UBC Faculty Research and Publications

Generating GPS activity spaces that shed light upon the mobility habits of older adults: a descriptive… Hirsch, Jana A; Winters, Meghan; Clarke, Philippa; McKay, Heather Dec 12, 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12942_2014_Article_624.pdf [ 719.55kB ]
JSON: 52383-1.0223419.json
JSON-LD: 52383-1.0223419-ld.json
RDF/XML (Pretty): 52383-1.0223419-rdf.xml
RDF/JSON: 52383-1.0223419-rdf.json
Turtle: 52383-1.0223419-turtle.txt
N-Triples: 52383-1.0223419-rdf-ntriples.txt
Original Record: 52383-1.0223419-source.json
Full Text

Full Text

RESEARCH Open AccessGenerating GPS activity spoHeinenanlives [2] and is central to them conducting commercial,cultural, and social activities [3-5]. Older adults’ well-beinglocal neighborhood, or beyond) was positively associatedwith diminished cognitive decline and Alzheimer’s DiseaseINTERNATIONAL JOURNAL OF HEALTH GEOGRAPHICSHirsch et al. International Journal of Health Geographics 2014, 13:51http://www.ij-healthgeographics.com/content/13/1/51Burnaby, British Columbia V5A 1S6, CanadaFull list of author information is available at the end of the articleand quality of life are also closely linked to their mobility[5-8]. The demographic shift toward an aging population isunprecedented in western society and demands novel solu-tions that evaluate and promote the mobility of olderadults. These solutions may be embedded within transpor-tation systems and planning [9-13].[18,19], lower risk of both death [20] and becoming morefrail [21]. Older age, being female, and having physicallimitations have been associated with smaller life-spaces,as is having had a stroke, high depressive symptoms, andbeing obese [22,23]. Higher education, better lower ex-tremity function and muscle strength were associated lar-ger life-spaces [22,23]. Importantly, the ability to driveplays a key role in the mobility of older adults as capturedusing life-space measures [23-26].* Correspondence: mwinters@sfu.ca3Faculty of Health Sciences, Simon Fraser University, 8888 University Drive,Daily Path Area) using GPS data from 95 older adults in Vancouver, Canada. Calculated activity space areas andcompactness were compared across sociodemographic and resource characteristics.Results: Area measures derived from the three different approaches to developing activity spaces were highlycorrelated. Participants who were younger, lived in less walkable neighborhoods, had a valid driver’s license, hadaccess to a vehicle, or had physical support to go outside of their homes had larger activity spaces. Mobility spacecompactness measures also differed by sociodemographic and resource characteristics.Conclusions: This research extends the literature by demonstrating that GPS tracking can be used as a valuabletool to better understand the geographic mobility patterns of older adults. This study informs potential ways tomaintain older adult independence by identifying factors that influence geographic mobility.Keywords: Global positioning systems (GPS), Geographic information systems (GIS), Activity space, Mobility,Neighborhood attributes, Older adultsBackgroundMobility is defined as the “ability to move oneself (eitherindependently or by using assistive devices or transpor-tation) within environments that expand from one’s hometo one’s neighborhood and regions beyond” [1]. Mobility iskey to older adults leading active, healthy, independentIt is essential to identify effective tools to describe olderadult mobility so as to better understand the influence ofneighborhood on their health and mobility [14]. “Life-space” is a frequently used measure of older adult mobility[15-17]. This self-reported measure of the extent of recenttravel (using thresholds such as: within the home, into theupon the mobility habitsa descriptive analysisJana A Hirsch1,2, Meghan Winters3*, Philippa Clarke4 andAbstractBackground: Measuring mobility is critical for understandfunctioning. Global Positioning Systems (GPS) may represcompare mobility patterns in older adults.Methods: We generated three types of activity spaces (St© 2014 Hirsch et al.; licensee BioMed Central.Commons Attribution License (http://creativecreproduction in any medium, provided the orDedication waiver (http://creativecommons.orunless otherwise stated.aces that shed lightf older adults:ather McKay5g neighborhood influences on older adults’ health andt an important opportunity to measure, describe, anddard Deviation Ellipse, Minimum Convex Polygon,This is an Open Access article distributed under the terms of the Creativeommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andiginal work is properly credited. The Creative Commons Public Domaing/publicdomain/zero/1.0/) applies to the data made available in this article,Hirsch et al. International Journal of Health Geographics 2014, 13:51 Page 2 of 14http://www.ij-healthgeographics.com/content/13/1/51Global Positioning Systems (GPS) technology may pro-vide a means to calculate geographic range as a measureof mobility. GPS has been used to objectively characterizelife-space [27-32] and also to detect outside physical activ-ity [33]. Most research using GPS for older adult mobilitycomes from one study—the Senior Tracking (SenTra)project, based in Germany and Israel. SenTra used GPSpoints or metrics of out-of-home behaviors (e.g. numberof visited places, time spent outside the home, and dis-tance traveled from home) to assess mobility patterns in asample of older adults who were cognitively impaired orhad Alzheimer’s disease, and compared this clinical groupwith community-dwelling older adults [34-47]. It maynot be possible to generalize outcomes from SenTraacross geographically and culturally diverse settings, orwith older adults who are independent and able to maketheir own travel decisions. Therefore, there is a need to in-vestigate GPS applications in community-dwelling olderadults in the North American context, given the limitedknowledge base on measurement approaches to define thegeographic patterns (i.e. shape) and extent (i.e. size) ofolder adult mobility in this population. Additionally, thepotential to characterize older adults’ geographic extent orpattern was not explored in any of these studies, leavingquestions about where older adults travel.To represent daily mobility, neighborhood studies ofphysical activity have used GPS-based “activity spaces” asan individual-based measure of spatial behavior [48-52].Activity spaces differ from the life-space measure, in thatthey focus on neighborhood (out of home) behavior only,rather than mobility both within and beyond the home.Additionally, with the exception of recent efforts toincorporate GPS [27-32], the majority of life-space studiesare based on self-reported travel extents (within the home,into the local neighborhood, or beyond), not spatially-located travel data. Thus, the expansion of activity spacesto the investigation of older adult mobility will give add-itional insight into the community factors and resourcesthat shape neighborhood activity. It is hypothesized thatactivity spaces may vary in size and shape across differentpopulations, such as those with low incomes or differentage groups [50,53]. To date the utility of GPS to measure,describe, and compare mobility patterns in older adultshas not been fully explored. Previous work proposed usinga “mobility envelope”, the length of the outer perimeter ofspatial excursions made by individuals, as an outcomemeasure for mobility studies in older adults [54]. By inte-grating various metrics of geographic extent from activityspace studies, the mobility envelope concept could be ad-vanced. For example, additional metrics evaluating differ-ent dimensions of individuals’ geographic scope may beuseful for understanding mobility. In particular, in urbanplanning the shape or “compactness” of activity spaces is ametric of how circular a polygon is and is a conceptthought to illustrate the capacity of neighborhoods to pro-vide opportunities to “live, work, shop, and socialize at thelocal scale” [53]. Specifically, compactness has been shownto vary across different travel modes [55]. However, com-parisons