Electronic Supplementary Appendix S3. Summary of Study Characteristics

Electronic Supplementary Appendix S3. Summary of study characteristics

First author
(year) / Participant characteristics / Study length / Predictor(s) measured / Physical activity measure / Statistical results and effect sizes / Time varying covariation / QR (/15)
Armitage (2005)
[47] / N=94 members from private gym
Mean age: 37.6
Males: 41
Females: 53 / 3 months / TPB
1) Intention
2) Subjective norm
3) Self-efficacy/PBC / Self-report and objective measure (number of times gym card swiped)
- classified as "did not attend this week" and "did attend this week" / Intention - NS
SN – NS
PBC/SE – Significant β-.49 / No / 9
Beenackers (2012)
[86] / N = 1289
Mean age: 40.7
Males: 37.8%
Females: 62.2%
Western Australia
Residential Environment Project (RESIDE) / 2 years / Self-report and objective measures of built environment GIS used to assess: 1) Residential density 2) Street connectivity
3) land-use mix
4) Transport destination 5) Recreation destinations Self-report variables:
6) Access to mixed services 7) Neighbourhood aesthetics 9) Traffic hazards 10) Crime hazards 11) Hilly streets 12) Major barriers present 13) Local parking difficult 14) Access to park 15)Access to cycling paths 16) Traffic speed usually slow 17) Traffic-slowing devices
18) Pedestrian crossings present
19) Streets are well lit at night 20) Few cul-de-sacs 21) Intersection distance short 21) Many four way intersections 22) Many alternative routes 23) Number of transport destinations 23) Number of recreation destinations
24) Intrapersonal factors - attitude 25) Self-efficacy 26) Interpersonal factors -social influence
27) Intention / Self-report Cycling
Neighbourhood Physical Activity questionnaire
Dichotomous variable yes/no to cycle in a usual week / Connectivity – NS
Accessibility – Small ES OR=1.54
Residential density – Small ES OR = 1.52
Number of recreation destinations NE
Aesthetics – NS
(Safety) Crime hazards – NS
Well-lite streets - NS
Social influence - NE
Intention - NE
Self-efficacy - NS / Yes / 12
Bell
(2005)
[36];
Brown
(2003)
[106];
Brown
(2009)
[50] / N = 7281 – 15,935
Ages 18-78 yrs
Females only
Australian Longitudinal Study on Women's Health / 4 years; 2 surveys / Life Domains: 1)Employment status 2)Relationship status (marriage/cohabitation) 3) Motherhood status
4) Divorce
5)Age
6) Income
7) Ethnicity (not speaking English at home)
/ Self-report/study created
Survey 1: Two adapted questions from Australian National Heart Foundation Risk Factor Prevalence Study
Survey 2: Questions from Active Australia campaign surveys Categorized as (1) Inactive (2) Active
Both then calculated a PA score based on a MET score
Change in PA was categorized as "consistently sedentary", "consistently low active", "consistently active", "decreasing", or "increasing" / Employment – NS
Marriage predicted a decline in PA OR = 1.5;
Becoming a mother predicted a decline in PA OR = 1.6;
Remain a mother compared to remain childless predicted decline in PA OR = 1.3
Parenthood – birth of first child OR = 1.67
Birth of second child OR = 1.52
Divorce: NS, NE
Age: NS
Income: NS
Ethnicity – n/a / Yes / 11
12
13
Boon-Heinonen (2011)
[79]
Kip
(2002)
[129] / M = 4532 - 5115 Ages - 18-30 years
Males – 55%
Female – 45% CARDIA study
United States / 15 years
4 assessments / Socioeconomic status
1) Marriage
2)Education
3) Income
4) Parenthood
5) Change in family history of health (heart attack or stroke) / Self-report
13 categories of moderate and vigorous
recreational sports, exercise, leisure, and occupational activities / Marriage – Onset of marriage significant and meaningful for decline in PA
Education – Higher education level significant and meaningful for PA increase
Income – Significant and meaningful for decrease in income related to decrease in PA among African Americans
Parenthood – Significant and meaningful among females
Change in family history of health did not predict a change in PA / Yes / 9
13
Borodulin (2012)
[82] / N = 718
Mean age: 62.5
National Mini Finland health survey / 22 years / 1) Smoking
2)Occupational status
4) alcohol
5) Education / Self-report
Regular moderate to vigorous Leisure time physical activity / Smoking – NS
Employment – NS
Alcohol – NS
Education – NS / Yes / 12
