- Split View
-
Views
-
Cite
Cite
Anna W. Wright, Makeda Austin, Carolyn Booth, Wendy Kliewer, Systematic Review: Exposure to Community Violence and Physical Health Outcomes in Youth, Journal of Pediatric Psychology, Volume 42, Issue 4, May 2017, Pages 364–378, https://doi.org/10.1093/jpepsy/jsw088
- Share Icon Share
Abstract
Objective To systematically review the evidence for associations between exposure to community violence and physical health outcomes in children and adolescents. Methods A thorough search of multiple online databases and careful consideration of inclusion and exclusion criteria yielded a final 28 studies for detailed review. In addition to review of findings, studies were rated on overall quality based on study design. Results Seven categories of physical health outcomes emerged, including asthma/respiratory health, cardiovascular health, immune functioning, hypothalamic–pituitary–adrenal axis functioning, sleep problems, weight, and a general health category. There were mixed findings across these categories. Evidence for a positive association between community violence exposure and health problems was strongest in the cardiovascular health and sleep categories. Conclusion There is reason to believe that community violence exposure has an effect on some areas of physical health. Additional well-designed research that focuses on mechanisms as well as outcomes is warranted.
Introduction
Each year, millions of children around the world witness or directly experience community violence (Fowler, Tompsett, Braciszewski, Jacques-Tiura, & Baltes, 2009), defined as exposure to intentional, interpersonal violent acts experienced directly or indirectly in a public setting (Kliewer, 2016a). For example, among adolescents aged 12–17 years, a study using a U.S. national household probability sample found that an estimated 38% had witnessed community violence in their lifetime, defined as seeing someone shot with a gun, cut or stabbed with a knife, sexually victimized, mugged or robbed, threatened with a weapon, or beaten up so badly that they required medical attention (Zinzow et al., 2009). In terms of victimization, in 2014, ∼3 million persons aged ≥12 years in the United States (1.1%) experienced at least one violent victimization (Truman & Langton, 2015). Thus, numerous youth living in the United States are exposed annually to violence that can and does lead to physical injury. In addition, there is a range of other physical health outcomes youth might experience as a consequence of exposure to community violence (ECV).
Scholars have studied violence exposure because it is associated with significant psychological distress in youth, as meta-analyses and reviews (e.g., Fowler et al., 2009) have shown. However, relative to the psychological, emotional, and even cognitive and academic costs associated with ECV, considerably less attention has been paid to the physical health consequences of exposure.
There are a number of reasons to hypothesize positive associations between ECV and physical health problems. First, many of the correlates of ECV including low socioeconomic status (SES) and neighborhood disadvantage are linked to poor physical health in youth (El-Sheikh et al., 2013; Evans & Kim, 2007; Schreier & Chen, 2013), suggesting that links between ECV and physical health are likely as well. Further, given that ECV is stressful, as demonstrated by positive associations with posttraumatic stress and anxiety symptoms, the wear and tear on the body resulting from this stress could present as physical health problems. Recent data demonstrating attenuated cortisol responses over time among low-income youth with ECV suggest this is a plausible pathway linking exposure and health (Aiyer, Heinze, Miller, Stoddard, & Zimmerman, 2014; Kliewer, 2016b).
The purpose of this article was to systematically review the extant literature linking ECV and physical health outcomes in youth. Because previous reviews of the community violence literature have not considered physical health outcomes in youth, we intentionally cast a wide net incorporating a broad range of health outcomes to address this question. To maintain a focused approach, we choose to concentrate on health outcomes rather than behaviors. Our search strategy included health outcomes most prevalent in childhood and adolescence. Our review also includes an evaluation of the quality of the available research, and offers suggestions for next steps in advancing this area of work.
Method
To ensure transparent reporting, this systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines (Moher, Liberati, Tetzlaff, & Altman, 2009), which include a 27-item checklist and a four-phase flow diagram illustrating identified, screened, eligible, and included studies.
Definition of Variables
Consistent with other reviews of the effects of ECV on youth (Fowler et al., 2009; Ozer, Lavi, Douglas, & Wolf, 2015), community violence was defined as violence occurring outside of the home but did not include acts of war and terrorism. Both victimization by community violence (i.e., being the recipient of an intentional act intended to cause harm) and witnessing community violence (i.e., seeing an event that involved loss or injury of some kind, including death) were included in the review. Peer victimization and bullying are unique subtypes of community violence, and there have been several recent reviews linking them to components of physical health (e.g., Midei & Matthews, 2011; van Geel, Goemans, & Vedder, 2016). As a result, we excluded articles focusing solely on peer victimization or bullying from our review.
Physical health outcomes were defined as any physiological characteristic or process that indicated the presence or absence of illness or injury. Thus, a construct was included in our review if it could be considered a health outcome on its own merits, although it also might impact other health outcomes. Most health behaviors did not fit this definition and were excluded. Several outcomes included in our review are ones which can be measured objectively using biological markers, although use of objective measures was not mandatory. Physical health outcomes considered were those which are most common in childhood and adolescence, identified in part through consulting pediatric psychology texts (e.g., Roberts, Koocher, Routh, & Willis, 1993). Our finalized list of health outcomes to search included sleep problems, asthma and respiratory health, cardiac health, weight, injury and morbidity, stress response functioning, cancer, pain, immune functioning, traumatic brain injury, and generalized health outcomes such as growth or somatic complaints. For the purpose of this review, we did not include studies that focused solely on mental health, despite recognizing that physical and mental health are intertwined.
Search Strategy
A literature search was conducted using the following databases: PsycINFO, PubMed, CINAHL Complete, and Web of Science. Searches were conducted between January 10, 2015 and February 25, 2015. The following ECV-specific search items were used to identify articles: community violence, violence exposure, experienced violence, neighborhood violence, victimization, witnessing violence, direct exposure to violence, and indirect exposure to violence. The following physical health outcome-specific search terms were used: health outcomes, physical health, sleep problem, asthma, sleep disorder, obesity, sleep disturbance, hypertension, morbidity, allostatic load, cancer, disease, death, and injury. An additional update of the literature was conducted between July 3, 2016 and July 8, 2016. Previous searches were rerun, to catch newly published studies, and additional health outcome-specific search terms were added at this time, including pain, headache, somatic complaints, chronic pain, cardiovascular disease, immune functioning, stress response, respiratory illness, accidents, traumatic brain injury, and diabetes. All possible combinations of ECV and health outcome-specific search terms were searched within each database.
Inclusion and Exclusion Criteria
To meet inclusion criteria, studies had to adhere to the following parameters: The study (1) assessed youth 0–18 years old at time of initial investigation; (2) used a measure of community violence and assessed child ECV via child, mother, or third-party report; (3) assessed physical health as an outcome; (4) was written in the English language; (5) was a peer-reviewed article; (6) used quantitative or semiquantitative measures; and (7) consisted of a sample size of ≥20, as we were not confident that samples <20 could make reliable statements regarding associations between violence exposure and health outcomes. Articles that inferred youth ECV by virtue of living in a specific neighborhood or community were excluded. Studies that examined witnessing violence, direct victimization, or a combination were included.
Study Selection
Quality Ratings
To assess the strength of the evidence, we adapted a quality rating procedure from a systematic review investigating interpersonal violence and obesity (Midei & Matthews, 2011). Midei and Matthews (2011) outlined a thorough, objective way to evaluate important study design elements and compare across numerous studies. Their rating scale allows for the weighting of various design elements based on importance in contributing to valid, reliable outcomes. For example, use of objective measures holds more weight than large sample size, although both are taken into consideration in overall score. We rated the same five components as Midei and Matthews, with slight adaptations to ratings of measures and covariates to be more applicable to the range of reviewed studies. The five components were sample size, measure of ECV, measure of physical health outcome, study design, and inclusion of covariates. For sample size, n ≥ 500 was rated 1 and n < 500 was rated 0. For measures of ECV, an objective measurement was rated 2 and a self-report measure was rated 0. For measures of physical health outcome, a validated measure was rated 2 and a nonobjective measure was rated 0. For study design, a longitudinal design was rated 2 and a cross-sectional design was rated 0. For covariates, the inclusion of covariates (e.g., race, age, sex, and parent education level) was rated 1 and no covariates was rated 0. Full-text articles that met inclusion criteria were included regardless of quality rating. See Table I for study details and quality ratings.