of compactness among different individuals (andwithin a particular type of activity space) may reveal im-portant information on the role of driving in older adultmobility.Therefore our objectives are twofold: to create andcompare different types of geographic activity spaces forcommunity dwelling older adults, so as to clarify the ex-tent and pattern of their mobility; and then within eachactivity space approach, to assess individual sociodemo-graphic and resource characteristics that are associatedwith larger (or smaller) or more (or less) compact activityspaces. We hypothesize that the size and shape of activityspaces will vary by the approach used to create them.Within each activity space type we hypothesize that par-ticipants who are younger, healthier, and with better ac-cess to transportation (driving or material resources forgoing outside) will have larger activity spaces. Further-more, we hypothesize that those who drive less will havemore compact activity spaces, and also that participantswho live in more walkable neighborhoods will have morecompact activity spaces, since amenities may be closer tohome.ResultsParticipants had GPS data for a mean of 3.5 days (stand-ard deviation (SD) 1.7 days; median 3.0 interquartilerange (IQR) 3.0) with a mean of 14656.3 GPS points (SD10232.8; median 12132, IQR 12401). Participants traveledbetween 1 and 12 trips each day they were tracked, with amean of 13.2 total trips (SD 7.8; median 13.0, IQR 12.0)per participant.The areas derived from the three different approaches todeveloping activity spaces were highly associated; correl-ation coefficients ranged from ρ = 0.96 (Standard DeviationEllipse (SDE) area vs Daily Path Area (DPA) area) toρ = 0.98 (Minimum Convex Polygon (MCP) area vs DPAarea and MCP area vs SDE area). However, values for com-pactness varied greatly between approaches; although com-pactness values derived from MCP and SDE approacheswere highly correlated (ρ = 0.82, p < 0.0001), compactnessvalues derived using DPA were not correlated with eitherMCP or SDE approaches (ρ = 0.07 p = 0.49 and ρ = −0.01p = 0.92, respectively). Activity spaces were larger than thetraditional buffers (200-meters or 1/8-mile; 400-meters or¼-mile; 800-meters or ½-mile) used for neighborhood re-search (Figure 1).In terms of size, DPA generated the smallest and MCPthe largest (Table 1) activity space areas. Patterns for ac-tivity space area by sociodemographic group and re-source characteristics were consistent across approaches.1 TyntiHirsch et al. International Journal of Health Geographics 2014, 13:51 Page 3 of 14http://www.ij-healthgeographics.com/content/13/1/51Activity space areas were generally larger for youngerparticipants, those in less walkable neighborhoods, those4.43.310.57.822.30510152025SDE MCPPercent of Activity Space Covered by Residential BufferActivity SpaceFigure 1 Percentage of activity space covered by traditional residewith valid driver’s licenses, those with access to a vehiclein the past 7 days, and those with physical support to gooutside.Compactness values generated using SDE and MCPapproaches were much higher than were DPA-generatedvalues (Table 2), signifying more circular shapes by theseapproaches. Men had more compact SDE and evidenceof more compact MCP, compared with women. Com-pactness values generated using SDE and MCP approacheswere higher in areas with lower walkability (as mea-sured by Walk Score), while compactness assessed usingthe DPA approach was higher in areas of higher walk-ability. Participants who had a valid driver’s license,and had access to a vehicle in the past 7 days had lesscompact values generated using the DPA approach. Par-ticipants who very much liked to walk outside had lowervalues for compactness if generated using SDE or MCPapproaches.Log-linear models demonstrated that participants wholived in less walkable neighborhoods, who had access to avehicle or who had physical support to go outside had sig-nificantly larger activity spaces (Table 3). Linear modelsdemonstrated limited associations between sociodemo-graphic or resource characteristics and compactness ofactivity space. Only three sociodemographic or resourcecharacteristics were associated with compactness of activ-ity spaces: living in walkable neighborhoods, liking to walkoutside, and access to a vehicle. Using the SDE and MCPapproaches, participants who lived in somewhat walkable3.39.26.819.2DPApe200m Buffer400m Buffer800m Bufferal buffers (200-meter, 400-meter, 800-meter).neighborhoods had higher values for compactness andthose who reported they very much like to walk outsidehad less compact activity spaces. Participants living in lesswalkable neighborhoods and with access to a vehicle hadless compact DPA-generated activity spaces.DiscussionThis research is one of the first studies to utilize GPStracking to create activity spaces as a means to assessolder adult mobility. We extend the literature by demon-strating that GPS tracking can be used to create threedifferent types of activity spaces as a valuable tools to bet-ter understand the geographic mobility patterns of olderadults. Not surprisingly, older adults deemed most mobilebased on their age, the walkability of their neighbor-hoods, and whether or not they drove, had the largest ac-tivity spaces. The trends in activity space areas were notdifferent, regardless of what approach was used to gener-ate them. However, shape of activity spaces (measured ascompactness) varied by approach.Walking and cycling trips are often extend beyond trad-itional buffer sizes (1 mile, ½ mile) used to represent neigh-borhoods where older adults live [56]. We demonstratedthat GPS may enable a more precise way to operationalizeneighborhoods than residential buffers or administrativeunits [51,57-62] as it better captures the locations individ-uals actually visit rather than a presumed neighborhoodTable 1 GPS-based activity space areas of Walk The Talk Study participants (n = 95) by sociodemographic group andresource characteristicsSociodemographic group n SDE area (in hectares) MCP area (in hectares) DPA area (in hectares)Median (IQR) pa Median (IQR) pa Median (IQR) paAll 95 1121.9 (3900.8) 1753.6 (7097.9) 837.2 (1389.1)Sex 0.3467 0.4616 0.6060Female 63 1183.8 (4903.2) 1753.6 (7256.0) 837.2 (1603.6)Male 32 972.7 (2355.6) 1752.6 (4974.1) 839.8 (1147.7)Age (years) 0.0289 0.0617 0.129365-69 26 3071.1 (8827.8) 4764.5 (12583.2) 1448.6 (1588.1)70-74 29 1909.6 (3362.5) 1952.4 (6744.7) 958.4 (1150.5)75-79 26 680.5 (1423.0) 1031.3 (2325.3) 566.5 (937.2)80+ 14 483.2 (1397.3) 791.8 (2624.7) 532.6 (548.3)Race 0.2727 0.3351 0.4617Non-white 18 693.5 (5189.3) 1163.0 (5545.7) 786.4 (1486.7)White 77 1379.6 (3539.9) 1896.0 (6854.3) 837.2 (1353.0)Education 0.8895 0.7534 0.8814Secondary school or less 26 982.5 (2522.6) 1456.0 (2396.1) 846.1 (838.7)Some or completed trade/technical school or college 36 1652.6 (7491.1) 2837.4 (9796.7) 949.9 (1853.6)Some university or higher 33 887.0 (3896.6) 1569.1 (5537.3) 837.2 (1192.6)Marital status 0.5785 0.8197 0.8308Not married 88 1160.3 (3735.0) 1817.2 (7119.1) 833.8 (1489.2)Married 7 616.6 (5038.3) 1317.0 (4865.3) 1035.2 (1085.0)Living with someone else 0.7248 0.9715 0.7171No 80 1129.3 (3764.0) 1731.2 (7124.3) 781.8 (1600.1)Yes 15 942.8 (4024.0) 2777.3 (4865.3) 1035.2 (1079.6)Dog ownership 0.5570 0.8752 0.9120No 84 1160.3 (4036.8) 1794.8 (7134.7) 830.2 (1396.0)Yes 11 887.0 (2538.0) 1753.6 (5488.9) 837.2 (1280.9)Walkabilityb 0.0182 0.0096 0.0116Car dependent (0–49) 19 3018.3 (16101.5) 7982.4 (23676.8) 2037.2 (2720.9)Somewhat walkable (50–69) 24 1194.6 (2898.6) 1888.4 (6268.4) 1037.9 (1355.4)Very walkable (70–89) 27 1614.0 (4074.9) 2243.0 (7064.9) 751.0 (1268.4)Walker’s paradise (90–100) 25 605.4 (1830.6) 1008.9 (2615.6) 542.4 (773.5)Length of time in neighborhood 0.4415 0.6514 0.7263≤ 2 years 27 1593.0 (4628.6) 1912.1 (7060.9) 958.4 (1592.3)Between 2 and up to 6 years 28 839.8 (5695.7) 1535.3 (9962.9) 794.1 (1476.8)Between 6 and up to 9 years 17 942.8 (3631.5) 2243.0 (6116.3) 752.4 (1498.6)> 9 years 23 770.4 (3672.6) 1096.2 (7321.2) 701.9 (1310.5)Valid driver’s license 0.0624 0.0501 0.0309No 23 617.6 (2543.3) 1053.7 (3075.3) 544.5 (1013.9)Yes 72 1218.1 (4266.3) 1904.0 (6986.1) 942.5 (1529.9)Access to a vehicle 0.0061 0.0027 0.0010No 37 617.6 (2002.8) 1008.9 (2928.5) 504.7 (900.9)Yes 56 1613.7 (4678.6) 2501.9 (7009.7) 997.4 (1511.6)Hirsch et al. International Journal of Health Geographics 2014, 13:51 Page 4 of 14http://www.ij-healthgeographics.com/content/13/1/51ud467831297368504890385904daHirsch et al. International Journal of Health Geographics 2014, 13:51 Page 5 of 14http://www.ij-healthgeographics.com/content/13/1/51Table 1 GPS-based activity space areas of Walk The Talk Stresource characteristics (Continued)Social support/companionship to go outsideNo 44 982.5 (34Yes 51 1183.8 (3Physical support to go outsideNo 51 887.0 (27Yes 44 1942.6 (8Like to walk outsidecLess than very much 28 813.5 (42Very much 67 1136.8 (3Confidence walking outsidecLess than very much 20 684.3 (69Very much 75 1183.8 (3Falls in past 6 monthsNo 75 1252.4 (3Yes 20 1004.4 (3Use of a mobility aid for walkingNo 78 1152.8 (3Yes 17 1002.7 (6Abbreviations: Interquartile range (IQR), Standard Deviation Ellipse using one stanboundary. GPS has recently become a more popular op-tion for measuring neighborhood exposure and context inhealth studies [33,50,63-67]. As the analytic approaches toGPS data develop, this technology may become a powerfultool to accurately and precisely describe people’s interac-tions with geographic space, including older adult mobil-ity. It is possible that different types of activity spacesmay be more appropriate depending on the applicationor the research question being addressed. Since the MCP-generated activity space is bounded by the outermost GPSpoints, it captures the envelope of the extreme extent oftravel and thus may include large geographic areas thatare not visited by, or important to, an individual [68]. Incontrast, SDE-generated spaces may be more useful if onewishes to assess the direction and general shape of a per-son’s travel area, without introducing potential error intro-duced by using geographically distant points. In addition,SDE approaches may indicate the frequency that an indi-vidual visits a geographic area as more points are gen-erated in that location - this ‘pulls’ the SDE-generatedactivity space toward that more often visited geographicarea. However, since SDE-generated activity spaces are bydefinition ellipse shaped, they may also capture a substan-tial area that an individual may not have visited. Finally,we recommend using DPA to generate an activity space ifa research question is about the destinations participants200-meter buffer (DPA).ap-value from Kruskal-Wallis non-parametric one-way Analysis of Variance (ANOVA)bMeasured by Street Smart Walk Score for home address.cLess than very much (1–4 on a 5-point scale); Very much (5 on a 5-point scale).y participants (n = 95) by sociodemographic group and0.5381 0.3413 0.2613.7) 1456.8 (5476.1) 741.0 (1265.3)9.5) 1896.0 (7054.8) 958.4 (1532.4)0.0537 0.0405 0.0510.0) 1305.4 (3766.4) 695.0 (893.0)6.2) 2937.3 (11791.8) 1047.8 (1757.7)0.8864 0.9967 0.8929.7) 1222.3 (6828.0) 741.7 (1469.7)3.4) 1896.0 (7124.4) 854.9 (1402.9)0.6580 0.7116 0.8444.3) 1168.8 (11618.3) 716.5 (1919.9)2.5) 1896.0 (6947.3) 849.2 (1355.9)0.8301 0.8444 0.91640.8) 1952.4 (7118.7) 854.9 (1389.1)4.5) 1596.8 (6563.9) 824.3 (1407.8)0.5248 0.5185 0.36916.6) 1888.4 (7125.3) 852.1 (1421.1)7.0) 1241.5 (9690.6) 752.4 (1556.3)rd deviation (SDE), Minimum Convex Polygon (MCP), Daily Path Area using apass in daily travel, as it relies solely on buffering theroutes actually traveled.The consistency of activity space area across the three dif-ferent approaches aligns with a previous study that showedassociations across MCP, SDE, and line-based buffers [52].However, no previous studies assessed the shape (compact-ness) of GPS-derived activity spaces. In our study, patternsof compactness differed depending on the approach usedto generate activity space. This signals the need to choosean approach to generating an activity space that is relevantto the research question being asked. Of interest, com-pactness generated using the SDE and MCP approach washigher when walkability was lower, whereas for DPA-generated compactness was higher in neighborhoods withhigher Walk Score. This finding may illustrate thatparticipants in neighborhoods with lower walkabilityare clustering their trips to a nearby retail area. Thus, inneighborhoods with lower walkability the area of all threeactivity space types becomes larger, and the compactnessof the SDE and MCP is higher. However, in this same sce-nario, DPA-generated activity spaces become more elon-gated, potentially with multiple trips heading in the samedirection, which results in a less compact DPA-generatedactivity space. Our findings using DPA are consistent withwork indicating that small and compact activity spaces in-crease the likelihood of walking and cycling [55].or Wilcoxon Rank Sum test across sociodemographic and resource categories.Table 2 GPS-based activity space compactness of Walk The Talk Study participants (n = 95) by sociodemographicgroup and resource characteristicsSociodemographic group n SDE compactness (0 to 1) MCP compactness (0 to 1) DPA compactness (0 to 1)Mean (SD) pa Mean (SD) pa Mean (SD) paAll 95 0.77 (0.15) 0.75 (0.11) 0.34 (0.21)Sex 0.0506 0.0919 0.6597Female 63 0.75 (0.15) 0.74 (0.11) 0.33 (0.21)Male 32 0.81 (0.12) 0.77 (0.09) 0.35 (0.21)Age (years) 0.9822 0.9412 0.211565-69 26 0.77 (0.14) 0.75 (0.10) 0.29 (0.21)70-74 29 0.77 (0.16) 0.74 (0.12) 0.31 (0.18)75-79 26 0.77 (0.15) 0.74 (0.11) 0.40 (0.23)80+ 14 0.79 (0.13) 0.76 (0.09) 0.38 (0.21)Race 0.3052 0.3293 0.1429Non-white 18 0.74 (0.12) 0.73 (0.11) 0.41 (0.27)White 77 0.78 (0.15) 0.75 (0.11) 0.33 (0.19)Education 0.1546 0.0172 0.4758Secondary school or less 26 0.77 (0.16) 0.74 (0.11) 0.30 (0.11)Some or completed trade/technical school or college 36 0.81 (0.12) 0.79 (0.08) 0.37 (0.27)Some university or higher 33 0.74 (0.16) 0.71 (0.12) 0.34 (0.19)Marital status 0.8909 0.4912 0.9096Not married 88 0.77 (0.15) 0.75 (0.11) 0.34 (0.21)Married 7 0.78 (0.15) 0.78 (0.10) 0.35 (0.25)Living with someone else 0.8379 0.5398 0.8590No 80 0.77 (0.14) 0.75 (0.11) 0.34 (0.21)Yes 15 0.77 (0.16) 0.76 (0.10) 0.33 (0.21)Dog Ownership 0.9119 0.7896 0.6049No 84 0.77 (0.14) 0.75 (0.11) 0.34 (0.20)Yes 11 0.77 (0.19) 0.76 (0.12) 0.37 (0.29)Walkabilityb 0.0721 0.0325 0.0244Car dependent (0–49) 19 0.77 (0.13) 0.73 (0.11) 0.24 (0.14)Somewhat walkable (50–69) 24 0.83 (0.13) 0.79 (0.08) 0.31 (0.18)Very walkable (70–89) 27 0.77 (0.14) 0.77 (0.09) 0.36 (0.23)Walker’s paradise (90–100) 25 0.72 (0.16) 0.70 (0.13) 0.42 (0.23)Length of time in neighborhood 0.0659 0.4341 0.5547≤ 2 years 27 0.77 (0.15) 0.75 (0.10) 0.29 (0.18)Between 2 and up to 6 years 28 0.76 (0.12) 0.74 (0.11) 0.36 (0.24)Between 6 and up to 9 years 17 0.71 (0.16) 0.72 (0.11) 0.34 (0.21)> 9 years 23 0.83 (0.14) 0.77 (0.11) 0.37 (0.21)Valid driver’s license 0.2126 0.9521 0.0175No 23 0.74 (0.14) 0.75 (0.10) 0.43 (0.26)Yes 72 0.78 (0.15) 0.75 (0.11) 0.31 (0.19)Access to a vehicle 0.5099 0.7018 0.0007No 37 0.76 (0.15) 0.75 (0.11) 0.43 (0.25)Yes 56 0.78 (0.14) 0.74 (0.11) 0.28 (0.16)Hirsch et al. International Journal of Health Geographics 2014, 13:51 Page 6 of 14http://www.ij-healthgeographics.com/content/13/1/51TdaHirsch et al. International Journal of Health Geographics 2014, 13:51 Page 7 of 14http://www.ij-healthgeographics.com/content/13/1/51Table 2 GPS-based activity space compactness of Walk Thegroup and resource characteristics (Continued)Social support/companionship to go outsideNo 44 0.78 (0.15)Yes 51 0.76 (0.14)Physical support to go outsideNo 51 0.78 (0.14)Yes 44 0.76 (0.15)Like to walk outsidecLess than very much 28 0.84 (0.11)Very much 67 0.74 (0.15)Confidence walking outsidecLess than very much 20 0.78 (0.13)Very much 75 0.77 (0.15)Falls in past 6 monthsNo 75 0.77 (0.15)Yes 20 0.77 (0.15)Use of a mobility aid for walkingNo 78 0.78 (0.14)Yes 17 0.74 (0.15)Abbreviations: Standard Deviation (SD), Standard Deviation Ellipse using one stanA number of factors surfaced as important to olderadults’ mobility. First, it is intuitive and supported byprevious studies that an