Article title: Objectively measured walkability and active transport and weight-related outcomes in adults: a systematic review
Journal name: International Journal of Public Health
Author names: Gerlinde Grasser (a), Delfien van Dyck (c, d), Sylvia Titze (b), Willibald Stronegger (a)
Affiliations:
a Institute of Social Medicine and Epidemiology, Medical University of Graz, Graz, Austria
b Institute of Sport Science, University of Graz, Graz, Austria
c Faculty of Medicine and Health Sciences, Department of Movement and Sport Sciences, Ghent University, Ghent, Belgium
d Research Foundation Flanders, Brussels, Belgium
E-mail corresponding author:
Appendix 1: Summary of quality assessment
1 / Berke et al, 2007 / 0 / 0 / 2 / 1 / Poor
2 / Bodea et al, 2008 / 0 / 2 / 1 / 0 / Poor
3 / Boer et al, 2007 / 0 / 0 / 0 / 2 / Poor
4 / Brown et al, 2009 / 0 / 1 / 1 / 2 / Fair
5 / Cerin et al, 2007 / 0 / 1 / 2 / 1 / Fair
6 / Chatman, 2009 / 0 / 1 / 0 / 2 / Poor
7 / Coombes et al, 2010 / 0 / 0 / 1 / 2 / Poor
8 / Forsyth et al, 2007 / 1 / 2 / 1 / 0 / Fair
9 / Forsyth et al, 2008 / 1 / 2 / 1 / 2 / Good
10 / Forsyth et al, 2009 / 1 / 2 / 2 / 1 / Good
11 / Frank et al, 2004 / 0 / 2 / 0,5 / 0,5 / Fair
12 / Frank et al, 2006 / 0 / 1 / 1,5 / 1 / Fair
13 / Frank et al, 2007 / 0 / 2 / 0 / 1 / Poor
14 / Frank et al, 2008 / 0 / 2 / 0,5 / 1 / Fair
15 / Frank et al, 2009a / 0 / 2 / 1 / 1 / Fair
16 / Huang et al, 2009 / 1 / 1 / 0 / 0 / Poor
17 / Kitamura et al, 1997 / 0 / 1 / 0 / 1 / Poor
18 / Lee and Moudon, 2006 / 0 / 0 / 2 / 1 / Poor
19 / Li et al, 2008 / 1 / 1 / 2 / 1 / Fair
20 / Li et al, 2009* / -- / -- / -- / -- / Good
21 / Lopez, 2007 / 0 / 0 / 1 / 0,5 / Poor
22 / McGinn et al, 2007 / 0 / 2 / 2 / 1 / Fair
23 / Oakes et al, 2007 / 1 / 2 / 2 / 1 / Good
24 / Owen et al, 2007 / 0 / 1 / 2 / 2 / Fair
25 / Owen et al, 2010 / 0 / 1 / 2 / 2 / Fair
26 / Pouliou and Elliot, 2010 / 2 / 2 / 1 / 1 / Good
27 / Rundle et al, 2007 / 0 / 2 / 2 / 2 / Good
28 / Rundle et al, 2009 / 0 / 2 / 2 / 2 / Good
29 / Sallis et al, 2009 / 0 / 1 / 1 / 2 / Fair
30 / Scott et al, 2009 / 0 / 2 / 0,5 / 1 / Fair
31 / Smith et al, 2008 / 0 / 1 / 1 / 2 / Fair
32 / Van Dyck et al, 2010a / 1 / 1 / 2 / 1 / Fair
33 / Van Dyck et al, 2010b / 1 / 1 / 1,6 / 2 / Good
34 / Zick et al, 2009 / 0 / 1 / 1 / 2 / Fair
* Li et al, 2009: the rating for the only publication with a longitudinal design was as follows: Are study participants adequately described (age, sex, baseline for outcomes)? Met; Are intervention/exposed and control/non exposed group similar? Not met; Are valid/reliable/standardized measures used for the outcome measure? Met; What is the follow-up? Met; Were confounders controlled? Met; Global rating: Good;
Quality criteria for publications with cross sectional design (2=met, 1=partly met, 0=not met): What is the response rate? 2=>60%, 1=41-60%, 0=<40% or not known; Is the sample surveyed representative of the population for whom the result will be generalised? 2=if stated that representative or comparison with population provided; 1=if some minor limitations; 0=not met or not known; Are valid or reliable or standardised measures used for the outcome measure? 2=IPAQ, BRFSS, measured weight and height; 1=self-reported weight and height, 0=not met or not known (Note: in cases where there were more than one outcome measure, each outcome measure was rated separately and an average score was given); Were confounders controlled? 2=on individual and neighbourhood level, 1=on individual level; 0=not met or not known (Note: in cases where more than one analysis was conducted, each analysis was rated separately and an average score was given.); Global rating: good: 3 ratings “met” OR 2 ratings “met” and 2 ratings “partly met”, fair: 2 ratings “met” OR 1 rating “met” and at least 2 ratings “parly met”, poor: others (note: if the average score was between 2/1/0 (eg. 0.5), it was rounded up.)