Reference . | Sample . | Study design . | Community violence measure . | Physical health outcome(s) . | Covariates . | Quality ratings . | Resultsa . |
---|---|---|---|---|---|---|---|
Asthma or respiratory health | |||||||
Clougherty et al. 2007 | N = 413; M + F | L | My child’s ETV scale (Selner-O’ Hagan, Kindlon, Buka, 1998) | Parent-reported asthma diagnosis | Maternal asthma, Edu, smoking before and after pregnancy, child G, A | 7 | 0 |
Age 5–8 | |||||||
52% HI; 44% WH | |||||||
Cohen et al. 2008 | N = 1,213; M + F | L | Modified version of the ECV Scale (Richters & Saltzman, 1990) | Parent- or child-reported past-year asthma diagnosis and/or lifetime allergic rhinitis diagnosis | A, G, SES, maternal/paternal history of asthma, caregiver-perceived stress | 8 | 0 |
Age 6.8–16.6 | |||||||
100% Puerto Rican | |||||||
Hellyer et al. 2013 | N = 365; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Parent- and child-reported asthma prevalence | A, G, R | 3 | 0 |
Age 9–11 | |||||||
51.5% HI; 26.8% AF; 50.7% WH | |||||||
Ramratnam et al. 2015 | N = 466; M + F | C | Exposure to Violence Survey—modified version | Asthma (physician diagnosed in past year) and allergens (serum levels of IgE specific to dust mite, cockroach, cat dander, dog dander, and mouse urinary protein | A, G, SES, parental asthma, prematurity, environmental tobacco smoke, distance from major roadway | 5 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Rosas-Salazar et al. 2016 | N = 747; M + F | C | Exposure to Violence Survey—one item (Selner-O’ Hagan et al., 1998) | Caretaker-reported asthma; laboratory-measured IgE levels | Child A, R, G, annual household I, maternal history of asthma, early life exposure to environmental tobacco smoke, exposure to gun violence, study site, child BMI, child obesity, child history of premature birth, Caesarian section birth, child low birth weight, parent Edu, current exposure to environmental tobacco smoke, breast-feeding | 4 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Sternthal et al. 2010 | N = 2,071; M + F | L | My child’s ETV scale (Selner-O’ Hagan et al., 1998) | Parent-reported lifetime asthma diagnosis | A, G, R, maternal Edu, maternal asthma, family violence in the home, neighborhood disadvantage, social disorder, collective efficacy | 8 | + |
Age 0–9 | |||||||
14.5% WH; 34.1% AF; 32.1% Mexican HI; 15.7% Non-Mexican HI; 3.7% OT | |||||||
Swahn and Bossarte 2006 | N = 1,943; M + F | C | Victimization—three items | Parent-reported lifetime asthma diagnosis and attack in the past year | G, R, grade, SS, self-perceived overweight status, exercising patterns | 6 | + |
No age range reported in article | |||||||
2003 Nationally representative sample | |||||||
Wilson et al. 2005 | N = 769; M + F | C | Modified version of ECV Scale (Richters & Saltzman, 1990) | Self-reported URI symptoms in the past 2 months | None | 5 | + |
Age 16–20 | |||||||
45% AA; 9% AS; 24% HI; 5% WH; 17% OT | |||||||
Yakubovich et al. 2016 | N = 6,002; M + F | C | Adapted version of the child ECV checklist (Richters & Martinez, 1990) | Self-reported asthma in the past year | A, G, urban/rural location, and province | 2 | + |
Age 10–17 | |||||||
100% AF | |||||||
Cardiovascular health | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Laboratory-measured BP and pulse | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Ford and Browning 2014 | N = 7,971; M + F | L | Victimization—three items; witnessing violence—one item | Hypertension status | Child R, A, foreign birth, general health, perpetrator of violence past, pregnancy status, inappropriate cuff, antihypertensive med use, resting HR; parent Edu, MS, two-parent household, economic hardship, public assistance | 4 | + |
Age 11–17 | |||||||
1994–1995 Nationally representative sample | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Mean HR, HR variability, SBP, and DBP | G, R | 5 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Wilson et al. 1998 | N = 40; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | SBP, DBP, and ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
Wilson et al. 2002 | N = 56; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | Daytime and nighttime ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
General health | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Teacher- and self-reported somatic complaints | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, Anx/Dep, number living in child’s home, HCA; maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Fredland et al. 2008 | N = 309; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Saltzman, 1990) | Self-reported physical health | G, R, poverty, maturation | 3 | 0 |
Age 11–14 | |||||||
95% AA | |||||||
Graham-Bermann and Seng 2005 | N = 160; M + F | C | Survey of ECV scale, parent version | Parent-reported physical health | G, monthly I, minority status, MS | 3 | 0 |
Age 4–6 | |||||||
52% WH; 44% AA or MI; 4% OT | |||||||
Halpern et al. 2013 | N = 8,531; F | C | Witnessing victimization—one itemDirect victimization—five items | Somatic complaints in the past year | A at Wave 1, self-reported R, parent Edu | 2 | + |
Age 11–21 | |||||||
24% AA; 57.9% WH; 18.1% HI | |||||||
Hart et al. 2013 | N = 409; M + F | L | The CREV ( Cooley et al., 1995) | Clinically significant somatic complaints | A, G, R, SES | 5 | + |
Age 8–13 | |||||||
85.6% AA | |||||||
Swisher and Warner 2013 | N = 20,103; M + F | L | Witnessing violence and violent victimization composite—five items | Youth-reported survival expectancies | A, G, R, immigrant status | 4 | + |
Mean age 15.7 at W1 | |||||||
1995 Nationally representative sample | |||||||
HPA axis functioning | |||||||
Aiyer et al. 2014 | N = 266; M + F | L | Witnessing violence and violent victimization combined—five items | Salivary cortisol AUC | G, R, interview start time | 7 | + |
Age 13.9–16.9 | |||||||
80% AA; 17% WH; 3% MI | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Salivary cortisol levels | G, R | 3 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Suglia et al. 2010 | N = 43; M + F | L | Modified version of the ECV scale (Richters & Saltzman, 1990) | Salivary cortisol levels, cortisol AUC, and diurnal slope | A, G, R, SES | 7 | 0 |
Age 7–13 | |||||||
54% WH; 46% HI | |||||||
Immune functioning | |||||||
Kacanek et al. 2016 | N = 268; M + F | C | Life Events Checklist—seven items (LEC; Kang, Mellins, Dolezal, Elkington, & Abrams, 2011) | Laboratory-measured unsuppressed viral load and CD4% | Child A, G, R, child knowledge of HIV status, annual household I, caregiver Edu, caregiver MS, caregiver relationship to child, presence of caregiver’s partner/spouse in home, HIV clinical characteristics, highly active antiretroviral therapy, adherence/nonadherence | 3 | + |
Age 8–15 | |||||||
71% AA; 21% HI | |||||||
Sleep | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Martinez, 1990) | Self-reported frequency of difficulty sleeping | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, anx/dep, number living in child’s home, HCA, maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Self-reported sleep disturbances | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Kliewer and Lepore 2015 | N = 362; M + F | L | Survey of children’s ECV scale (Richters & Saltzman, 1990) | Self-reported sleep problems | A, G, life events, family structure, intervention condition, school | 5 | + |
Age 11–14 | |||||||
51% HI; 34% AA; 8% WH; 5% AM; 6% AS | |||||||
Rubens et al. 2014 | N = 144; M + F | C | Witnessing violence—five items | Self-reported sleep problems | G, time spent in the United States | 1 | + |
Age 14–19 | |||||||
100% HI | |||||||
Spilsbury et al. 2014a | N = 46; M + F | L | REVS (van Dulmen et al., 2008)—items related to witnessing violence only | Self-reported and actigraphy-measured sleep quantity and quality | A, G, I, time between index-event and baseline REVS score, vacation status, PTS score, CWWVP service use | 7 | + |
Mean age 11.4% | |||||||
60.9% AA; 23.9% WH; 15.2% OT | |||||||
Umlauf et al. 2015 | N = 12,263; M + F | L | Witnessing violence—two items | Self-reported sleep problems | A, G, hopelessness | 4 | + |
Age 9.75–19.25 | |||||||
93.5% AA; 6% MI; <1% WH or AS | |||||||
Weight | |||||||
Gooding et al. 2015 | N = 147; M + F | C | Screen for Adolescent Violence Exposure (SAVE; Hastings & Kelley, 1997) | BMI | R, parent Edu, SS, weekly physical activity, frequency of fast food consumption, frequency of sugar-sweetened beverages consumption, internalizing problems | 1 | 0 |
Age 13–17 | |||||||
42.9% WH; 17.7% AA; 17% HI; 22.5% OT |
Reference . | Sample . | Study design . | Community violence measure . | Physical health outcome(s) . | Covariates . | Quality ratings . | Resultsa . |
---|---|---|---|---|---|---|---|
Asthma or respiratory health | |||||||
Clougherty et al. 2007 | N = 413; M + F | L | My child’s ETV scale (Selner-O’ Hagan, Kindlon, Buka, 1998) | Parent-reported asthma diagnosis | Maternal asthma, Edu, smoking before and after pregnancy, child G, A | 7 | 0 |
Age 5–8 | |||||||