older adult’s ability to drive willinfluence the size of their activity space [23-26], and thatthe activity space of older adult will be smaller comparedwith a younger person, on average [22]. Sex did not sur-face as a differentiating factor in our study, in contrastwith previous reports that women had smaller activityspaces than did men [22,23]. However, as only one-thirdof our sample were men it is possible we lacked the stat-istical power to test this association. Second, when anolder person is no longer able to drive, physical supportto maintain mobility within the neighborhood becomesincreasingly important. Given its apparent role in ourstudy around encouraging older adults to travel withintheir neighborhoods, physical support may represent aneffective intervention in future studies. Third, our resultssupport previous findings – that neighborhood attributes(i.e. higher street connectivity, proximity to destinations,and traffic conditions, and parks) are associated with in-creased mobility among older adults [65,69-71]. Althoughseemingly counterintuitive at first glance, our finding thatthose living in higher walkability areas had smaller activityspaces could reflect closer proximity and access to amen-ities, and the use of different modes of transport. That is,highly walkable neighborhoods are more likely to have200-meter buffer (DPA).ap-value from one-way Analysis of Variance (ANOVA) across sociodemographic andbMeasured by Street Smart Walk Score for home address.cLess than very much (1–4 on a 5-point scale); Very much (5 on a 5-point scale).alk Study participants (n = 95) by sociodemographic0.5297 0.9556 0.58380.75 (0.11) 0.35 (0.21)0.75 (0.10) 0.33 (0.21)0.3954 0.3574 0.12880.76 (0.10) 0.37 (0.22)0.74 (0.11) 0.31 (0.19)0.0021 0.0037 0.91440.80 (0.08) 0.34 (0.22)0.73 (0.11) 0.34 (0.21)0.8907 0.7597 0.40280.74 (0.09) 0.38 (0.26)0.75 (0.11) 0.33 (0.19)0.8087 0.7402 0.75830.75 (0.11) 0.34 (0.21)0.74 (0.11) 0.33 (0.20)0.3118 0.9095 0.15270.75 (0.11) 0.35 (0.22)0.75 (0.11) 0.27 (0.12)rd deviation (SDE), Minimum Convex Polygon (MCP), Daily Path Area using adestinations that older people deem important, within eas-ier walking access. As walking trips are often shorter thandriving trips, this would reduce distance traveled to thesedestinations. However, this study cannot disentangle thecompeting and related concepts of walkability and carusage. On one hand, car usage was associated with largeractivity spaces, potentially indicating greater mobility, yetwalkability was associated with smaller spaces, potentiallyindicating that the local neighborhood environment issufficient to fulfill daily activities and amenities. Futurework creating activity spaces by mode or considering thedistance traveled or frequency of trips within an activityspace may help tease apart these complex and comple-mentary elements of older adult mobility. All of these re-sults are novel and in our view, are worth pursuing infuture trials that evaluate different groups of older adults(we recruited older adults with low incomes) who resideacross diverse built environment settings.The life-space literature indicates that greater mobilityis related to a wide range of favorable health outcomes[18-21]. This study indicates that larger activity spaces areassociated with a number of different resources such asyounger age, access to a vehicle, or physical support forgoing outside. One possibility is that the geographic mo-bility of individuals, as measured either in life-space oractivity space, is in fact a proxy for personal resources.resource categories.Table 3 Associations between sociodemographic groups, resource characteristics and GPS-based activity space area and compactness of Walk The Talk StudyParticipants (n = 95)Area CompactnessSDE MCP DPA SDE MCP DPAPercent difference(95% CL)Percent difference(95% CL)Percent difference(95% CL)Mean difference(95% CL)Mean difference(95% CL)Mean difference(95% CL)Male −25.5 (−69.8, 83.6) −19.0 (−66.6, 96.6) −15.0 (−47.3, 37.2) 0.03 (−0.03, 0.09) 0.02 (−0.02, 0.07) 0.02 (−0.07, 0.11)Age (years)a65-69 125.8 (−43.4, 801.1) 91.7 (−50.8, 647.7) 46.1 (−30.0, 204.5) −0.07 (−0.17, 0.02) −0.04 (−0.12, 0.03) −0.08 (−0.22, 0.07)70-74 71.1 (−54.5, 543.6) 63.4 (−55.6, 501.1) 30.8 (−35.3, 164.2) −0.06 (−0.16, 0.03) −0.05 (−0.12, 0.01) −0.06 (−0.20, 0.08)75-79 −43.3 (−85.0, 114.5) −40.3 (−83.9, 120.7) −17.1 (−59.1, 68.0) −0.08 (−0.17, 0.02) −0.05 (−0.12, 0.01) 0.03 (−0.11, 0.17)EducationaSome or completed trade/technicalschool or college−15.2 (−69.9, 138.6) −4.5 (−65.5, 164.1) −1.9 (−43.4, 69.9) 0.05 (−0.02, 0.12) 0.06 (0.00, 0.11) 0.05 (−0.06, 0.15)Some university or higher −5.6 (−67.3, 172.5) 0.7 (−64.5, 185.7) 5.1 (−40.1, 84.6) −0.01 (−0.08, 0.07) −0.02 (−0.07, 0.04) 0.02 (−0.09, 0.13)WalkabilityaCar Dependent (0–49) 447.8 (54.8, 1838.7) 525.2 (80.4, 2066.1) 146.4 (26.0, 381.9) 0.02 (−0.07, 0.11) 0.01 (−0.06, 0.07) −0.16 (−0.30, −0.03)Somewhat Walkable (50–69) 213.7 (0.6, 877.8) 237.7 (10.4, 932.7) 102.3 (10.6, 269.9) 0.09 (0.01, 0.17) 0.07 (0.01, 0.13) −0.11 (−0.23, 0.01)Very Walkable (70–89) 225.5 (7.4, 886.6) 267.0 (23.3, 991.7) 95.8 (8.7, 252.7) 0.03 (−0.05, 0.11) 0.05 (−0.01, 0.10) −0.09 (−0.21, 0.02)Length of time in NeighborhoodaBetween 2 and up to 6 years ------b ------b ------b 0.00 (−0.07, 0.08) 0.00 (−0.06, 0.05) 0.04 (−0.07, 0.15)Between 6 and up to 9 years ------b ------b ------b −0.06 (−0.15, 0.03) −0.03 (−0.09, 0.03) −0.03 (−0.16, 0.10)> 9 years ------b ------b ------b 0.06 (−0.02, 0.14) 0.02 (−0.04, 0.07) 0.03 (−0.08, 0.15)Have a valid driver’s license 46.4 (−62.0, 463.3) 50.5 (−60.0, 466.4) 17.7 (−42.4, 140.8) 0.03 (−0.07, 0.12) −0.01 (−0.08, 0.06) −0.01 (−0.15, 0.13)Have access to a vehicle 285.1 (25.0, 1085.9) 304.4 (33.8, 1122.4) 139.8 (32.0, 335.6) −0.02 (−0.09, 0.06) −0.02 (−0.07, 0.04) −0.16 (−0.27, −0.04)Have physical support to go outside 184.2 (18.5, 581.2) 184.9 (20.6, 573.2) 75.0 (10.0, 178.4) −0.03 (−0.09, 0.03) −0.02 (−0.06, 0.02) −0.08 (−0.17, 0.01)Like to walk outside very much ------b ------b ------b −0.09 (−0.15, −0.03) −0.06 (−0.11, −0.02) −0.03 (−0.12, 0.06)Abbreviations: Confidence Limits (CL), Standard Deviation Ellipse using one standard deviation (SDE), Minimum Convex Polygon (MCP), Daily Path Area using a 200-meter buffer (DPA). Bold values indicate estimateswith p < 0.05.aReference categories: 80+ years old; secondary school or less; Walker’s Paradise (Walk Score 90–100); living in neighborhood less than 2 years.bNot tested in models of area due to lack of significance in bivariate analysis.Hirschetal.InternationalJournalofHealthGeographics2014,13:51Page8of14http://www.ij-healthgeographics.com/content/13/1/51with self-reported characteristics and trip identifica-gaging independently with their community. Specifically,of walkability, measured using Street Smart Walk ScoreHirsch et al. International Journal of Health Geographics 2014, 13:51 Page 9 of 14http://www.ij-healthgeographics.com/content/13/1/51this work highlights the role of neighborhood walkability,driving patterns, and physical support to go outside as im-portant factors in determining the size of older adult activ-ity spaces. Identifying an approach that best captures theactivity space of older adults may be useful for future workaimed at isolating features of the neighborhood environ-ment that support older people ‘aging in place’ or inform-tion. Although older adults may experience some dis-comfort while wearing GPS devices [72], they werealso highly compliant in wearing them. There was noclear association between any characteristics of olderadults and their level of compliance with wearingGPS [73]. This study did not examine specific desti-nations or resources within activity spaces, althoughwork building on this can illuminate factors that con-tribute to the capacity of neighborhoods to provideopportunities to older adults. Finally, bias associatedwith selective daily mobility may be a barrier tocausal inference when using GPS to assess neighbor-hood exposure [74].ConclusionThere are many different ways to represent geographicactivity spaces where individuals travel to and spendtheir time. However, outcomes and interpretations