8
Appendix 2: Summary of the characteristics of the reviewed publications
Ref. / Author / Country / Sample / Study / Exposure / Outcome / Association1 / Berke et al, 2007 / US / n=740
age:65+
by sex / Adult Changes in Thought Cohort Study / Walkability index (all buffers) / BMI / Women: n.s., 100m OR 0.78, 95% CI 0.54 1.16, 500m OR 0.80, 95% CI 0.53 1.19, 1000m OR 0.75, 95% CI 0.48 1.17;
Men: n.s., 100m OR 0.99, 95% CI 0.74 1.32, 500m OR 1.02, 95% CI 0.75 1.38, 1000m OR 0.93, 95% CI 0.67 1.28;
2 / Bodea et al, 2008 / US / n=10,150
age: 21+ / SMARTRAQ / Entropy index (continous) / Overweight / n.s., beta=-0.230, p=0.09
Entropy index (continous) / Obesity / n.s., beta=-0.320, p=0.06;
Entropy index (quartiles) / Obesity / -, beta =-0.055, p=0.02;
3 / Boer et al, 2007 / US / n=29,724
age: 5+ / National Personal Transportation Survey / Business diversity score / Walking for transport / +, 3 vs. 4 businesses: OR 1.24, 95% CI 1.07 1.44;
Block length / Walking for transport / +, <600 feet vs. 600-804 feet: OR: 1.26, 95% CI 1.04 1.52;
Ratio of four-way intersections to all intersections / Walking for transport / +, <25% vs. 25-49%: OR 1.35, 95% CI 1.13 1.60; 25-49% vs. 50-74%: OR 1.40, 95% CI 1.09 1.78;
4 / Brown et al, 2009 / US / n=5,000
age: 25-64
by sex / Salt Lake County drivers / Gross population density / BMI / Women: n.s., beta=-0.02;
Men: n.s., beta=-0.03;
Entropy index / BMI / Women: n.s., beta=0.00 for 2 land use types, beta=-0.04 for 3 land use types, beta=-0.03 for 6 land use types;
Men: n.s., beta=-0.02 for 2 land use types; -, beta=-0.04 for 3 land use types, p<0.05; -, beta=-0.04 for 6 land use types, p<0.05;
Intersection density / BMI / Women: n.s., beta=0.00, p=0.72;
Men: +, beta=0.02, p=0.01;
Gross population density / Overweight / Women: n.s., beta=0.00, p=0.98;
Men: n.s., beta=0.02, p=0.78;
Entropy index / Overweight / Women: n.s., beta=0.00 for 2 land use types, beta=-0.04 for 3 land use types, beta=-0.04 for 6 land use types;
Men: n.s., beta=-0.02 for 2 land use types, beta=-0.03 for 3 land use types, beta=-0.03 for 6 land use types;
Intersection density / Overweight / Women: n.s., beta=0.00, p=0.39;
Men: n.s., beta=0.01, p=0.21;
Gross population density / Obesity / Women: n.s., beta=0.00, p=0.83;
Men: n.s., beta=0.00, p=0.08;
Entropy index / Obesity / Women: n.s., beta=-0.01 for 2 land use types, beta=-0.02 for 3 land use types, beta=-0.02 for 6 land use types;
Men: n.s., beta=-0.05 for 2 land use types; -, beta=-0.06 for 3 land use types, p<0.05; -, beta=-0.07 for 6 land use types, p<0.05;
Intersection density / Obesity / Women: n.s., beta=0.00, p=0.73;
Men: n.s., beta=0.01, p=0.15;
5 / Cerin et al, 2007 / AUS / n=2,369
age: 20-65 / PLACE / Entropy index / Walking for transport / n.s., bx=30, 95%CI -165 224, p=0.765, bx2=-63, 95% CI -318 192, p=0.639;
Composition of land uses / Walking for transport / +, commercial vs. recreational areas 39.6 more weekly minutes, 95% CI 0.4 78.9, p=0.048;
6 / Chatman, 2009 / US / n=999
age: adult / Time Use Survey / Number of intersections / Overall active transportation / +, RR: 1.03, p=0.01;
7 / Coombes et al, 2010 / UK / n=6,803