52% HI; 44% WH | |||||||
Cohen et al. 2008 | N = 1,213; M + F | L | Modified version of the ECV Scale (Richters & Saltzman, 1990) | Parent- or child-reported past-year asthma diagnosis and/or lifetime allergic rhinitis diagnosis | A, G, SES, maternal/paternal history of asthma, caregiver-perceived stress | 8 | 0 |
Age 6.8–16.6 | |||||||
100% Puerto Rican | |||||||
Hellyer et al. 2013 | N = 365; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Parent- and child-reported asthma prevalence | A, G, R | 3 | 0 |
Age 9–11 | |||||||
51.5% HI; 26.8% AF; 50.7% WH | |||||||
Ramratnam et al. 2015 | N = 466; M + F | C | Exposure to Violence Survey—modified version | Asthma (physician diagnosed in past year) and allergens (serum levels of IgE specific to dust mite, cockroach, cat dander, dog dander, and mouse urinary protein | A, G, SES, parental asthma, prematurity, environmental tobacco smoke, distance from major roadway | 5 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Rosas-Salazar et al. 2016 | N = 747; M + F | C | Exposure to Violence Survey—one item (Selner-O’ Hagan et al., 1998) | Caretaker-reported asthma; laboratory-measured IgE levels | Child A, R, G, annual household I, maternal history of asthma, early life exposure to environmental tobacco smoke, exposure to gun violence, study site, child BMI, child obesity, child history of premature birth, Caesarian section birth, child low birth weight, parent Edu, current exposure to environmental tobacco smoke, breast-feeding | 4 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Sternthal et al. 2010 | N = 2,071; M + F | L | My child’s ETV scale (Selner-O’ Hagan et al., 1998) | Parent-reported lifetime asthma diagnosis | A, G, R, maternal Edu, maternal asthma, family violence in the home, neighborhood disadvantage, social disorder, collective efficacy | 8 | + |
Age 0–9 | |||||||
14.5% WH; 34.1% AF; 32.1% Mexican HI; 15.7% Non-Mexican HI; 3.7% OT | |||||||
Swahn and Bossarte 2006 | N = 1,943; M + F | C | Victimization—three items | Parent-reported lifetime asthma diagnosis and attack in the past year | G, R, grade, SS, self-perceived overweight status, exercising patterns | 6 | + |
No age range reported in article | |||||||
2003 Nationally representative sample | |||||||
Wilson et al. 2005 | N = 769; M + F | C | Modified version of ECV Scale (Richters & Saltzman, 1990) | Self-reported URI symptoms in the past 2 months | None | 5 | + |
Age 16–20 | |||||||
45% AA; 9% AS; 24% HI; 5% WH; 17% OT | |||||||
Yakubovich et al. 2016 | N = 6,002; M + F | C | Adapted version of the child ECV checklist (Richters & Martinez, 1990) | Self-reported asthma in the past year | A, G, urban/rural location, and province | 2 | + |
Age 10–17 | |||||||
100% AF | |||||||
Cardiovascular health | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Laboratory-measured BP and pulse | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Ford and Browning 2014 | N = 7,971; M + F | L | Victimization—three items; witnessing violence—one item | Hypertension status | Child R, A, foreign birth, general health, perpetrator of violence past, pregnancy status, inappropriate cuff, antihypertensive med use, resting HR; parent Edu, MS, two-parent household, economic hardship, public assistance | 4 | + |
Age 11–17 | |||||||
1994–1995 Nationally representative sample | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Mean HR, HR variability, SBP, and DBP | G, R | 5 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Wilson et al. 1998 | N = 40; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | SBP, DBP, and ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
Wilson et al. 2002 | N = 56; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | Daytime and nighttime ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
General health | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Teacher- and self-reported somatic complaints | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, Anx/Dep, number living in child’s home, HCA; maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Fredland et al. 2008 | N = 309; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Saltzman, 1990) | Self-reported physical health | G, R, poverty, maturation | 3 | 0 |
Age 11–14 | |||||||
95% AA | |||||||
Graham-Bermann and Seng 2005 | N = 160; M + F | C | Survey of ECV scale, parent version | Parent-reported physical health | G, monthly I, minority status, MS | 3 | 0 |
Age 4–6 | |||||||
52% WH; 44% AA or MI; 4% OT | |||||||
Halpern et al. 2013 | N = 8,531; F | C | Witnessing victimization—one itemDirect victimization—five items | Somatic complaints in the past year | A at Wave 1, self-reported R, parent Edu | 2 | + |
Age 11–21 | |||||||
24% AA; 57.9% WH; 18.1% HI | |||||||
Hart et al. 2013 | N = 409; M + F | L | The CREV ( Cooley et al., 1995) | Clinically significant somatic complaints | A, G, R, SES | 5 | + |
Age 8–13 | |||||||
85.6% AA | |||||||
Swisher and Warner 2013 | N = 20,103; M + F | L | Witnessing violence and violent victimization composite—five items | Youth-reported survival expectancies | A, G, R, immigrant status | 4 | + |
Mean age 15.7 at W1 | |||||||
1995 Nationally representative sample | |||||||
HPA axis functioning | |||||||
Aiyer et al. 2014 | N = 266; M + F | L | Witnessing violence and violent victimization combined—five items | Salivary cortisol AUC | G, R, interview start time | 7 | + |
Age 13.9–16.9 | |||||||
80% AA; 17% WH; 3% MI | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Salivary cortisol levels | G, R | 3 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Suglia et al. 2010 | N = 43; M + F | L | Modified version of the ECV scale (Richters & Saltzman, 1990) | Salivary cortisol levels, cortisol AUC, and diurnal slope | A, G, R, SES | 7 | 0 |
Age 7–13 | |||||||
54% WH; 46% HI | |||||||
Immune functioning | |||||||
Kacanek et al. 2016 | N = 268; M + F | C | Life Events Checklist—seven items (LEC; Kang, Mellins, Dolezal, Elkington, & Abrams, 2011) | Laboratory-measured unsuppressed viral load and CD4% | Child A, G, R, child knowledge of HIV status, annual household I, caregiver Edu, caregiver MS, caregiver relationship to child, presence of caregiver’s partner/spouse in home, HIV clinical characteristics, highly active antiretroviral therapy, adherence/nonadherence | 3 | + |
Age 8–15 | |||||||
71% AA; 21% HI | |||||||
Sleep | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Martinez, 1990) | Self-reported frequency of difficulty sleeping | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, anx/dep, number living in child’s home, HCA, maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Self-reported sleep disturbances | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Kliewer and Lepore 2015 | N = 362; M + F | L | Survey of children’s ECV scale (Richters & Saltzman, 1990) | Self-reported sleep problems | A, G, life events, family structure, intervention condition, school | 5 | + |
Age 11–14 | |||||||
51% HI; 34% AA; 8% WH; 5% AM; 6% AS | |||||||
Rubens et al. 2014 | N = 144; M + F | C | Witnessing violence—five items | Self-reported sleep problems | G, time spent in the United States | 1 | + |
Age 14–19 | |||||||
100% HI | |||||||
Spilsbury et al. 2014a | N = 46; M + F | L | REVS (van Dulmen et al., 2008)—items related to witnessing violence only | Self-reported and actigraphy-measured sleep quantity and quality | A, G, I, time between index-event and baseline REVS score, vacation status, PTS score, CWWVP service use | 7 | + |
Mean age 11.4% | |||||||
60.9% AA; 23.9% WH; 15.2% OT | |||||||
Umlauf et al. 2015 | N = 12,263; M + F | L | Witnessing violence—two items | Self-reported sleep problems | A, G, hopelessness | 4 | + |
Age 9.75–19.25 | |||||||
93.5% AA; 6% MI; <1% WH or AS | |||||||
Weight | |||||||
Gooding et al. 2015 | N = 147; M + F | C | Screen for Adolescent Violence Exposure (SAVE; Hastings & Kelley, 1997) | BMI | R, parent Edu, SS, weekly physical activity, frequency of fast food consumption, frequency of sugar-sweetened beverages consumption, internalizing problems | 1 | 0 |
Age 13–17 | |||||||
42.9% WH; 17.7% AA; 17% HI; 22.5% OT |
Note. M = male; F = female; AM = American Indian or Alaskan Native; AS = Asian; AF = African; AA = Black or African–American; HI = Hispanic/Latino; WH = White or of Caucasian decent; MI = mixed race; OT = other; C = cross-sectional; L = longitudinal; URI = upper respiratory infection; BP = blood pressure; BMI = body mass index; PTS = posttraumatic stress symptoms; HCA = history of child abuse; Anx/Dep = anxiety or depression; SS = smoking status; CWWVP, Children Who Witness Violence Program; SBP = systolic blood pressure; SES = socioeconomic status; R = race/ethnicity; G = sex; A = Age; Edu = education level; MS = marital status; Emp = employment status; I = income; DBP = diastolic blood pressure; HR = heart rate; HRV = heart rate variability; AUC = area under the curve; REVS = Recent Exposure to Violence Scale; CREV = Children’s Report of Exposure to Violence; TISH = Things I Have Seen and Heard.
