Table 1.
Studies self-identifying as natural experiments which met inclusion criteria
| Author, year | Geography | Type | Exposure | Exposed group | Unexposed group | Intervention outcome | Study design | Primary data |
|---|---|---|---|---|---|---|---|---|
| Allais et al., 2015 [27] | France | P | Food labelling | Consumers of supermarket stores | No control group (pre exposure) |
Sale of healthy food items |
Interrupted time series | No |
| Barlow et al., 2017 [28] | Canada | P | Tariff/trade agreement | Residents in country with tax on sugar sweeteners | Residents in countries with no tax on sugar sweeteners | Sales of sugar foods/beverage | Pre/post cohort + control | No |
| Barradas et al., 2017 [29] | Colombia | C | PA programme (park based) | Residents near park with PA programme |
Residents near parks without programme; with existing programme |
Quality of life | Post only | Yes |
| Bernatchez et al., 2015 [30] | Canada | E | Bike share programme | City residents near bike-share docks | City residents further from bike-share docks | Programme awareness | Repeated cross-sectional | Yes |
| Brandt et al., 2017 [31] | United States | P | Food regulation | Residents in counties with TFA restrictions | Residents in counties with no TFA restrictions | Hospital admission for cardiovascular events | Pre/post cohort + control | No |
| Cawley et al., 2007 [32] | United States | P | PE education | Students in high schools with PE requirements | Students in high schools without PE requirements | PA, weight | Post only | No |
| Cawley et al., 2010 [33] | United States | P | Social welfare benefit | Elderly receiving inflated social security benefits | No control group (variation in exposure) | Weight | Repeated cross-sectional | No |
| Christian et al., 2013 [34] | Australia | E |
Built environment (housing relocation) |
Residents in a PA-supportive neighbourhood | residents in conventional or hybrid neighbourhood | Walking | Pre/post cohort + control | Yes |
| Christian et al., 2017 [35] | Australia | E | Built environment (housing relocation) | Residents in a new neighbourhood | No control group | Walking | Pre/post cohort | Yes |
| Cohen et al., 2012 [36] | United States | E | Green space (outdoor gym equipment) | Park users in parks with new fitness equipment | Park users in parks without equipment | PA METs | Pre/post cohort + control | Yes |
| Cornelsen et al., 2017 [37] | United Kingdom | P | Tax (sugar) | Customers of restaurant chain | No control group (pre exposure) | Sugar-sweetened beverage sales | Interrupted time series | No |
| Cranney et al., 2016 [38] | Australia | E |
Green space (outdoor gym equipment) |
Park users in parks with new fitness equipment | No control group (pre exposure) | PA |
Repeated cross-sectional; qualitative interviews |
Yes |
| Datar & Nicosia, 2017 [39] | United States | P | Restricted access unhealthy food | Children living in states with school policies on discretionary foods | Children in states without policies | BMI | Post only | No |
| Dill et al., 2014 [40] | United States | E | Built environment (bicycle infrastructure) | Residents in streets with bike paths | Residents near streets without paths | PA/AT | Pre/post cohort + control | Yes |
| Dubowitz et al., 2015 [41] | United States | E | Food environment (access to healthy food outlet) | Residents in neighbourhoods with no full-service supermarket | No control group (baseline data only) | Healthy food purchases | Pre only | Yes |
| Elbel et al., 2011 [42] | United States | P | Food labelling | Consumers (low-income kids/teens) of fast-food outlets in city with mandatory food labelling | Consumers (low-income kids/teens) of fast-food outlets in city with no food labelling | Food consumption - fast food | Pre/post cohort + control | Yes |
| Ferguson et al., 2017 [43] | Australia | p | Food affordability (pricing) | Residents in remote Aboriginal communities | No control group (variation in exposure) | F&V, general food/beverage sales |
Interrupted time series; repeat cross-sectional; Qualitative interviews |
Mixed |
| Frew et al., 2014 [44] | United Kingdom | C | PA programme (access to services) | City residents with limited access to free PA services | Hypothetical (as if there was no free access) | PA-related diseases, QALYS | Pre/post cohort | Yes |
| Fu & VanLandingham, 2012 [45] | United States | O | Migration | Vietnamese immigrants in America | Vietnamese immigrant returnees and native (never-leavers) | BMI, waist-hip circumference | Post only | Yes |
| Gee et al., 2015 [46] | United States | O | Migration | Filipinos in America | Native Filipinos | BMI, waist-hip circumference | Pre/post cohort + control | Yes |
| Giles-Corti et al., 2013 [47] | Australia | E |
Built environment (housing relocation) |
Residents in a PA-supportive neighbourhood | residents in conventional or hybrid neighbourhood | AT and recreational PA | Pre/post cohort + control | Yes |
| Hobin et al., 2017 [48] | Canada | P | Food labelling | Consumers of supermarkets with nutrition labelling | Consumers of supermarkets without nutrition labelling |
Sale of healthy food items |
Pre/post cohort + control; qualitative interviews |
Mixed |
| Jancey et al., 2016 [49] | Australia | E | Built environment (workplaces) | Office workers move to new building | No control group (pre exposure) | PA | Repeated cross-sectional | Yes |