Boutelle (2004)
[49] / N = 1146
Mean age: 37.6
Females: 917 Males: 229
Community based sample out of the U.S. Pound of Prevention Study / 4 years
Assessments taken annually (after controlled for intervention) / 1) Gender
2) Age
3) BMI
4) Employment
5) Ethnicity
6) Income group
7) Marital status
8) Education
9) Smoking status
10) TV/day
11) Social support (family) 12) Social support (friend) 13) Exercise level at baseline / Self-report
Vigorous activity 5 Questions from the Physical Activity History questionnaire (PAH)
Measure vigorous LTPA / Gender- NS
Age – NS
Employment – NS
Ethnicity adoption NS; maintenance sig h =.25 (compared to non-Caucasian)
Income – Maintenance h=.34; adoption NS
Marriage NS
Education - Maintenance h = .24
Adoption NS
Smoking - Maintenance h = .20 Adoption h = .33
Social support - NS / No / 12
Chung
(2009)
[77]
Margolis (2013)
[73]
Shaw
(2008)
[57] / N = 9935 - 16,606
Mean age: 60.3
Females: 53%; non-Hispanic whites: 74%
Health and Retirement Study (HRS) 1992 U.S. / 6 year study
Follow-up every 2 years / 1) Income
2) Divorce
3) Gender
4) Ethnicity
5) Marital status
6) Education level
7) Age
8) Work status / Self-report
Dichotomous self- report with vigorous physical activity definition / Income significant and small effect for decreased PA and lower income
Divorce - NS
Gender – NS
Ethnicity – NS
Marital status - NS
Education - NS
Age: NE
Employment - NS / No / 12
10
13
Coogan (2009)
[37] / N = 3206
Ages 21-69 African American women
Black Women's Health Study New York, Los Angeles, Chicago U.S. / 6 years Assessed biennially / Environment 1) Change in housing density / Self-report
Total hours walking per week and vigorous activity
7 response categories / Housing density NS / Yes / 9
Cornelio (2008)
[60] / N = 1246 Males: 567 Females: 679
(4 different age groups) Data from Cornelia Health Interview Survey Follow-up Study Barcelona, Catalonia Spain. / 8 years 1994 baseline and follow up in 2002 / 1) Gender
2) Age
3) Education level 4)Social class (occupation) 5) Smoking
6) Alcohol consumption 7) Health-related chronic conditions
8) Self perceived health 9) BMI / Self-report
Adapted questionnaire from US Health Interview Survey and the Welsh Heart Health Study.
Physical activity classified as sedentary, light-moderate, and vigorous
MVPA / No effects found for any variables / No / 13
Courneya (2001)
[87]
Plotnikoff (2001)
[130]
Plotnikoff (2009)
[131]
Rhodes (2006)
[100] / N = 683 –1602
Mean age: 40.59 Females: 54%
Ottawa-Carleton, Canada / Assessments made at baseline, 6 months and 1 year / 1) Intention
2) Perceived control 3) Attitude
4) Subjective norm
5) Social support
6) Cognitive and behavioral processes of change / Self-report single item question "Do you engage in regular vigorous physical activity for at least 20 minutes each week?" - Regular defined as 'at least 3 times'. / Intention was a significant and meaningful predictor for maintenance (.69); NS for adoption
Perceived behavioural control – NS
Attitude - NS at Time 1, but significant and meaningful at Time 2 for adoption of PA (.43)
Subjective norm – NS
Social support – Adoption – NS; Social support predicted lower chance of relapse in second six month follow-up (.36)
More positive attitude also predicted preparers to regular exercise at 6 months (.34); Intention, attitude and social support all predicted regression from action/maintenance stage. / No / 11
13
14
12
Dawson (2007)
[67] / N = 551
Mean age: 64.4
Females: 72.1%
Participants from England and Scotland Recruited from "Walking the Way to Health" and "Paths to Health" / Longitudinal 12 month follow-up / Neighbourhood environment - Barriers to neighbourhood walking 1) Health problems 2) No one to walk with
3) Aesthetics
4) No pavement 5) Nowhere to go nearby (connectivity, proximity) 8) Safety
9) Worry about being knocked over by cyclist 10) Worry about tripping over broken paving stones 11) Too much traffic 12) Too much pollution
13) Gender / Self-report
British Heart Foundation's "Daily Activities Questionnaire" adapted from validated American measure / No effects found for any variables. / No / 11
De Moor (2006)
[84] / N = 19,288
Mean age: 33.0 Males: 8773
Females: 10.515
Netherlands / 11 years / 1) Personality (neuroticism) / Self-report