+ Refers to a statistically significant relationship between ECV and health outcome; 0, no significant relationship between ECV and health outcome.
Reference . | Sample . | Study design . | Community violence measure . | Physical health outcome(s) . | Covariates . | Quality ratings . | Resultsa . |
---|---|---|---|---|---|---|---|
Asthma or respiratory health | |||||||
Clougherty et al. 2007 | N = 413; M + F | L | My child’s ETV scale (Selner-O’ Hagan, Kindlon, Buka, 1998) | Parent-reported asthma diagnosis | Maternal asthma, Edu, smoking before and after pregnancy, child G, A | 7 | 0 |
Age 5–8 | |||||||
52% HI; 44% WH | |||||||
Cohen et al. 2008 | N = 1,213; M + F | L | Modified version of the ECV Scale (Richters & Saltzman, 1990) | Parent- or child-reported past-year asthma diagnosis and/or lifetime allergic rhinitis diagnosis | A, G, SES, maternal/paternal history of asthma, caregiver-perceived stress | 8 | 0 |
Age 6.8–16.6 | |||||||
100% Puerto Rican | |||||||
Hellyer et al. 2013 | N = 365; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Parent- and child-reported asthma prevalence | A, G, R | 3 | 0 |
Age 9–11 | |||||||
51.5% HI; 26.8% AF; 50.7% WH | |||||||
Ramratnam et al. 2015 | N = 466; M + F | C | Exposure to Violence Survey—modified version | Asthma (physician diagnosed in past year) and allergens (serum levels of IgE specific to dust mite, cockroach, cat dander, dog dander, and mouse urinary protein | A, G, SES, parental asthma, prematurity, environmental tobacco smoke, distance from major roadway | 5 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Rosas-Salazar et al. 2016 | N = 747; M + F | C | Exposure to Violence Survey—one item (Selner-O’ Hagan et al., 1998) | Caretaker-reported asthma; laboratory-measured IgE levels | Child A, R, G, annual household I, maternal history of asthma, early life exposure to environmental tobacco smoke, exposure to gun violence, study site, child BMI, child obesity, child history of premature birth, Caesarian section birth, child low birth weight, parent Edu, current exposure to environmental tobacco smoke, breast-feeding | 4 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Sternthal et al. 2010 | N = 2,071; M + F | L | My child’s ETV scale (Selner-O’ Hagan et al., 1998) | Parent-reported lifetime asthma diagnosis | A, G, R, maternal Edu, maternal asthma, family violence in the home, neighborhood disadvantage, social disorder, collective efficacy | 8 | + |
Age 0–9 | |||||||
14.5% WH; 34.1% AF; 32.1% Mexican HI; 15.7% Non-Mexican HI; 3.7% OT | |||||||
Swahn and Bossarte 2006 | N = 1,943; M + F | C | Victimization—three items | Parent-reported lifetime asthma diagnosis and attack in the past year | G, R, grade, SS, self-perceived overweight status, exercising patterns | 6 | + |
No age range reported in article | |||||||
2003 Nationally representative sample | |||||||
Wilson et al. 2005 | N = 769; M + F | C | Modified version of ECV Scale (Richters & Saltzman, 1990) | Self-reported URI symptoms in the past 2 months | None | 5 | + |
Age 16–20 | |||||||
45% AA; 9% AS; 24% HI; 5% WH; 17% OT | |||||||
Yakubovich et al. 2016 | N = 6,002; M + F | C | Adapted version of the child ECV checklist (Richters & Martinez, 1990) | Self-reported asthma in the past year | A, G, urban/rural location, and province | 2 | + |
Age 10–17 | |||||||
100% AF | |||||||
Cardiovascular health | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Laboratory-measured BP and pulse | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Ford and Browning 2014 | N = 7,971; M + F | L | Victimization—three items; witnessing violence—one item | Hypertension status | Child R, A, foreign birth, general health, perpetrator of violence past, pregnancy status, inappropriate cuff, antihypertensive med use, resting HR; parent Edu, MS, two-parent household, economic hardship, public assistance | 4 | + |
Age 11–17 | |||||||
1994–1995 Nationally representative sample | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Mean HR, HR variability, SBP, and DBP | G, R | 5 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Wilson et al. 1998 | N = 40; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | SBP, DBP, and ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
Wilson et al. 2002 | N = 56; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | Daytime and nighttime ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
General health | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Teacher- and self-reported somatic complaints | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, Anx/Dep, number living in child’s home, HCA; maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Fredland et al. 2008 | N = 309; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Saltzman, 1990) | Self-reported physical health | G, R, poverty, maturation | 3 | 0 |
Age 11–14 | |||||||
95% AA | |||||||
Graham-Bermann and Seng 2005 | N = 160; M + F | C | Survey of ECV scale, parent version | Parent-reported physical health | G, monthly I, minority status, MS | 3 | 0 |
Age 4–6 | |||||||
52% WH; 44% AA or MI; 4% OT | |||||||
Halpern et al. 2013 | N = 8,531; F | C | Witnessing victimization—one itemDirect victimization—five items | Somatic complaints in the past year | A at Wave 1, self-reported R, parent Edu | 2 | + |
Age 11–21 | |||||||
24% AA; 57.9% WH; 18.1% HI | |||||||
Hart et al. 2013 | N = 409; M + F | L | The CREV ( Cooley et al., 1995) | Clinically significant somatic complaints | A, G, R, SES | 5 | + |
Age 8–13 | |||||||
85.6% AA | |||||||
Swisher and Warner 2013 | N = 20,103; M + F | L | Witnessing violence and violent victimization composite—five items | Youth-reported survival expectancies | A, G, R, immigrant status | 4 | + |
Mean age 15.7 at W1 | |||||||
1995 Nationally representative sample | |||||||
HPA axis functioning | |||||||
Aiyer et al. 2014 | N = 266; M + F | L | Witnessing violence and violent victimization combined—five items | Salivary cortisol AUC | G, R, interview start time | 7 | + |
Age 13.9–16.9 | |||||||
80% AA; 17% WH; 3% MI | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Salivary cortisol levels | G, R | 3 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Suglia et al. 2010 | N = 43; M + F | L | Modified version of the ECV scale (Richters & Saltzman, 1990) | Salivary cortisol levels, cortisol AUC, and diurnal slope | A, G, R, SES | 7 | 0 |
Age 7–13 | |||||||
54% WH; 46% HI | |||||||
Immune functioning | |||||||
Kacanek et al. 2016 | N = 268; M + F | C | Life Events Checklist—seven items (LEC; Kang, Mellins, Dolezal, Elkington, & Abrams, 2011) | Laboratory-measured unsuppressed viral load and CD4% | Child A, G, R, child knowledge of HIV status, annual household I, caregiver Edu, caregiver MS, caregiver relationship to child, presence of caregiver’s partner/spouse in home, HIV clinical characteristics, highly active antiretroviral therapy, adherence/nonadherence | 3 | + |
Age 8–15 | |||||||
71% AA; 21% HI | |||||||
Sleep | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Martinez, 1990) | Self-reported frequency of difficulty sleeping | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, anx/dep, number living in child’s home, HCA, maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Self-reported sleep disturbances | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Kliewer and Lepore 2015 | N = 362; M + F | L | Survey of children’s ECV scale (Richters & Saltzman, 1990) | Self-reported sleep problems | A, G, life events, family structure, intervention condition, school | 5 | + |
Age 11–14 | |||||||
51% HI; 34% AA; 8% WH; 5% AM; 6% AS | |||||||
Rubens et al. 2014 | N = 144; M + F | C | Witnessing violence—five items | Self-reported sleep problems | G, time spent in the United States | 1 | + |
Age 14–19 | |||||||
100% HI | |||||||