age: 16+ / Bristol Quality of Life in your Neighbourhood survey / Herfindahl-Hirschmann Index / Overweight or obesity / n.s., (no further information provided);
Ratio of junctions to cul-de-sacs / Overweight or obesity / n.s., (no further information provided);
Number of junctions per km road / Overweight or obesity / -, p<0.001 (no further information provided);
8 / Forsyth et al, 2007 / US / n=716
age: nip / Twin Cities Walking Study / Gross population density / Walking for transport / +, all street network and circular buffers: r=0.1 to 0.2, p<0.001; 805mx805m grid: r=0.5 to 0.6, p<0.001;
Employment density / Walking for transport / n.s., for 805mx805m grid and for 400m street network buffers (no further information provided); +, for 200m street network buffers and circular buffers: r=0.1, p<0.05;
Housing unit density / Walking for transport / +, all street network and circular buffers: r=0.1 to 0.2, p<0.05, 805mx805m grid: r=0.5 to 0.6, p<0.001;
Population per developed land area / Walking for transport / +, all street network and circular buffers: r=0.1 to 0.2, p<0.001, 805mx805m grid: r=0.5 to 0.6, p<0.001;
Residential population in residential parcels / Walking for transport / +, all street network and circular buffers: r=0.1 to 0.2, p<0.05, 805mx805m grid: r=0.5 to 0.6, p<0.001;
Population plus employment per unit land area / Walking for transport / +, all street network and circular buffers: r=0.1 to 0.2, p<0.05, 805mx805m grid: r=0.5 to 0.7, p<0.001;
9 / Forsyth et al, 2008 / US / n=716
age: nip / Twin Cities Walking Study / Employment density / Walking for transport / n.s., (no further information provided);
Entropy index / Walking for transport / n.s., (no further information provided);
Herfindahl-Hirschmann Index / Walking for transport / n.s., (no further information provided);
Proportion of dissimilar land uses among grid cells in an area / Walking for transport / n.s., (no further information provided);
Intersection density / Walking for transport / +, r=0.4 to 0.6, p<0.05;
Ratio of four- and three-way intersections to all intersections / Walking for transport / +, for four-way intersections and the connected node ratio: r=0.4 to 0.5, p<0.05;
-, for three-way intersections: r=-0.4 to 0.5, p<0.05;
10 / Forsyth et al, 2009 / US / n=716
age: nip
by sex / Twin Cities Walking Study / Gross population density / Walking for transport / Both (only white): +, OR: 1.97, 95% CI 1.14 3.43, p=0.02;
Women: n.s., (no further information provided);
Men: +, OR: 1.78, 95% CI 1.12 2.83, p=0.02;
Block size / Walking for transport / Both (only white): +, OR: 0.81, 95% CI 0.47 1.37, p=0.42;
Women: n.s., (no further information provided);
Men: +, OR: 0.95, 95% CI 0.51 1.77, p=0.51;
11 / Frank et al, 2004 / US / n=7,134
age: nip
by sex / SMARTRAQ / Entropy index / Walking for transport / Women: +, r=0.051, p<0.001;
Men: +, r=0.046, p=0.01;
Intersection density / Walking for transport / Women: +, r=0.084, p<0.001;
Men: +, r=0.062, p<0.001;
Entropy index / BMI / Women: -, r=-0.086, p<0.001;
Men: -, r=-0.110, p<0.001, lowest vs. highest quartile decrease in BMI of 1.34 kg/m2;
Intersection density / BMI / Women: n.s., r=-0.018, p=0.28;