mayvary based on the approach used to generate an activityspace. It is important to use an approach tailored to theneeds of a specific research question and outcomes. Somefactors we identified as important to geographic mobilityof older adults may be used to inform interventions andto design policies that support older adults living and en-Additional work, examining changes in activity spaces asolder adults transition through life changes (e.g. retire-ment, loss of a spouse, move to a more walkable neigh-borhood, driving cessation) may help to tease apart thecomplex connections between older adult resources, geo-graphic mobility, and health outcomes.Strengths and limitationsOur study has a number of strengths, including thecharacterization of a sample of older adults with lowincome, the examination of multiple methods to cre-ate activity space, and the provision of sufficient codeto utilize these methods in other studies. We acknow-ledge that our study also has several limitations. Itwas not possible to draw strong conclusions given therelatively small and select sample, the cross-sectionaldesign, and potential measurement error associateding interventions and policies that support older adultsliving independently in the community.(www.walkscore.com) (ntotal = 2000), to ensure that partic-ipants were recruited across a range of built environments.Recruitment was done via telephone between January andFebruary 2012. Individuals were excluded if they were di-agnosed with dementia, left their home less than once in atypical week, were unable to understand or speak English,were unable to walk more than ten meters with or withouta mobility aid (e.g. cane, walker), or were unable to partici-pate in a mobility assessment involving a four meter walk.Measurement was conducted between March and May2012. At the end of the measurement sessions, partici-pants were instructed regarding wear of accelerometersand completion of travel diaries. A sub-group (n = 107) ofparticipants received GPS and were instructed as to theiruse. The study was approved by the University of BritishColumbia’s Clinical Research Ethics Board (certificate:H10-02913).Travel dataHome locations were geocoded based on participant re-ported home address. Participants’ travel patterns andphysical activity was assessed using travel diaries, QStarzDatalogger BT-Q1000XT GPS sensors (Semsons, Arcadia,CA, USA; recording at 1 s) and ActiGraph GT3X-Plus tri-axial accelerometers (ActiGraph LLC, Fort Walton Beach,FL, USA), respectively, over the 7 days immediately fol-lowing measurement sessions. For travel diaries, partici-pants were instructed to record for each trip: start andend locations and times, reason for travel, mode of travel,and others who accompanied them. For the GPS sensors,the vibration sensor was activated to preserve memoryand battery life; participants were not asked to chargedevices so the data collection period was a function ofbattery life. GPS data were downloaded using the QStarzMethodsSampleParticipants were recruited to take part in Walk the Talk(WTT), a cross-sectional study (n = 161) that evaluates theassociation between the built environment and the mobil-ity and health of low-income older adults. Participants res-ide in eight cities in Metropolitan Vancouver (Burnaby,New Westminster, North Vancouver, Richmond, Surrey,Vancouver, West Vancouver, White Rock). Methods forWTT are described elsewhere [75], but briefly: WTT basepopulation consists of 5806 households that receive a Shel-ter Aid for Elderly Renters (SAFER) rental subsidy fromBC Housing, had a head of household aged ≥ 65 years, anda telephone number on file with BC Housing (Figure 2).Households were sampled using a random stratified de-sign, selecting 200 households from within each decileData Viewer software. Of participants who were givenGPS, 97.2% wore them and of these, we acquired valid dataHirsch et al. International Journal of Health Geographics 2014, 13:51 Page 10 of 14http://www.ij-healthgeographics.com/content/13/1/51from 93.3% (n = 97). For accelerometry, data were down-loaded using the ActiLife software.There is little consensus regarding best-practices for pro-cessing GPS data [76]. Previously, customized automatedalgorithms were used to identify destinations and tripsfrom GPS data [77-80]. However, to capitalize on data ac-quired from travel diaries and to address broader questionsrelated to multi-modal trips, we coded GPS data manuallyfor this study, as has been done by our team [81] andothers [82]. In brief, the 1-sec GPS data were first time-aligned with accelerometer data and then processed usingArcGIS tracking analyst in concert with travel diaries todefine the start and end points of trips based on trip speed,distance, duration, and accelerometry-defined activity level.Tracks had to be of ≥30 s in duration and ≥100 m dis-tance to be considered a trip. Trip start was identifiedas the first GPS point outside of home or leaving theFigure 2 Walk The Talk (WTT) Participant recruitment and flow for GP(Burnaby, New Westminster, North Vancouver, Richmond, Surrey, VancouveRenters rental subsidy from BC Housing, have a head of household aged gBC Housing. Participants were considered lost after telephone contact if thparticipation. GPS data was considered invalid if the unit was turned to theprevious trips’ destination location where speed ≥1 km/hand distance >0 m of movement. Changes from these cri-teria indicated trip stop time, allowing for pauses of<5 min (e.g. at a stop light, bus stop). Two participantswho did not log at least one out-of-home trip were ex-cluded. Thus, the final sample size was n = 95 men andwomen who provided 333 days of recorded GPS data. Weremoved trips outside the metropolitan Vancouver area soas to represent participant movement within the region.GPS activity spacesThere are a number of different ways, derived from geog-raphy and ecology, to analyze geographic behaviors usingpoint data [67]. We analyzed trip-related GPS point data(n = 1,392,347), aggregated by individual, using Python2.7.2 (Python Software Foundation, www.python.org) andArcPy for ArcGIS 10.1 (ESRI, Redlands, CA, USA). WeS data. Source population comprised of households in our study arear, West Vancouver, White Rock) that receive a Shelter Aid for Elderlyreater than or equal to 65 years, and a telephone number on file withey could not be reached again after expression of interest in studyoff position by the participant.represented activity space using three different approachesfor each participant. They were; 1. Standard DeviationEllipse (SDE), 2. Minimum Convex Polygon (MCP), and 3.Daily Path Area (DPA) (Figure 3). SDE, a commonly usedmeasures of activity space, measures the directional distri-bution of a series of points [49,50,52,62,83-85]. Similar toothers [49,50], we used a one-SDE that contains 68% ofall GPS points. MCP, sometimes referred to as “homeranges”, represents the smallest polygon that contains allGPS points [62,86], with the outermost points serving asvertices [68]. DPA were adapted from previous literature[50,52,87]. We created them by buffering all of an in-dividual’s trips by 200-meters. We conducted sensitiv-ity analyses on activity spaces with and without water. Areaswere highly correlated (Spearman’s ρ >0.99, p < 0.0001)and results were consistent across measures with andwithout water (not presented).For the three activity space polygons we calculated twodimensions of activity space: 1. area (hectares) and 2.compactness. Area and perimeter were generated using“Calculate Geometry” in ArcGIS. Compactness is a meas-across individual sociodemographic and resource charac-teristics can highlight determinants of the shape, or localorientation, of travel.We provide python code to create area and compact-ness across the three activity space measures, both withand without water, in the Additional file 1 that supportsthis paper.Sociodemographic and resource characteristicsParticipants self-reported sociodemographic and resourcecharacteristics during measurement sessions. Self-reportedage (65–69 years; 70–74 years; 75–79 years; 80+ years),race (White; non-White), education level (completedsecondary school or less; some trade/technical school orcollege through completed trade/technical school or col-lege diploma; some university or higher), marital status(single; married; widowed, separated or divorced), co-habitation with someone, dog ownership, current validdriver’s license, and vehicle at their disposal were assessedvia questionnaire. We assessed neighborhood walkabilityusing Street Smart Walk Score, a single measure that ac-Hirsch et al. International Journal of Health Geographics 2014, 13:51 Page 11 of 14http://www.ij-healthgeographics.com/content/13/1/51ure of how circular a polygon is; a value near 1 indicatesthe activity space is similar to a circle while a value near 0indicates an elongated space, more closely resembling aline [53,55]. Compactness is calculated as the ratio of theperimeter of a circle with the same area to perimeter ofthe observed activity space. Compactness values may berelated to the activity space approach (e.g., SDE would beexpected to be more compact that DPA), however, withina given activity space type the comparison of compactnessFigure 3 Example of three types of activity spaces.counts for distance to popular amenities and street design(www.walkscore.com). We categorized walkability basedon cut-off categories as recommended by designers ofStreet Smart Walk Score (car dependent 0–49, somewhatwalkable 50–69, very walkable 70–89, walker’s paradise90–100). In-depth description of Walk Score can be foundelsewhere [88,89]. Participants also reported how longthey lived in their current neighborhood (classified intoquartiles: less than 2 years; between 2 and up to 6 years;Hirsch et al. International Journal of Health Geographics 2014, 13:51 Page 12 of 14http://www.ij-healthgeographics.com/content/13/1/51between 6 and up to 9 years; more than 9 years), whetherpeople in their lives offered support related to going out-side in their neighborhood (no or don’t know; yes, peoplethat offer physical support (drive places); yes, people thatoffer social support/companionship; yes, people that offerboth physical and social support), how much they liked towalk outside (not at all, not much or neutral; somewhat;very much), how confident they were walking in theirneighborhood (not at all, not much or neutral; somewhat;very much), and whether they had any falls in the past6 months or used a mobility aid for walking.Statistical methodsWe examined the correlation between each of the threeactivity spaces using Spearman’s and Pearson’s correlationcoefficients (ρ) for area and compactness, respectively.We describe area using medians and IQR due to non-normal distribution. We describe compactness as meanand SD. We used one-way Analysis of Variance (ANOVA),Kruskal-Wallis non-parametric one-way ANOVA orWilcoxon Rank Sum test to test for differences in areaand compactness of each activity space across sociodemo-graphic and resource categories, as appropriate. We usedlinear Ordinary Least Squares (OLS) regression models, toassess the associations between sociodemographic and re-source characteristics of participants and log-transformedarea or non-transformed compactness of activity spacesafter adjustment for other potential variables. Variableswere included in the simultaneous model based on a-priori hypothesis (sex, age, education) or if they wereassociated (p < 0.1) to dependent variables in bivariateanalyses. To enhance interpretability, results from re-gression models for area have been retransformed andpresented as percentage differences. We conducted allstatistical analyses using SAS software, Version 9.3(SAS Institute Inc., Cary, NC, USA).Additional fileAdditional file 1: We have provided a generic python script toenable others to easily create activity spaces using their own GPSdata: Activity_Space_Processing_TEMPLATE_supplement.py.AbbreviationsANOVA: Analysis of variance; DPA: Daily path area; GPS: Global positioningsystem; IQR: Interquartile range; MCP: Minimum convex polygon;SAFER: Shelter aid for elderly renters; SenTra: Senior tracking; SD: Standarddeviation; SDE: Standard deviation ellipse; WTT: Walk the talk.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsJH conceived of the study, designed and executed activity spacecalculations, performed the statistical analyses, and drafted the manuscript.MW shaped the conceptualization of activity spaces, supervised the activityspace calculations, advised the statistical analyses, and critically revised themanuscript. PC and HM participated in designing and coordinating the WalkThe Talk (WTT) study and critically revised the manuscript to ensure it wasrelevant to the field of older adult mobility. All authors read and approvedthe final manuscript.AcknowledgementsOngoing research was funded by the Canadian Institutes of Health Research(CIHR) (Grant #108607 “Walk the Talk: Transforming the Built Environment toEnhance Mobility in Seniors Team”). The authors thank community partners,BC Housing, and study participants for their valuable collaborations andinvolvement. The authors graciously acknowledge Anna Chudyk for codingof the travel diaries and Karen Schellenberg, Vivian Chung, Christine Vossand Morgan Schinkel for their contributions to the GPS and accelerometrydata processing. The content is solely the responsibility of the authors anddoes not necessarily represent the official views of the CIHR.Author details1Centre for Hip Health and Mobility and Department of Medicine, Universityof British Columbia, 2635 Laurel Street, Vancouver, British Columbia V5Z1 M9, Canada. 2Carolina Population Center, University of North Carolina atChapel Hill, 206 West Franklin St, Chapel Hill, NC 27516, USA. 3Faculty ofHealth Sciences, Simon Fraser University, 8888 University Drive, Burnaby,British Columbia V5A 1S6, Canada. 4Institute for Social Research, University ofMichigan, 426 Thompson Street, Ann Arbor, MI 48104, USA. 5Centre for HipHealth and Mobility and Department of Family Practice, University of BritishColumbia, 2635 Laurel Street, Vancouver, British Columbia V5Z 1 M9, Canada.Received: 10 October 2014 Accepted: 30 November 2014Published: 12 December 2014References1. Webber SC, Porter MM, Menec VH: Mobility in older adults: a comprehensiveframework. The Gerontologist 2010, 50(4):443–450. doi:10.1093/geront/gnq013.2. World Health Organization: Global age-Friendly Cities: A Guide. Geneva,Switzerland: World Health Organization; 2007.3. Alsnih R, Hensher DA: The mobility and accessibility expectations of seniorsin an aging population. Transp Res A Policy Pract 2003, 37(10):903–916.4. Sugiyama T, Thompson CW: Outdoor environments, activity and thewell-being of older people: conceptualising environmental support.Environment and Planning A 2007, 39(8):1943.5. Ziegler F, Schwanen T: ‘I like to go out to be energised by differentpeople’: an exploratory analysis of mobility and wellbeing in later life.Ageing Soc 2011, 31(05):758–781.6. Abraham A, Sommerhalder K, Abel T: Landscape and well-being: a scopingstudy on the health-promoting impact of outdoor environments. Int JPublic Health 2010, 55(1):59–69.7. Spinney JE, Scott DM, Newbold KB: Transport mobility benefits andquality of life: a time-use perspective of elderly Canadians. Transp Policy2009, 16(1):1–11.8. Metz D: Mobility of older people and their quality of life. Transp Policy2000, 7(2):149–152.9. Paez A, Scott D, Potoglou D, Kanaroglou P, Newbold KB: Elderly mobility:demographic and spatial analysis of trip making in the Hamilton CMA,Canada. Urban Stud 2007, 44(1):123–146.10. Golob TF, Hensher DA: The trip chaining activity of Sydney residents: across-section assessment by age group with a focus on seniors. J TranspGeogr 2007, 15(4):298–312.11. Scott DM, Newbold KB, Spinney JE, Mercado R, Páez A, Kanaroglou PS:New insights into senior travel behavior: the Canadian experience.Growth Change 2009, 40(1):140–168.12. Rosenbloom S: Meeting transportation needs in an aging-friendly community.Generations 2009, 33(2):33–43.13. Broome K, Worrall L, McKenna K, Boldy D: Priorities for an age-friendly bussystem. Can J Aging 2010, 29(03):435–444.14. Yen IH, Michael YL, Perdue L: Neighborhood environment in studiesof