Spilsbury et al. 2014a | N = 46; M + F | L | REVS (van Dulmen et al., 2008)—items related to witnessing violence only | Self-reported and actigraphy-measured sleep quantity and quality | A, G, I, time between index-event and baseline REVS score, vacation status, PTS score, CWWVP service use | 7 | + |
Mean age 11.4% | |||||||
60.9% AA; 23.9% WH; 15.2% OT | |||||||
Umlauf et al. 2015 | N = 12,263; M + F | L | Witnessing violence—two items | Self-reported sleep problems | A, G, hopelessness | 4 | + |
Age 9.75–19.25 | |||||||
93.5% AA; 6% MI; <1% WH or AS | |||||||
Weight | |||||||
Gooding et al. 2015 | N = 147; M + F | C | Screen for Adolescent Violence Exposure (SAVE; Hastings & Kelley, 1997) | BMI | R, parent Edu, SS, weekly physical activity, frequency of fast food consumption, frequency of sugar-sweetened beverages consumption, internalizing problems | 1 | 0 |
Age 13–17 | |||||||
42.9% WH; 17.7% AA; 17% HI; 22.5% OT |
Reference . | Sample . | Study design . | Community violence measure . | Physical health outcome(s) . | Covariates . | Quality ratings . | Resultsa . |
---|---|---|---|---|---|---|---|
Asthma or respiratory health | |||||||
Clougherty et al. 2007 | N = 413; M + F | L | My child’s ETV scale (Selner-O’ Hagan, Kindlon, Buka, 1998) | Parent-reported asthma diagnosis | Maternal asthma, Edu, smoking before and after pregnancy, child G, A | 7 | 0 |
Age 5–8 | |||||||
52% HI; 44% WH | |||||||
Cohen et al. 2008 | N = 1,213; M + F | L | Modified version of the ECV Scale (Richters & Saltzman, 1990) | Parent- or child-reported past-year asthma diagnosis and/or lifetime allergic rhinitis diagnosis | A, G, SES, maternal/paternal history of asthma, caregiver-perceived stress | 8 | 0 |
Age 6.8–16.6 | |||||||
100% Puerto Rican | |||||||
Hellyer et al. 2013 | N = 365; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Parent- and child-reported asthma prevalence | A, G, R | 3 | 0 |
Age 9–11 | |||||||
51.5% HI; 26.8% AF; 50.7% WH | |||||||
Ramratnam et al. 2015 | N = 466; M + F | C | Exposure to Violence Survey—modified version | Asthma (physician diagnosed in past year) and allergens (serum levels of IgE specific to dust mite, cockroach, cat dander, dog dander, and mouse urinary protein | A, G, SES, parental asthma, prematurity, environmental tobacco smoke, distance from major roadway | 5 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Rosas-Salazar et al. 2016 | N = 747; M + F | C | Exposure to Violence Survey—one item (Selner-O’ Hagan et al., 1998) | Caretaker-reported asthma; laboratory-measured IgE levels | Child A, R, G, annual household I, maternal history of asthma, early life exposure to environmental tobacco smoke, exposure to gun violence, study site, child BMI, child obesity, child history of premature birth, Caesarian section birth, child low birth weight, parent Edu, current exposure to environmental tobacco smoke, breast-feeding | 4 | + |
Age 9–14 | |||||||
100% Puerto Rican | |||||||
Sternthal et al. 2010 | N = 2,071; M + F | L | My child’s ETV scale (Selner-O’ Hagan et al., 1998) | Parent-reported lifetime asthma diagnosis | A, G, R, maternal Edu, maternal asthma, family violence in the home, neighborhood disadvantage, social disorder, collective efficacy | 8 | + |
Age 0–9 | |||||||
14.5% WH; 34.1% AF; 32.1% Mexican HI; 15.7% Non-Mexican HI; 3.7% OT | |||||||
Swahn and Bossarte 2006 | N = 1,943; M + F | C | Victimization—three items | Parent-reported lifetime asthma diagnosis and attack in the past year | G, R, grade, SS, self-perceived overweight status, exercising patterns | 6 | + |
No age range reported in article | |||||||
2003 Nationally representative sample | |||||||
Wilson et al. 2005 | N = 769; M + F | C | Modified version of ECV Scale (Richters & Saltzman, 1990) | Self-reported URI symptoms in the past 2 months | None | 5 | + |
Age 16–20 | |||||||
45% AA; 9% AS; 24% HI; 5% WH; 17% OT | |||||||
Yakubovich et al. 2016 | N = 6,002; M + F | C | Adapted version of the child ECV checklist (Richters & Martinez, 1990) | Self-reported asthma in the past year | A, G, urban/rural location, and province | 2 | + |
Age 10–17 | |||||||
100% AF | |||||||
Cardiovascular health | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Laboratory-measured BP and pulse | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Ford and Browning 2014 | N = 7,971; M + F | L | Victimization—three items; witnessing violence—one item | Hypertension status | Child R, A, foreign birth, general health, perpetrator of violence past, pregnancy status, inappropriate cuff, antihypertensive med use, resting HR; parent Edu, MS, two-parent household, economic hardship, public assistance | 4 | + |
Age 11–17 | |||||||
1994–1995 Nationally representative sample | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Mean HR, HR variability, SBP, and DBP | G, R | 5 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Wilson et al. 1998 | N = 40; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | SBP, DBP, and ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
Wilson et al. 2002 | N = 56; M + F | C | Modified version of ECV scale (Richters & Saltzman, 1990) | Daytime and nighttime ambulatory BP | A, G, BMI | 5 | + |
Age 11–18 | |||||||
100% AA | |||||||
General health | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version of “TISH” scale—eight items (Richters & Martinez, 1990) | Teacher- and self-reported somatic complaints | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, Anx/Dep, number living in child’s home, HCA; maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Fredland et al. 2008 | N = 309; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Saltzman, 1990) | Self-reported physical health | G, R, poverty, maturation | 3 | 0 |
Age 11–14 | |||||||
95% AA | |||||||
Graham-Bermann and Seng 2005 | N = 160; M + F | C | Survey of ECV scale, parent version | Parent-reported physical health | G, monthly I, minority status, MS | 3 | 0 |
Age 4–6 | |||||||
52% WH; 44% AA or MI; 4% OT | |||||||
Halpern et al. 2013 | N = 8,531; F | C | Witnessing victimization—one itemDirect victimization—five items | Somatic complaints in the past year | A at Wave 1, self-reported R, parent Edu | 2 | + |
Age 11–21 | |||||||
24% AA; 57.9% WH; 18.1% HI | |||||||
Hart et al. 2013 | N = 409; M + F | L | The CREV ( Cooley et al., 1995) | Clinically significant somatic complaints | A, G, R, SES | 5 | + |
Age 8–13 | |||||||
85.6% AA | |||||||
Swisher and Warner 2013 | N = 20,103; M + F | L | Witnessing violence and violent victimization composite—five items | Youth-reported survival expectancies | A, G, R, immigrant status | 4 | + |
Mean age 15.7 at W1 | |||||||
1995 Nationally representative sample | |||||||
HPA axis functioning | |||||||
Aiyer et al. 2014 | N = 266; M + F | L | Witnessing violence and violent victimization combined—five items | Salivary cortisol AUC | G, R, interview start time | 7 | + |
Age 13.9–16.9 | |||||||
80% AA; 17% WH; 3% MI | |||||||
Murali and Chen 2005 | N = 115; M + F | C | Exposure to Violence Interview (Thompson et al., 2002) | Salivary cortisol levels | G, R | 3 | + |
Age 16–19 | |||||||
55% AA; 42% WH; 3% OT | |||||||
Suglia et al. 2010 | N = 43; M + F | L | Modified version of the ECV scale (Richters & Saltzman, 1990) | Salivary cortisol levels, cortisol AUC, and diurnal slope | A, G, R, SES | 7 | 0 |
Age 7–13 | |||||||
54% WH; 46% HI | |||||||
Immune functioning | |||||||
Kacanek et al. 2016 | N = 268; M + F | C | Life Events Checklist—seven items (LEC; Kang, Mellins, Dolezal, Elkington, & Abrams, 2011) | Laboratory-measured unsuppressed viral load and CD4% | Child A, G, R, child knowledge of HIV status, annual household I, caregiver Edu, caregiver MS, caregiver relationship to child, presence of caregiver’s partner/spouse in home, HIV clinical characteristics, highly active antiretroviral therapy, adherence/nonadherence | 3 | + |
Age 8–15 | |||||||