Study designed question
Calculated as “exerciser” or “non-exerciser” / Exercisers scored higher on Extraversion (d=.32) and lower on neuroticism (d=0.14) / No / 8
Dishman (2010)
[91] / N=497 Mean age: 47
Males: 36.6%
Hawaii / Two years
Assessment every 6 months / 1) Processes of change (combined – doesn’t work for analysis)
2) Pros and cons (attitude/outcome expectancy)
3) Self-efficacy
4) Social support
/ Self-report
MVPA
IPAQ / NS for adoption of PA for social support
NS for adoption of PA for self-efficacy
NS for attitude/outcome expectancy / Yes / 14
Droomers (2001)
[78]
Slingerland (2007)
[132] / N =971 – 3793
Cohort 1 ages < 45 years
Cohort 2 ages > 45 years
GLOBE Study Netherlands / Prospective study over 6 years 1991 - 1997 / Self-report
1) Health status (at least one problem)
2) Emotional problems 3) Energy levels 4) Isolation 5) Mobility
6) Pain
7) Sleep
8) Self-perceived health 9) Obesity
10) Alcohol consumption 11) Smoking
12) Work responsibilities 13) Family responsibilities/parenthood 14) Locus of control 15) Neuroticism 16) Life events 17) Long lasting difficulties with health
18) Difficulties with relationships 19) Income
20) Financial problems 21) Situational difficulties 2) Neighbourhood circumstances 23) Housing problems 24) Education
25) Income
1) Employment status / Self-report Divided into 4 categories:
1) Completely sedentary
2) Lightly active 3) Moderately active 4) Highly active / Income – Meaningful for decrease in PA OR= 2.91 Small - Medium ES
Education – Higher education predicted increase in PA OR = 2.41
Employment – NS
Smoking – NS for maintenance for adults under 45 years, but meaningful for PA relapse over 45 years OR=1.56 Small
Alcohol – NS
Personality (neuroticism) – Meaningful for drop out of PA for middle aged adults (OR=1.87) but NS for older middle aged adults
Parenthood – Meaningful effect for parents less likely to maintain PA compared to non-parents (OR=1.87)
Safety – NS for overall PA; Extreme neighbourhood safety circumstances associated with meaningful effects of lower PA for people over 45 years
Aesthetics – Meaningful effects for extreme neighbourhood circumstances compared to ideal neighbourhood situations associated with lower levels of PA (OR = 2.04) / No / 13
12
Dunton (2008)
[92] / N= 246 (332 was total sample, only 246 for analysis of change)
Ages: 35 – 65
Males: 129
Females: 200
Southern California and Minnesota / 3 months / Self-report
1) Anticipated effective consequence of PA behavior success and failure (affective judgement) / Self-report short version of IPAQ / Affective judgement predicted an increase in PA adoption among sedentary individuals OR = 1.83 / 11
Eaton
(1993)
[62]
/ N = 1081
Mean age: 42.2 Two New England Communities 1986-1991 / 4 years / Self-report Exercise Performance: 1) Previous success with exercise
2) Previous success with weight loss
3) Participation in any programs
4) participation in exercise program Health Knowledge and beliefs: 5) Participation in weight reduction program 6) Health belief that exercise reduces CHD 7) Suggest exercise to others 8) Perceived high risk for CHD Social Support: 9) Children recommend exercise 10) Organization membership CHD Risk Factors: 11) Present cigarette smoking 12) Diabetes mellitus 13) Hypertension 14) HDL> 35mg 15) BMI (>120%) 16) Cholesterol 17) Work outside home 18) Married 19) Education more than 12 years
20) age / Self-report
MVPA
One item from Paffenbarger Physical Activity Questionnaire / Social support – Adopters h = .18; maintainers women -.58; men – 1.74
Attitude/outcome expectancy and PA adoption Small ES
Sex NS
Age – Sig for adoption among men; NS for maintenance
Marriage – NS for adoption or maintenance
Education significant and meaningful ES - .24 for adopters; .15 maintenance
Smoking NS / No / 12
Eng
(2005)
[34] / N = 38,865 Males only Ages: 40-75
Health Professionals Follow up study (U.S.) / Longitudinal repeated measures collected at 4 year intervals / Self-report 1)Marital status (divorce) / Self-report on activities converted into METs / Divorce NS / Yes / 11
Evenson (2002)
[74]
Ranchod (2013)
[64] / N = 6812 – 7782
Ages 45-64
Atherosclerosis Risk in Communities
United States / 2-6 year follow up / 1) Ethnicity
2) Education
3) Built environment (accessibility/connectivity)