health of older adults: a systematic review. Am J Prev Med 2009,37(5):455–463.15. Baker PS, Bodner EV, Allman RM: Measuring life‐space mobilityin community‐dwelling older adults. J Am Geriatr Soc 2003,51(11):1610–1614.16. May D, Nayak U, Isaacs B: The life-space diary: a measure of mobility inold people at home. Disabil Rehabil 1985, 7(4):182–186.Hirsch et al. International Journal of Health Geographics 2014, 13:51 Page 13 of 14http://www.ij-healthgeographics.com/content/13/1/5117. Peel C, Baker PS, Roth DL, Brown CJ, Bodner EV, Allman RM: Assessingmobility in older adults: the UAB study of aging life-space assessment.Phys Ther 2005, 85(10):1008–1019.18. Crowe M, Andel R, Wadley VG, Okonkwo OC, Sawyer P, Allman RM: Life-spaceand cognitive decline in a community-based sample of African Americanand Caucasian older adults. J Gerontol A: Biol Med Sci 2008, 63(11):1241–1245.19. James BD, Boyle PA, Buchman AS, Barnes LL, Bennett DA: Life space andrisk of Alzheimer disease, mild cognitive impairment, and cognitivedecline in Old Age. Am J Geriatr Psychiatry 2011, 19(11):961–969.20. Boyle PA, Buchman AS, Barnes LL, James BD, Bennett DA: Associationbetween life space and risk of mortality in advanced Age. J Am GeriatrSoc 2010, 58(10):1925–1930.21. Xue Q-L, Fried LP, Glass TA, Laffan A, Chaves PH: Life-space constriction,development of frailty, and the competing risk of mortality the Women'shealth and aging study I. Am J Epidemiol 2008, 167(2):240–248.22. Snih SA, Peek KM, Sawyer P, Markides KS, Allman RM, Ottenbacher KJ:Life-space mobility in Mexican Americans aged 75 and older. J AmGeriatr Soc 2012, 60(3):532–537.23. Byles JE, Leigh L, Vo K, Forder P, Curryer C: Life space and mental health:a study of older community-dwelling persons in Australia. Aging MentHealth 2014, 19(2):1–9. doi:10.1080/13607863.2014.917607.24. Sikder S, Pinjari A: Immobility levels and mobility preferences of theelderly in the united states. Transportation Research Record: Journal of theTransportation Research Board 2012, 2318(−1):137–147.25. Zeitler E, Buys L: Mobility and out-of-home activities of older peopleliving in suburban environments:‘Because I'm a driver, I don't have a problem'.Ageing Soc 2014, 1–24. http://dx.doi.org/10.1017/S0144686X13001086.26. Shah RC, Maitra K, Barnes LL, James BD, Leurgans S, Bennett DA: Relation ofdriving status to incident life space constriction in community-dwelling olderpersons: a prospective cohort study. J Gerontol A: Biol Med Sci 2012, 67(9):984–989.27. Tung JY, Rose RV, Gammada E, Lam I, Roy EA, Black SE, Poupart P:Measuring life space in older adults with mild-to-moderate Alzheimer'sdisease using mobile phone GPS. Gerontology 2013, 60(2):154–162.28. Wan N, Lin G: Life-space characterization from cellular telephonecollected GPS data. Comput Environ Urban Syst 2013, 39:63–70.29. Boissy P, Brière S, Hamel M, Jog M, Speechley M, Karelis A, Frank J, Vincent C,Edwards R, Duval C: Wireless Inertial Measurement Unit With GPS(WIMU-GPS)—Wearable Monitoring Platform for Ecological Assessment ofLifespace and Mobility in Aging and Disease. In Engineering in Medicineand Biology Society, EMBC, 2011 Annual International Conference of theIEEE: 2011. Boston, MA: IEEE; 2011:5815–5819. http://dx.doi.org/10.1109/IEMBS.2011.6091439.30. Wan N, Qu W, Whittington J, Witbrodt BC, Henderson MP, Goulding EH,Schenk AK, Bonasera SJ, Lin G: Assessing smart phones for generatinglife-space indicators. Environ Plann Plann Des 2013, 40(2):350–361.31. Liddle J, Ireland D, McBride SJ, Brauer SG, Hall LM, Ding H, Karunanithi M,Hodges PW, Theodoros D, Silburn PA: Measuring the lifespace of people withParkinson’s disease using smartphones: proof of principle. JMIR mhealth anduhealth 2014, 2(1):e13.32. Schenk AK, Witbrodt BC, Hoarty CA, Carlson RH Jr, Goulding EH, Potter JF,Bonasera SJ: Cellular telephones measure activity and lifespace incommunity-dwelling adults: proof of principle. J Am Geriatr Soc 2011,59(2):345–352.33. Kerr J, Marshall S, Godbole S, Neukam S, Crist K, Wasilenko K, Golshan S,Buchner D: The relationship between outdoor activity and health inolder adults using GPS. Int J Environ Res Publ Health 2012, 9(12):4615–4625.34. Herrmann SD, Snook EM, Kang M, Scott CB, Mack MG, Dompier TP, Ragan BG:Development and validation of a movement and activity in physicalspace score as a functional outcome measure. Arch Phys Med Rehabil2011, 92(10):1652–1658.35. Rosenberg DE, Huang DL, Simonovich SD, Belza B: Outdoor builtenvironment barriers and facilitators to activity among midlife andolder adults with mobility disabilities. The Gerontologist 2013,53(2):268–279. doi:10.1093/geront/gns119.36. Buys L, Snow S, van Megen K, Miller E: Transportation behaviours of olderadults: an investigation into car dependency in urban Australia.Australas J Ageing 2012, 31(3):181–186.37. Wettstein M, Wahl H-W, Diehl MK: A multidimensional view of out-of-homebehaviors in cognitively unimpaired older adults: examining differentialeffects of socio-demographic, cognitive, and health-related predictors.Eur J Ageing 2014, 11(2):141–153.38. Shoval N, Auslander G, Cohen-Shalom K, Isaacson M, Landau R, Heinik J:What can we learn about the mobility of the elderly in the GPS era?J Transp Geogr 2010, 18(5):603–612.39. Kaspar R, Oswald F, Wahl H-W, Voss E, Wettstein M: Daily mood andOut-of-home mobility in older adults does cognitive impairment matter?J Appl Gerontol 2012, doi:10.1177/0733464812466290.40. Wahl H-W, Wettstein M, Shoval N, Oswald F, Kaspar R, Issacson M, Voss E,Auslander G, Heinik J: Interplay of cognitive and motivational resourcesfor out-of-home behavior in a sample of cognitively heterogeneousolder adults: findings of the SenTra project. J Gerontol Ser B Psychol SciSoc Sci 2013, 68(5):691–702. doi:10.1093/geronb/gbs106.41. Oswald F, Wahl H-W, Voss E, Schilling O, Freytag T, Auslander G, Shoval N,Heinik J, Landau R: The Use of tracking technologies for the analysis ofoutdoor mobility in the face of dementia: first steps into a project andsome illustrative findings from Germany. Journal of Housing For the Elderly2010, 24(1):55–73.42. Wettstein M, Wahl H-W, Shoval N, Oswald F, Voss E, Seidl U, Frölich L,Auslander G, Heinik J, Landau R: Out-of-home behavior and cognitiveimpairment in older adults: findings of the SenTra project. J ApplGerontol 2012, doi:10.1177/0733464812459373.43. Werner S, Auslander GK, Shoval N, Gitlitz T, Landau R, Heinik J: Caregivingburden and out-of-home mobility of cognitively impaired care-recipientsbased on GPS tracking. Int Psychogeriatr 2012, 24(11):1836–1845.44. Wettstein M, Wahl H-W, Shoval N, Auslander G, Oswald F, Heinik J:Identifying mobility types in cognitively heterogeneous older adultsbased on GPS-tracking: what discriminates best? J Appl Gerontol 2013,doi:10.1177/0733464813512897.45. Wettstein M, Wahl H-W, Shoval N, Auslander G, Oswald F, Heinik J:Cognitive status moderates the relationship between out-of-homebehavior (OOHB), environmental mastery and affect. Arch GerontolGeriatr 2014, 59(1):113–121.46. Shoval N, Wahl H-W, Auslander G, Isaacson M, Oswald F, Edry T, Landau R,Heinik J: Use of the global positioning system to measure the out-of-home mobility of older adults with differing cognitive functioning.Ageing Soc 2011, 31(05):849–869.47. Shoval N, Auslander GK, Freytag T, Landau R, Oswald F, Seidl U, Wahl H-W,Werner S, Heinik J: The use of advanced tracking technologies for theanalysis of mobility in Alzheimer's disease and related cognitive diseases.BMC Geriatr 2008, 8(1):7.48. Perchoux C, Chaix B, Cummins S, Kestens Y: Conceptualization andmeasurement of environmental exposure in epidemiology: accountingfor activity space related to daily mobility. Health & place 2013, 21:86–93.49. Sherman JE, Spencer J, Preisser JS, Gesler WM, Arcury TA: A suite ofmethods for representing activity space in a healthcare accessibilitystudy. Int J Health Geogr 2005, 4(1):24.50. Zenk SN, Schulz AJ, Matthews SA, Odoms-Young A, Wilbur J, Wegrzyn L,Gibbs K, Braunschweig C, Stokes C: Activity space environment anddietary and physical activity behaviors: a pilot study. Health Place2011, 17(5):1150–1161.51. Matthews SA, Yang T-C: Spatial polygamy