71% AA; 21% HI | |||||||
Sleep | |||||||
Bailey et al. 2005 | N = 268; M + F | C | Child-adapted version “TISH” scale—eight items (Richters & Martinez, 1990) | Self-reported frequency of difficulty sleeping | G, SES, child prenatal and postnatal alcohol exposure, chronic medical condition, anx/dep, number living in child’s home, HCA, maternal A, MS, Edu, psychopathology, life stress | 3 | + |
Age 6–7 | |||||||
100% AA | |||||||
Cooley-Quille and Lorion 1999 | N = 64; M +F | C | CREV (Cooley et al., 1995) | Self-reported sleep disturbances | G | 4 | + |
Age 16–18 | |||||||
84% AA | |||||||
Kliewer and Lepore 2015 | N = 362; M + F | L | Survey of children’s ECV scale (Richters & Saltzman, 1990) | Self-reported sleep problems | A, G, life events, family structure, intervention condition, school | 5 | + |
Age 11–14 | |||||||
51% HI; 34% AA; 8% WH; 5% AM; 6% AS | |||||||
Rubens et al. 2014 | N = 144; M + F | C | Witnessing violence—five items | Self-reported sleep problems | G, time spent in the United States | 1 | + |
Age 14–19 | |||||||
100% HI | |||||||
Spilsbury et al. 2014a | N = 46; M + F | L | REVS (van Dulmen et al., 2008)—items related to witnessing violence only | Self-reported and actigraphy-measured sleep quantity and quality | A, G, I, time between index-event and baseline REVS score, vacation status, PTS score, CWWVP service use | 7 | + |
Mean age 11.4% | |||||||
60.9% AA; 23.9% WH; 15.2% OT | |||||||
Umlauf et al. 2015 | N = 12,263; M + F | L | Witnessing violence—two items | Self-reported sleep problems | A, G, hopelessness | 4 | + |
Age 9.75–19.25 | |||||||
93.5% AA; 6% MI; <1% WH or AS | |||||||
Weight | |||||||
Gooding et al. 2015 | N = 147; M + F | C | Screen for Adolescent Violence Exposure (SAVE; Hastings & Kelley, 1997) | BMI | R, parent Edu, SS, weekly physical activity, frequency of fast food consumption, frequency of sugar-sweetened beverages consumption, internalizing problems | 1 | 0 |
Age 13–17 | |||||||
42.9% WH; 17.7% AA; 17% HI; 22.5% OT |
Note. M = male; F = female; AM = American Indian or Alaskan Native; AS = Asian; AF = African; AA = Black or African–American; HI = Hispanic/Latino; WH = White or of Caucasian decent; MI = mixed race; OT = other; C = cross-sectional; L = longitudinal; URI = upper respiratory infection; BP = blood pressure; BMI = body mass index; PTS = posttraumatic stress symptoms; HCA = history of child abuse; Anx/Dep = anxiety or depression; SS = smoking status; CWWVP, Children Who Witness Violence Program; SBP = systolic blood pressure; SES = socioeconomic status; R = race/ethnicity; G = sex; A = Age; Edu = education level; MS = marital status; Emp = employment status; I = income; DBP = diastolic blood pressure; HR = heart rate; HRV = heart rate variability; AUC = area under the curve; REVS = Recent Exposure to Violence Scale; CREV = Children’s Report of Exposure to Violence; TISH = Things I Have Seen and Heard.
+ Refers to a statistically significant relationship between ECV and health outcome; 0, no significant relationship between ECV and health outcome.
We did not complete a formal meta-analysis of the relevant literature for several reasons. First, there was considerable heterogeneity in the assessment of ECV with some investigators using sophisticated analytical techniques to model exposure while others quantified exposure with a single item. Second, because a prior review had not been conducted in this area and we intentionally cast a wide net with regard to the physical health outcomes included in the review, there was considerable heterogeneity in outcomes across studies. Consequently, in our review, we summarize patterns observable across the range of health outcomes studied.
Results
Twenty-eight studies were carefully reviewed and included in our final consideration and evaluation of results. Studies fell into seven categories of physical health outcomes: asthma or respiratory health, cardiovascular health, immune functioning, hypothalamic–pituitary–adrenal (HPA) axis functioning, sleep, weight, and a general health category. Note that we included sleep and weight because they can be considered health outcomes in their own right, although they also impact other health conditions. Physical health outcomes including injury and morbidity, cancer, pain, and traumatic brain injury could not be reviewed because of a lack of research evaluating these areas of health as consequences of ECV. Two studies focused on multiple health outcomes and therefore are summarized in multiple categories. Twenty-one studies had significant findings, while seven did not. A variety of methodological strategies were used, with roughly 60% of the studies using cross-sectional and 40% of the studies using longitudinal designs. A wide range of youth age and racial backgrounds were sampled. Of the 28 studies, 17 used a validated measure of ECV, and 15 used an objective measure of health outcomes, as opposed to parent or self-report. All but two studies controlled for covariates. The demographics, study design, community violence measure, physical health measures, covariates, and quality ratings are summarized in Table I.
Asthma or Respiratory Health
With respect to the association between ECV and asthma or respiratory health, there does not seem to be a clear consensus among the research to date. Of the nine reviewed articles in this category, six found significant associations between ECV and the measured health outcomes (Ramratnam et al., 2015; Rosas-Salazar et al., 2016; Sternthal, Jun, Earls, & Wright, 2010; Swahn & Bossarte, 2006; Wilson, Rosenthal, & Austin, 2005; Yakubovich, Cluver, & Gie, 2016), while three did not (Cohen, Canino, Bird, & Celedon, 2008; Clougherty et al., 2007; Hellyer, Garrido, Petrenko, & Taussig, 2013). No clear pattern of results emerged when examining quality rating of studies, study design, or participant demographics.
Sternthal et al. (2010) found that medium to high levels of ECV were associated with increased asthma risk. After controlling for covariates (see Table I), being of African–American race remained independently associated with increased asthma risk. When collective efficacy was added into the model, however, the effects of race approached nonsignificance. Collective efficacy refers to the level of trust among residents and their perceived willingness to engage in collective action (Sternthal et al., 2010). The authors interpreted their results to mean that the feeling of low collective efficacy contributes to the increased burden of community violence felt by African–American youth. Race also was considered in Rosas-Salazar et al.’s (2016) study, revealing that incremental increases in African ancestry were associated with increased odds of asthma, but only among those exposed to gun violence.
Ramratnam et al. (2015) found that children who experienced a combination of hearing gunshots more than once and being afraid to leave home because of violence had 3.2 times greater odds of asthma than children who did not experience either. Yakubovich et al. (2016) conducted a study examining ECV as one of multiple risk factors for asthma. The final risk model showed that less severe household poverty, increased ECV, fewer outdoor tasks, and higher ratings of anxiety symptoms were significant predictors of increased asthma prevalence.
Secondary analyses by Swahn and Bossarte (2006) found that any victimization in the past year significantly increased odds of having an asthma attack in the same year. However, when results were stratified by urban versus rural setting, significant effects only were seen for youth living in urban areas. Another study investigating upper respiratory illness (URI) found that rates of URI symptoms during the first 2 months of college were associated positively with rates of adolescent ECV when they were in high school (Wilson et al., 2005).