4) Income
5) Age
/ Self-report
Baecke Questionnaire
Activities converted into METS semi-quantitative questionnaire adapted from the Cross-Cultural Activity Participation Study / Ethnicity NS
Education NS
Accessibility/convenience NS
Income NS
Age – NS / No / 10
13
Fortier
(2008)
[38] / N = 149 females Ottawa, Canada Mean age = 51.8 years / Longitudinal baseline and 6 month follow-up / 1) Intentions
2) Autonomous motivation 3) Controlled motivation
4) Subjective norms
5) Perceived behavioural control 6) Attitudes / Self-report
Community Health Activities Model Program for Seniors / Intention predicted PA change ES=.23
SN – NS
SE/PBC – NS
Attitude - NS / No / 9
French
(1996)
[83]
Schmitz (1997)
[56] / M = 3,672
Males = 1,692 Females = 1,980
Drawn from the Healthy Worker Project (Minnesota) / Longitudinal Surveyed at baseline and 2 years later / 1) Smoking status
2) Marital status
3)Education
4) Age
/ Self-report Categorized as leisure time or job-related PA Scale adapted from the Physical Activity History (PAH) / Smoking – NS
Marital status – NS
Divorce – NS
Education (higher) predicted increase in PA in women but NS for men
Age – Significant and meaningful for older age and decline in PA / Yes / 12
10
Fuchs
(1996)
[99] / N = 299
Mean age: 52.2
Females: 149 Berlin / Spring and summer/fall / 1) Intention
2) Self-efficacy
3) Social support / Self-report
Exercise behavior assessed by 5 items referring to activities or activity by groups (i.e. biking, endurance exercise, sports games) / Intention sig for adoption and maintenance .18
Self-efficacy – adoption NS
Social support sig in adoption .23 / No / 10
Godin
(2009)
[94] / N = 91
Mean age: 40.9
Females: 40
Males: 51
Quebec City, Canada / 3 months / 1) Perceived behavioral control/self-efficacy
2) Intention
3) Perceived built environment (accessibility and connectivity; aesthetics; / Self-report
Leisure-time PA "Within the last 3 months, how often did you participate in one or more physical activities of moderate intensity, totaling at least 30 min in a same day during your leisure time? (Validated measure by Gionet and Godin) / PBC/self-efficacy- NS
Intention was a significant and meaningful predictor of PA
Accessibility/connectivity – NS
Safety – NS
Aesthetics - NS / No / 10
Grace
(2006)
[107] / N= 243
Females only Mean age: 39.5 Canada / Maternity group and control group. Measured at baseline, maternity leave and returning to work (average 16 months) / Pregnancy/parenthood / Self-report PA Health-promoting lifestyle profile II / Motherhood - NS / No / 13
Gunnel
(2014)
[90] / N= 203 adults Mean age: 32.57
Female – 68%
Canada / 6 month follow up / 1) Attitude/outcome expectation / Self-report
Light, moderate and strenuous exercise Leisure time Exercise Questionnaire (LTEQ) / Attitude/outcome expectation - NS / Yes / 13
Hamer
(2012)
[46]
Kouvonen (2011)
[108] / N = 394 - 5395
Mean age: 54 WhiteHall study Europe / 13 year follow-up (1997 – 2010)
5 year follow up (phase 5 1997-1999 and phase 7 of study 2002-2004) / 1) Smoking
2) Education
3) Employment
4) Social support
/ Moderate to vigorous
Self-report at baseline, objectively measured with accelerometers at follow-up / Smoking – NS
Education significant and meaningful for higher levels predicting PA increase
Employment - NS
Social support NS OR = 1.34 (maintaining), social support (improving) OR = 1.25 / No / 11
12
Hawkley (2009)
[51] / N = 229
Ages 50 - 68 Chicago Health Longitudinal Study / Longitudinal Tested annually for 3 years Analysis carried out between subjects / 1) Loneliness (social isolation)
2) Age
3) Gender
4) Ethnicity
5) Education / Self-report Minnesota Leisure Time Physical Activity Questionnaire Categorized. Self-report MVPA / Higher levels of loneliness predicted a decrease in PA OR = 1.57
Age - NS
Gender – NS
Ethnicity – NS
Education - NS / No / 13
Higgins (1999)
[39] / N = 258
Female only
Mean age: 20.6 Australia / 18 month Longitudinal study / 1) Desire (strength, permanence, stability) (attitude/outcome expectancy)
2) Perceived behavioural control/ Self-efficacy
/ Self-report
Used a PA level scale from Weir, Jackson, and Pinkerton (1989) / Increased cons associated with decrease in PA = -.15
Higher self-efficacy/PBC associated with increase in PA = .31, .19 / Yes / 11
Hull
(2010)
[52] / N = 646
