TABLE 1.

Basic characteristics of five included studies

First author (year)	Study design a (period)	Study area b (scale)	Sample characteristics	Sample size	Sample age c	Dataset used	Statistical model
Ewing (2006) 33	C (1997)	US (N)	Adolescents	6760	12–17, 18–23 between 1997 and 2003	• Round‐1 (1997) data of 1997 National Longitudinal Survey of Youth (NLSY97)	Multilevel logistic regression
	C (2002)		Adolescents	5815		• Round‐6 (2002) data of 1997 National Longitudinal Survey of Youth (NLSY97)
	L (1997–2003)		Adolescents lived in the same county throughout survey	3367		• All consecutive rounds of NLSY97 from 1997 through 2003
	L (1997–2003)		Adolescents moved	3567		• The intercounty movers of all consecutive rounds of NLSY97 from 1997 to 2003
Guarnizo‐Herreño (2019) 34	C (2003) C (2007) C (2011–2012)	US (N)	Children and adolescents	129 781	10–17 between 2003 and 2012	National Surveys of Children's Health (NSCH)	Linear probability regression
Guettabi (2014) 35	L (1988–2008)	US (N)	Children	2324	2–17 between 1988 and 2008	• 1979 National Longitudinal Survey of Youth (NLSY79)• The NLSY79 Child Survey	• Linear regression • First difference estimation
Schwartz (2011) 36	C (2001–2008)	31‐county region of Pennsylvania, US (C31)	Children and adolescents	47 769	5–18 between 2001 and 2008	• Electronic health record (EHR): individual‐level data• Pennsylvania Spatial Data Access: (PASDA) place‐level data for Pennsylvania	Multilevel logistic regression
Seliske (2012) 37	C (2007, 2008)	33 census metropolitan areas (CMA), Canada (C33)	Adolescents	7017	12–19 in 2007, 2008	• Canadian Community Health Survey (CCHS)	Multilevel logistic regression

Abbreviation: NA, not available.

^aStudy design: C, cross‐sectional; L, longitudinal.

^bStudy scale: (N), National; (Nn), n countries; (C), city; (Cn), n cities.

^cSample age: age in baseline year for cohort studies or mean age in survey year for cross‐sectional studies.