Clougherty et al. (2007) found that ECV did not have a significant direct effect on child asthma diagnosis. However, high levels of air pollution predicted an asthma diagnosis only in children with elevated ECV levels. In a study of children living in foster care, ECV did not predict asthma over and above covariates (Hellyer et al., 2013). The researchers reasoned that because foster care children are already at a high risk of chronic health problems, variables such as ECV would not be a strong enough contributor to account for significantly more variance. Cohen et al. (2008) did not find significant effects of ECV on asthma and allergy symptoms with their sample. ECV was one level of exposure on a scale including witnessing neighborhood violence, direct victimization through neighborhood violence, and history of child abuse within the home. The researchers found significant effects for the victims of child abuse, but not the other two categories. Results were interpreted to mean that severity of traumatic stressor could be a moderator of the effect on asthma and allergies, with the assumption that child abuse is a more severe traumatic event than is ECV.
Cardiovascular Health
When it came to cardiovascular health, all studies found significant effects. Higher ECV was associated with elevated blood pressure (BP) in four of five studies (Ford & Browning, 2014; Murali & Chen, 2005; Wilson, Kliewer, Teasley, Plybon, & Sica, 2002; Wilson et al., 1998). Quality ratings of studies were in the 4–5 range. Wilson and colleagues’ (1998) study of 40 African–American youth found significant positive associations between ECV and systolic BP (SBP) levels, even when controlling for body mass index (BMI). Nondipping status, which is a lack of the typical nocturnal BP pattern involving lowered BP in the time from sleep to wake, has been correlated with worsened cardiovascular outcomes in adults (Viera et al., 2012). Wilson et al. (2002) found significant positive associations between ECV and BP dipping status, with sex differences. Males exhibiting nondipping status showed stronger positive associations between witnessing violence and daytime epinephrine levels than did females. Ford and Browning’s (2014) study found that compared with those without ECV in the previous year, participants with ECV in the year before Wave 1 had significantly increased odds of developing hypertension by Wave 4.
Murali and Chen (2005) examined various measures of cardiovascular activity and their association with ECV at baseline and in response to an acute stress task, a measure of cardiovascular reactivity. Measures of cardiovascular activity included SBP, diastolic BP (DBP), heart rate (HR), and HR variability (HRV)—a measure of the variation in the time interval between heartbeats, reflecting activation of the parasympathetic nervous system. They found that total ECV was significantly associated with higher basal DBP. Frequency of experienced violence, a subcomponent of total ECV, was associated with increased basal DBP and HR. In response to the acute stress task, they found that increased total ECV, and various subcomponents of ECV (frequency of experienced violence, proximity to violence, and total observed violence), was associated with decreased cardiovascular reactivity. Interestingly, ECV was positively associated with an HRV measure of cardiovascular reactivity in males only. Cooley-Quille and Lorion (1999) contrastingly showed a decrease in cardiovascular activity at baseline, measured via resting HR, in youth who had experienced the highest rates of ECV. In this study, ECV showed no effect on HR or BP during a stressful interview task.
General Health
Several studies reported research linking ECV with a variety of health outcomes, classified here as general health, offer mixed results. This is not surprising, as this category includes dramatically different “general” health outcome variables such as perceptions of overall physical health and somatic complaints. Four of these studies reported significant associations between ECV and general health (Bailey et al., 2005; Halpern et al., 2013; Hart, Hodgkinson, Belcher, Hyman, & Cooley-Strickland, 2013; Swisher & Warner, 2013) and two did not (Fredland, Campbell, & Han, 2008; Graham-Bermann & Seng, 2005). It is of note that the significant effects were found in the longitudinal studies and not in the cross-sectional studies. The longitudinal studies had larger sample sizes and were of slightly higher design quality.
Swisher and Warner (2013) found that with each unit increase in measured ECV, there was a 9.7% decrease in the odds of self-reported expectations of surviving to age 35 years. Hart et al. (2013) conducted secondary data analyses on 409 children, finding that ECV was associated with increased odds of having a borderline or clinically significant level of somatic complaints, measured by the Child Behavior Checklist, even when controlling for race, sex, age, and SES. Halpern et al. (2013) examined the effects of ECV and sexual violence on somatic symptoms in an entirely female sample and found that those who experienced both types of violence were the most likely to experience high somatic symptomatology. The next highest predictor of high symptomatology was for females who experienced only sexual violence, and third highest for those who had experienced ECV.
In a study of multiple health-related outcomes among urban African–American children, Bailey et al. (2005) found that victims of community violence had a 28% increased risk of appetite problems and a 174% increased risk of stomachache complaints compared with children who reported no victimization. Additionally, when participants were divided by level of witnessed community violence, children in the upper 50% levels of witnessed violence had a 57% increased risk of headache complaints compared with children in the lower 50%.
The above research was in contrast to that of Fredland et al. (2008), who conducted a study with predominantly African–American youth focused on the impact of ECV on a number of outcomes, including physical health outcomes. Physical health variables included asthma, diabetes, sleep disturbances, cold and flu symptoms, headaches, stomachaches, pain, and fatigue. The researchers did not find ECV to be significantly associated with any health measures. However, this was attributed to a lack of variability in level of ECV.
Graham-Bermann and Seng (2005) conducted a study with similar health outcomes to Fredland et al. (i.e., asthma, flu, cold, allergies, Attention-Deficit/Hyperactivity Disorder, somatic complaints, gastrointestinal problems, and headaches). However, in this study, health symptoms were summed to create ratings of overall child health status. After controlling for covariates (see Table I), analyses revealed that ECV was not a significant predictor of the number of health problems.
Immune Functioning
Our body’s immune system is the means by which we are able to fight off disease and maintain health. For years, studies have documented the negative effect stress plays on immune functioning, decreasing our ability to fight illness (Leonard, 1988). Kacanek et al. (2016) conducted a study with 268 youth with perinatal HIV (PHIV) who had experienced ECV. After controlling for multiple covariates (see Table I), results showed that youth who experienced ECV in the past year had elevated odds of unsuppressed viral load. Viral load is an indicator of the level of viral cells within the body. Unsuppressed viral load indicates a difficulty or inability of the body to fight viral infection and maintain health. Kacanek et al. (2016) highlighted the important fact that children born with PHIV are often also born into lower SES neighborhoods where they are more likely to encounter stressors including ECV, which hinder the body’s ability to maintain health over time.
HPA Axis Functioning
The HPA axis is the primary biological stress response system for physical and psychological stressors. When stressors are encountered, a cascade of signaling hormones such as cortisol is produced. Chronic stress exposure has been shown to result in HPA axis dysregulation consisting of both hyperreactivity, increased HPA axis activation, and hyporeactivity, or decreased HPA axis activation in response to a stressor (Rao, Hammen, Ortiz, Chen, & Poland, 2008). A variety of methods were used to operationalize HPA axis dysregulation by cortisol production in each study reviewed. Despite this, evidence suggests that ECV disrupts HPA axis functioning, as two of three studies found significant effects of ECV on cortisol measures (Aiyer et al., 2014; Murali & Chen, 2005). The two studies with significant findings involved a primarily African–American sample, while the one study without significant effects used a predominantly white sample (Suglia, Staudenmayer, Cohen, & Wright, 2010). Participant age and study quality rating revealed no patterns of significant effects.
A previously mentioned study by Murali and Chen (2005) also examined basal salivary cortisol concentration levels (nMol/L) and cortisol levels following a laboratory-based acute stress task (cortisol reactivity). Higher frequency of ECV was significantly associated with increased basal cortisol levels for all participants. Aiyer et al. (2014) measured salivary cortisol concentration (nMol/L) at three time points during a single day, and used cortisol area under the curve (AUC) as an estimate of total cortisol output over the course of that day. This study also supported HPA axis dysregulation in response to ECV as youth who experienced ECV exhibited lowered cortisol AUC. This relationship was moderated by sex, with males demonstrating a stronger negative relationship between ECV and cortisol AUC relative to females. Suglia et al.’s (2010) study did not find significant associations between ECV and any of the three cortisol measures assessed on three different days: cortisol concentration (nMol/L), cortisol AUC, and diurnal slope—the average rate of decline in cortisol concentration from waking to sleep.
Sleep
Sleep was another category of health in which all studies had significant findings, showing that sleep problems were positively associated with ECV (Bailey et al., 2005; Cooley-Quille & Lorion, 1999; Kliewer & Lepore, 2015; Rubens, Fite, Cooley, & Canter, 2014; Spilsbury, Babineau, Frame, Juhas, & Rork, 2014; Umlauf, Bolland, Bolland, Tomek, & Bolland, 2015). Study quality ratings ranged from 1 to 7, suggesting that sleep problems are significantly affected by ECV regardless of design quality.