Men: -, r=-0.089, p<0.001, lowest vs. highest quartile decrease in BMI of 1.21 kg/m2, p<0.001;
Entropy index / Obesity / Both: -, OR: 0.88, 95% CI 0.84 0.92, with each quartile increase in entropy index, p<0.000;
Intersection density / Obesity / Both: n.s., (no further information provided);
12 / Frank et al, 2006 / US / n=1,228
age: 20-65 / NQLS / Walkability index / Overall active transport / +, beta=0.3, p=0.000, explained variance: 8.35%;
Walkability index / BMI / -, beta=-0.11 kg/m2, p=0.00, explained variance: 1.14%;
13 / Frank et al, 2007 / US / n=3,543
age: adult / SMARTRAQ / Walkability index / Walking for transport / Subsample neighborhood selection: +, 2nd quartile: odds 0.79, 95% CI 0.43 1.44, 3rd quartile: odds 1.71, 95% CI 1.07 2.69, 4th quartile: odds 2.56, 95% CI 1.65 3.84;
subsample neighborhood preference: n.s., 2nd quartile: odds 0.72, 95% CI 0.39 1.31, 3rd quartile: odds 1.11, 95% CI 0.64 1.91, 4th quartile: odds 1.62, 95% CI 0.95 2.76;
Walkability index / Obesity / Subsample neighborhood selection: -, 2nd quartile: odds 0.98, 95% CI 0.70 1.38, 3rd quartile: odds 0.83, 95% CI 0.59 1.16, 4th quartile: odds 0.67, 95% CI 0.49 0.89;
subsample neighborhood preference: n.s., 2nd quartile: odds 0.77, 95% CI 0.50 1.19, 3rd quartile: odds 0.90, 95% CI 0.57 1.40, 4th quartile: odds 0.89, 95% CI 0.55 1.42;
14 / Frank et al, 2008 / US / n=13,065
age: 25+
by sex / SMARTRAQ / Household unit density / Walking for transport / Both: +, 1st (=low) tertile response index=0.43, 2nd tertile response index=0.74, 3rd (=high) tertile response index=1.78, χ2=406.0, p<0.001;
Number of land uses per acre / Walking for transport / Both: +, 1st tertile response index=0.52, 2nd and 3rd tertile response index=0.97, χ2=13.1, p<0.001 (in middle density area);
Both: +, 1st tertile response index=1.0, 2nd and 3rd tertile response index=2.29 (in highly connected areas);
Intersection density / Walking for transport / Both: +, 1st tertile response index=1.11, 2nd tertile response index=1.48, 3rd tertile response index=2.11, χ2=41.9, p<0.001 (in higher density area);
Household unit density / Overweight / Women: n.s., (no further information provided);
Men: -, 1st tertile response index=1.36, 2nd tertile response index=1.28, 3rd tertile response index=1.12, χ2=67.0, p<0.001;
Number of land uses per acre / Overweight / Both: n.s., (no further information provided);
Intersection density / Overweight / Both: n.s., (no further information provided);
Household unit density / Obesity / Women: +, 1st (=low) and 2nd tertile response index=0.71, 3rd (=high) tertile response index=1.12 (among women without a degree);
Men: n.s., (no further information provided);
Number of land uses per acre / Obesity / Women: n.s., (no further information provided);
Men: -, 1st tertile response index=1.01, 2nd and 3rd tertile response index=0.65, χ2=9.5, p<0.004 (among men with a degree in highly connected areas)
Intersection density / Obesity / Women: n.s., (no further information provided);
Men: -, 1st tertile response index=1.24, 2nd tertile response index=0.99, 3rd tertile response index=0.77, χ2=26.6,p<0.001 (among men with a degree)