60 to 79 (Colley et al. 2019). Median distances walked to destinations were observed to be shorter among older adults than among 50- to 64-year-olds, but the differences measured in metres were so small that similar-sized buffers were found suitable for research in all age groups (Sugiyama et al. 2019). Further, in younger as in older adults, the volume of PA relates to how features of the environment are perceived (Wallmann-Sperlich et al. 2014).
2.5 The social ecological approach to studying physical activity Relationships between environmental features and older adults’ physical activity can be conceptualised using a social ecological approach (Stokols 1992). Models constructed using this approach provide holistic ways to examine the interplay of individual, social, environmental, and policy features with physical activity.
Social ecological models were originally mainly applied in sociological research but were subsequently introduced in physical activity research to increase knowledge on environmental factors related to physical activity behaviour (Stokols 1992). The need for this had emerged after it was realised that individual-focused health promotion programs had limited capacity to influence physical activity levels at the population level. Instead, it was thought that measures taken to enhance environments to better support physical activity might yield better results, as they would reach a higher number of people simultaneously (Stokols 1992). In 2005, to increase the predictive capacity of social ecological models in environment-PA studies, Gilles-Corti et al. (2005b) called for better correspondence between measures of the environment and