and contextual exposures(SPACEs): promoting activity space approaches in research onplace and health. Am Behav Sci 2013, 57(8):1057–1081. doi:10.1177/0002764213487345.52. Starnes HA: Tests and development of perceived and objective builtenvironment measures for physical activity research. Purdue (West Lafayette,Indiana): 2012. http://docs.lib.purdue.edu/dissertations/AAI3544536/.53. Manaugh K, El-Geneidy A: What makes travel'local': Defining andunderstanding local travel behaviour. Journal of Transport and LandUse 2012, 5(3):15–17. doi:10.5198/jtlu.v5i3.300.54. Frank JS, Patla AE: Balance and mobility challenges in older adults:Implications for preserving community mobility. Am J Prev Med 2003,25(3, Supplement 2):157–163.55. Harding C, Patterson Z, Miranda-Moreno L: Activity Space Geometry andits Effect on Mode Choice. In 92nd Annual Meeting of the TransportationResearch Board, Washington, DC: 2013. ; 2013.56. Prins RG, Pierik F, Etman A, Sterkenburg RP, Kamphuis CBM, van Lenthe FJ:How many walking and cycling trips made by elderly are beyondcommonly used buffer sizes: Results from a GPS study. Health Place 2014,27:127–133.57. Kwan M-P: How GIS can help address the uncertain geographic contextproblem in social science research. Ann GIS 2012, 18(4):245–255.83. Newsome TH, Walcott WA, Smith PD: Urban activity spaces: Illustrationsand application of a conceptual model for integrating the time andspace dimensions. Transportation 1998, 25(4):357–377.84. Yuill RS: The standard deviational ellipse; an updated tool for spatialdescription. Geografiska Annaler Series B Human Geography 1971, 53(1):1971.doi:10.2307/490885.85. Lefever DW: Measuring geographic concentration by means of thestandard deviational ellipse. Am J Sociol 1926, 32(1):88–94. http://www.jstor.org/stable/2765249.86. Fan Y, Khattak AJ: Urban form, individual spatial footprints, and travel:examination of space-use behavior. Transportation Res Record: JTransportation Res Board 2008, 2082(1):98–106.87. Kwan MP: Gender and individual access to urban opportunities: a studyusing space–time measures. Prof Geogr 1999, 51(2):210–227.88. Hirsch JA, Diez Roux AV, Moore KA, Evenson KR, Rodriguez DA: ChangeHirsch et al. International Journal of Health Geographics 2014, 13:51 Page 14 of 14http://www.ij-healthgeographics.com/content/13/1/5158. Brownson RC, Hoehner CM, Day K, Forsyth A, Sallis JF: Measuring the builtenvironment for physical activity: state of the science. Am J Prev Med2009, 36(4):S99–S123. e112.59. Chaix B: Geographic life environments and coronary heart disease:a literature review, theoretical contributions, methodological updates, anda research agenda. Annu Rev Public Health 2009, 30:81–105.60. Chaix B, Kestens Y, Perchoux C, Karusisi N, Merlo J, Labadi K: An interactivemapping tool to assess individual mobility patterns in neighborhoodstudies. Am J Prev Med 2012, 43(4):440–450.61. Cummins S: Commentary: investigating neighbourhood effects onhealth—avoiding the ‘local trap’. Int J Epidemiol 2007, 36(2):355–357.62. Rainham D, McDowell I, Krewski D, Sawada M: Conceptualizing thehealthscape: contributions of time geography, location technologiesand spatial ecology to place and health research. Soc Sci Med 2010,70(5):668–676.63. Krenn PJ, Titze S, Oja P, Jones A, Ogilvie D: Use of global positioningsystems to study physical activity and the environment: a systematicreview. Am J Prev Med 2011, 41(5):508–515. doi:10.1016/j.amepre.2011.06.046.64. Oliver M, Duncan M, Badland H, Duncan J, MacRae D: Global positioningsystems in physical activity research: past, present, and future.http://discovery.ucl.ac.uk/1320088/.65. Vine D, Buys L, Aird R: The use of amenities in high density neighbourhoodsby older urban Australian residents. Landsc Urban Plan 2012, 107(2):159–171.66. Rodríguez DA, Cho G-H, Evenson KR, Conway TL, Cohen D, Ghosh-Dastidar B,Pickrel JL, Veblen-Mortenson S, Lytle LA: Out and about: association of thebuilt environment with physical activity behaviors of adolescent females.Health Place 2012, 18(1):55–62.67. Boruff BJ, Nathan A, Nijënstein S: Using GPS technology to (re)-examineoperational definitions of ‘neighbourhood’in place-based healthresearch. Int J Health Geogr 2012, 11(1):22.68. Worton B: A review of models of home range for animal movement.Ecol Model 1987, 38(3):277–298.69. Rosso AL, Auchincloss AH, Michael YL: The urban built environment andmobility in older adults: a comprehensive review. J Aging Res 2011,2011: http://dx.doi.org/10.4061/2011/816106.70. Clarke P, Ailshire JA, Bader M, Morenoff JD, House JS: Mobility disabilityand the urban built environment. Am J Epidemiol 2008, 168(5):506–513.71. Clarke P, Ailshire JA, Lantz P: Urban built environments andtrajectories of mobility disability: findings from a national sample ofcommunity-dwelling American adults (1986–2001). Soc Sci Med 2009,69(6):964–970.72. Zenk SN, Schulz AJ, Odoms-Young A, Wilbur J, Matthews SA, Gamboa C,Wegrzyn LR, Hobson S, Stokes C: Feasibility of using global positioningsystems (GPS) with diverse urban adults: Before and after data onperceived acceptability, barriers, and ease of use. J Phys Act Health2012, 9(7):924.73. Auslander GK, Werner S, Shoval N, Isaacson M, Landau R, Heinik J:Cooperation with complex research protocols: the use of globalpositioning systems with cognitively impaired elders–a preliminaryreport. Life 2010, 17(5.65):15–45.74. Chaix B, Meline J, Duncan S, Merrien C, Karusisi N, Perchoux C, Lewin A,Labadi K, Kestens Y: GPS tracking in neighborhood and health studies:a step forward for environmental exposure assessment, a step backwardfor causal inference? Health Place 2013, 21:46–51.75. Chudyk AM, Winters M, Moniruzzaman M, Ashe MC, Sims Gould J, McKay H:Destinations matter: the association between where older adults liveand their travel behavior. Journal of Transport and Health. In Press.76. Kerr J, Duncan S, Schipperjin J: Using global positioning systems in healthresearch: a practical approach to data collection and processing. Am JPrev Med 2011, 41(5):532–540.77. Thierry B, Chaix B, Kestens Y: Detecting activity locations from raw GPSdata: a novel kernel-based algorithm. Int J Health Geogr 2013, 12(1):14.78. Rodriguez D, Cho G-H, Elder J, Conway T, Evenson KR, Ghosh-Dastidar B,Shay E, Cohen DA, Veblen-Mortenson S, Pickrell J: Identifying walking tripsfrom GPS and accelerometer data in adolescent females. J Phys ActHealth 2012, 9(3):421.79. Ellis K, Godbole S, Marshall S, Lanckriet G, Staudenmayer J, Kerr J:Identifying active travel behaviors in challenging environments usingGPS, accelerometers, and machine learning algorithms. Frontiers in publichealth 2014, 2:36. doi:10.3389/fpubh.2014.00036.in walking and body mass index following residential relocation:the multi-ethnic study of atherosclerosis. Am J Public Health 2014,104(3):e49–e56.89. Hirsch JA, Moore KA, Evenson KR, Rodriguez DA, Roux AVD: Walk score®and transit score® and walking in the multi-ethnic study of atherosclerosis.Am J Prev Med 2013, 45(2):158–166.doi:10.1186/1476-072X-13-51Cite this article as: Hirsch et al.: Generating GPS activity spaces thatshed light upon the mobility habits of older adults: a descriptiveanalysis. International Journal of Health Geographics 2014 13:51.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistribution80. Hwang S, Hanke T, Evans C: Automated Extraction of CommunityMobility Measures from GPS Stream Data Using Temporal DBSCAN.In Computational Science and Its Applications–ICCSA 2013. Springer;2013:86–98. http://link.springer.com/chapter/10.1007/978-3-642-39643-4_7. ISBN 978- 3-642-39642-7.81. Voss C, Winters M, Frazer AD, McKay HA: They go straight home–don’tthey? Using global positioning systems to assess adolescentschool-travel patterns. Journal of Transport & Health 2014,doi:10.1016/j.jth.2014.09.013.82. Panter J, Costa S, Dalton A, Jones A, Ogilvie D: Development of methodsto objectively identify time spent using active and motorised modes oftravel to work: how do self-reported measures compare? Int J Behav NutrPhys Act 2014, 11(1):116.Submit your manuscript at www.biomedcentral.com/submit


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items