M: 52%
F: 48% Mean age: 24.7
University of Pittsburgh Physical Activity Study / 2 year follow-up / Self-report
1) Age
2) Gender
3) Ethnicity
4) Income
5) Education
6) Marital status
7) Having a child (parenthood)
/ Self-report Past Leisure Time Physical Activity Questionnaire MVPA / Age – NS
Gender – Significant and meaningful for males greater decline than females =.21
Ethnicity – NS
Income – NS
Education – NS
Marriage/cohabitation – NS
Parenthood – Significant and meaningful decline in PA among women -.48; Significant and meaningful decline in PA among males moderate effect size / Yes / 13
Ingledew (1998)
[72] / N = 425
Males: 282 Females: 143 Mean age: 37.8
British Government employees / 3 month / Exercise motivation 1)Attitude/outcome expectancy
2) Affective judgement
3) Gender / Self-report
Regular exercise in leisure time adapted, categorised with TTM / Gender – NS
Attitude/outcome expectancy – NS
Affective judgement – Large effect for overall PA change / No / 7
Kaplan
(1991)
[75] / N = 4025
Males: 1781
Females: 2238 Alameda Country Study Data came from the Human Population Laboratory's 24 -year study of adult residences Alameda Country, California / Longitudinal - interviewed in 1965 and 1974 / 1) Ethnicity/race
2) Family income
3) Occupation/employment
4) Marriage/cohabitation
5) Social isolation / Self-report
List of activities Constructed a 4 item Leisure-time physical activity index / Ethnicity – Significant for whites reporting more positive changes in PA compared to blacks
Income – Lower income related to decrease in PA in women but not men
Employment – Significant and meaningful for blue collar workers decrease PA more than white collar workers but not for women
Marriage/cohabitation – Significant and meaningful for marriage and PA increase
Social isolation – Significant and meaningful for PA decline / No / 7
Kern
(2010)
[68] / N = 1,409 Males: 798 Females: 611 Terman Life Cycle Study 1922 California / Longitudinal Took qualitative information to examine PA changes across 4 decades / 1) Measures of child and adult personality (neuroticism)
2) Sex / Self-report Determined MET from qualitative data / Neuroticism – For men significant and meaningful for lower neuroticism but NS for women
Extraversion were more active
Conscientiousness – NS
Sex - NS / No / 7
King
(1998)
[104]
Rimal
(2001)
[93] / N = 558 - 1033 302 women and 256 men
Ages 25 to 75 years Stanford Five City Project (California) / Assessments completed at 5 time points across 10 years / 1) Marital status
2) Exercise knowledge
3) Exercise self-efficacy
/ Self-report From interviewer administered single item question on 7 -point Likert Scale / Marital status (marriage)– NS
Marital status (divorce) – NS
Knowledge of PA - NS
PBC/self-efficacy – Significant and meaningful for higher PBC/self-efficacy and positive change in PA / Yes / 9
11
Klesges
(1992)
[35]
/ N=120
adults males (non-smokers n=101; smokers n = 19)
Mean age (non-smokers) 34.6 years, (smokers) 36.5 years) U.S. / Longitudinal 3 years / 1) Smoking / Self-report BAECKE / NS / No / 10
Koenman (2012)
[30] / N=186 Two groups, used the younger group Mean age 53.7 years Netherlands
LASA study – Longitudinal Aging Study Amsterdam / 10 years follow-up / Life events
1) Widowhood
2) Retirement
/ Self-report
Daily MVPA and sport or exercise in previous two weeks LASA physical activity questionnaire (LAPAQ) / N/a
Widowhood excluded only looked at in older age group
Excluded for retirement / No / 12
Kwan
(2010)
[101] / N = 127
Mean age: 22.4
67 females; 62 males
United States / 3 months / 1) Affective response / Self-report
3 study designed questions targeting voluntary aerobic exercise
SR 7 day Physical Activity Revall (PAR) / Positive affect during exercise predicted PA at follow up β = 0.62, p = 0.03 / Yes / 14
Laaksonen (2002)
[69] / N = 5494 Males: 2229 Females: 2852 Ages 20 - 64 years Finnish adults / Longitudinal Baseline and 7 years later / 1) Smoking
2) Sex / Self-report
Single item question on moderately strenuous leisure time PA on six point scale / Smoking – No meaningful effects
Sex - NS / No / 10
Lahti
(2011)
[53] / N = 7332 Ages 40-60 years City of Helsinki, Finland
Helsinki Health Study questionnaire / Survey 2000-2002 and again in 2007 / 1) Age
2) Sex / Self-report