| Jones-Smith et al., 2017 [50] | United States | O | Economic resources (casino developments) | Native American new-borns born in area near casinos (greater resources) | Native American new-borns near no casino (lower resources) | Birth weight (childhood obesity) | Repeated cross-sectional | No |
| Jürges et al., 2011 [51] | Germany | P | Education reform (years educated) | Country population years of education | No control (variation in exposure) | Overweight, obesity, smoking | Repeated cross-sectional | No |
| Kapinos et al., 2017 [52] | United States | O | Clinical (fetal mal-presentation at birth delivery) | Mothers who deliver by caesarean birth | Mothers who deliver by vaginal birth | Mothers’ postpartum BMI | Pre/post cohort + control | No |
| Kesten et al., 2014 [53] | United Kingdom | E | Built environment (AT infrastructure) | Commuters in neighbourhoods proximal to AT infrastructure | No control group (pre exposure) | Awareness of AT infrastructure | Qualitative interviews; media | Mixed |
| Kodish et al., 2016 [54] | United States | O | Economic resources (casino ownership) | Native American tribal communities with casino ownership profit | Tribal communities with little/no cash profit | Access to healthy food, PA facilities, etc. | Qualitative interviews | Yes |
| Lee et al., 2017 [55] | United States | E | Built environment (transfer to local school) | Students transferred to local school | No control group (recall pre exposure) | AT participation | Post only | Yes |
| Madsen, 2011 [56] | United States | P | Clinical (health (BMI) screening optional notification) | Students in schools with screening-notified parents | Students in schools with screening not notified | BMI | Repeated cross-sectional; qualitative interviews | No |
| Odoms-Young et al., 2014 [57] | United States | P | Food access (change to nutritional content of welfare food package) | Low-income families receiving food security packages | No control group (pre exposure) |
Dietary intake, Body weight |
Pre/post cohort | Yes |
| Ovrum et al., 2014 [58] | Norway | P | Food access (healthy food) | Families of children whose schools give free F&V | Families of children who do not have F&V programme or have optional programme | F&V consumption | Post only | Yes |
| Pollack et al., 2014 [59] | United States | P | Social housing | Public housing residents in scattered housing | Public housing residents in clustered housing | Diet, PA behaviour, BMI, perceived health | Post only | Yes |
| Ram et al., 2016 [60] | United Kingdom | E | Social housing | City residents from different housing sectors | No control group (baseline data only) | PA, BMI, self-reported health | Pre only | Yes |
| Sadler et al., 2013 [61] | United States | E | Food environment (access to healthy food outlet) | Residents in neighbourhoods with new supermarket | Residents in neighbourhood with no supermarket | F&V intake, food security | Pre/post cohort + control | Yes |
| Schultz et al., 2017 [62] | United States | E | Built environment (street redesign) | Low-income neighbourhood residents | No control group (pre exposure) | PA, park usage | Repeated cross-sectional | Yes |
| Simões et al., 2017 [63] | Brazil | C | PA programme (community) | Cities participating in the programme | Non-participatory and not-yet-participatory cities | PA | Pre/post cohort + control (step wedge) | Yes |
| Stanley et al., 2016 [64] | Canada | E | Public transport strike | University students using public transport and other transport modes | No control group (pre-exposure) | PA, AT participation | Pre/post cohort | Mixed |
| Stone et al., 2012 [65] | Canada | P | PA programme (school curriculum) | Primary students in school with daily PA policy (mandatory minimum PA minutes) | Primary students in school without PA curriculum | PA, BMI | Post only | Yes |
| Sutherland et al., 2010 [66] | United States | P | Food labelling | Consumers of supermarkets with shelf nutrition labelling | No control group (variation in exposure) | Sale of healthy food items | Interrupted time series | No |
| Torres et al., 2017 [67] | Colombia | C | PA programme (park-based) | Residents near park with new PA programme |
Residents near parks without programme or with established programme |
MVPA | Pre/post cohort + control | Yes |
| Tudor-Locke et al., 2008 [68] | Australia | E | Built environment (housing relocation) | Residents relocating home within a metropolitan area | No control group (pre exposure) | PA | Pre/post cohort | Yes |
| Veitch et al., 2012 [69] | Australia | E | Green space | Users of a refurbished park | Users an un-furbished park | PA, park usage | Pre/post cohort + control | Yes |
| Wagner et al., 2013 [70] | United States | C | Social living arrangements | Student at a university | No control group (variation in exposure) | Various health behaviours (PA; alcohol) | Post only | No |
| Watson et al., 2016 [71] | Canada | P | Economic (change in unemployment benefits) | Adult workforce population | No control group (variation in exposure) | Obesity | Pre/post cohort | No |
| Zick, 2014 [72] | United States | P | Daylight savings scheme | Residents in states with daylight savings | Residents in states with no daylight savings | MVPA | Pre/post cohort + control | No |
AT active travel, BMI body mass index, C community intervention, E environmental intervention, F&V fruit and vegetable intake, METs metabolic equivalent expenditures, MVPA moderate-to-vigorous physical activity, O other, P policy, PA physical activity, PE physical education, QALYs quality-adjusted life years, TFA trans-fatty acid