Cooley-Quille and Lorion (1999) examined sleep outcomes in addition to their aforementioned cardiovascular outcomes, finding a positive correlation between ECV and self-reported sleep disturbances. Bailey et al.’s (2005) previously mentioned study also examined sleep outcomes. Results revealed that children who had been victims of community violence had a 94% increased risk of self-reported “difficulty sleeping.” Additionally, Bailey et al. (2005) found that children in the upper 50% levels of witnessed violence had a 79% increased risk of sleep problems compared with children in the lower 50% levels.
Spilsbury et al. (2014) found that youth reporting witnessing a homicide had a twofold increase in “wake after sleep onset”—the amount of time they were awake again after falling asleep during the night—at baseline compared with children who did not. However, these results were no longer statistically significant at follow-up. Kliewer and Lepore (2015) found that both community violence victimization and witnessing were associated with self-reported sleep problems for a sample of urban middle school students. Rubens et al. (2014) also found a significant association between ECV and sleep problems, even when controlling for covariates (see Table I).
Umlauf and colleagues (2015) also found significant associations between ECV and sleep problems, although unlike Rubens et al. (2014), sex differences emerged. In their study, high ECV had the strongest negative effects on sleep for females. The relationship between ECV and sleep decreased in strength as adolescents got older, perhaps indicating physiological adaption and desensitization to the effects of ECV on sleep.
Weight
Only one study meeting review criteria examined weight as a physical health outcome. Gooding, Milliren, Austin, Sheridan, and McLaughlin (2015) investigated the effects of ECV on BMI, which is calculated using an individual’s height and weight and is considered a rough indicator of healthy weight status. The authors reported a “marginally significant” (p = .07) association between ECV and BMI, though this association was attenuated when controlling for race and parent education. Owing to the small sample size, lack of validated and objective measures, and cross-sectional design, this study only received a quality rating of 1. Future studies with higher design quality may find different results.
Discussion
This systematic review of ECV and physical health outcomes in children and adolescence yielded several findings with implications for researchers and health care practitioners. First, the strongest study finding was that ECV had reliable associations with elevated BP and sleep problems. Interestingly, BP and sleep problems are outcomes in and of themselves, but when compromised over a long period, contribute to other health problems. ECV also was associated with unsuppressed viral load and with dysregulated cortisol response in two studies, which also contribute to additional health problems. Although reliable associations were observed between ECV and these health outcomes, the mechanisms explaining these associations were not examined. In thinking about viable mechanisms that could link ECV and health outcomes, threat appraisals seemed appropriate. Threat appraisals are evaluations of what is “at stake” in a situation (Lazarus & Folkman, 1984). Based on Lazarus and Folkman’s (1984) theory of stress processing, researchers have demonstrated that youth cognitively process the experience of violence exposure differently (e.g., Kliewer & Sullivan, 2008; Kliewer, Lepore, Oskin, & Johnson, 1998; Schwartz & Proctor, 2000), and its impact is determined in part by adolescents’ appraisals of the negative implications violence exposure has for their well-being. The health outcomes most consistently linked to ECV in our review—asthma, BP, and sleep problems—may be more sensitive to threat appraisals than other health outcomes. That is, ECV may activate an appraisal process that then has a nearly immediate effect on body physiology. Further, for any association between ECV and a health outcome to be observed, some physiological process has to be activated. When viewed this way, it is less surprising that an association between ECV and physical activity did not exist: the mechanism connecting them is not as apparent. Notably, overall, the studies with the strongest associations between ECV and health outcomes had samples that might be described as disadvantaged, suggesting that threat appraisals indeed may be the putative mediator.
A reexamination of the pattern of findings in the asthma or respiratory health category using this lens may help to underscore the viability of threat appraisals as a mediating, or in some cases, a moderating, process. The Ramratnam et al. (2015) study found that children who experienced a combination of hearing gunshots more than once and being afraid to leave home because of violence had elevated odds of asthma relative to youth and children who did not experience either. In the Yakubovich et al. (2016) study, increased ECV and higher ratings of anxiety symptoms were significant predictors of increased asthma prevalence. In the Swahn and Bossarte (2006) article, only urban youth exposed to violence were at increased risk for having an asthma attack.
From a research perspective, testing the hypothesis that threat appraisals account for the linkages between ECV and BP, sleep problems, or other health problems would provide data that could inform interventions to reduce negative health effects. Although eliminating primary ECV would be optimal, once youth are exposed to violence, knowing the extent to which this exposure activates threat appraisal processes and how strongly appraisal processes are linked to physical health is a step in interrupting the path between ECV and compromised health. This information could be useful to health care practitioners, but also school counselors and individuals who interact with youth exposed to violence.
Limitations of the Literature Reviewed
One limitation of the literature reviewed is the wide variability in measures of ECV. The studies included used a combination of validated and nonvalidated community violence measures, making it difficult to determine patterns in the current data. A second limitation is that only two of the reviewed studies controlled for family-level violence or gun exposure. This lack of covariate inclusion makes it difficult to isolate the impact of ECV on negative child health outcomes. Across all studies reviewed, there was some diversity in areas such as sample age ranges and race. However, there was little diversity in terms of environment (urban setting vs. rural) or family SES. These variables could impact children’s interpretation of the violence they are exposed to, and could influence subsequent health outcomes. To date, there are a limited number of studies examining these alternative environments and living arrangements to allow for conclusions to be drawn regarding an environmental interaction with ECV and health outcomes.
Limitations of the Review Strategy
The present review searched for a wide range of physical health outcomes to evaluate the ways in which ECV impacts numerous areas of health. As a result of the wide variability of studies identified, we were unable to conduct a meta-analysis, and effect sizes could not be quantitatively compared. A second limitation of the review concerns our search strategy. Key health- and violence-related words were identified and searched for within manuscript titles, abstracts, and key words. However, it is possible that studies examining physical health outcomes related to ECV as a subfinding of a larger study did not mention these results in the title or abstract and thus were missed in the present search. Finally, restricting studies to those published in peer-reviewed articles leaves the review subject to reporting bias such that inconclusive or null findings are excluded.
Recommendations for Future Research
The results of the present review highlight a number of trends and gaps in the current research. Below we offer recommendations related to (1) range in physical health outcomes studied, (2) methodology, (3) assessment of processes that account for linkages between ECV and physical health outcomes, and (4) assessment of factors that attenuate or exacerbate associations between ECV and physical health outcomes.
As noted in the results, there were numerous physical health outcomes that have been overlooked in the research to date. These include injury and morbidity, cancer, pain, and traumatic brain injury. Some studies have examined injury and morbidity outcomes in relation to crime data, and have inferred youth ECV, but have failed to evaluate actual youth exposure rates as predictors of these outcomes. The listed physical health outcomes are among top health concerns prevalent in childhood and adolescence. The complete lack of evidence supporting or refuting any link between ECV and these areas of health leaves a hole in our understanding and therefore should be examined in future research.
With respect to methodology, future research could improve on choice of measures of ECV, the range of physical health outcomes studied, and the nature of samples assessed. Several validated measures of ECV exist, with Richters and Saltzman’s (1990)Survey of Exposure to Community Violence being used most widely. Thus, a critical point is that researchers use a validated measure of ECV. Additionally, it is of note that 36 otherwise eligible studies were excluded because they failed to measure youth ECV; instead, these studies measured caregiver ECV or caregiver perception of their child’s safety. While caregiver experience offers insight into familial perspectives on community violence, reliance on caregiver report hinders the development of conclusions about child experience of violence. Thus, using a validated measure specific to examining child ECV is necessary to improve current knowledge of the nature of associations and processes linking ECV and health outcomes in youth.
In terms of all health categories reviewed, the majority of research focused on asthma and sleep outcomes. Areas of health including HPA axis functioning, weight, and immune functioning still require significant investigation before it can be determined whether ECV has an effect on these health outcomes. The samples assessed evidenced gaps in sample diversity, particularly as pertains to rural versus urban setting. We recommend that future studies examine the association between ECV and youth health outcomes in rural and urban settings. It is possible that youth in urban settings interpret ECV differently than rural youth. All but two studies reviewed consisted of an urban sample. It would be useful for future studies to determine whether these effects are unique to urban areas or transferable to rural settings, to thereby inform intervention and dissemination of research.
Second, future studies should move beyond just evaluating the relationships between community violence and health on a descriptive level and begin to focus on processes that account for linkages between ECV and physical health outcomes. Few studies appear to have examined these underlying connections. Such work might take two directions. As noted above, based on stress and coping theory (Lazarus & Folkman, 1984), researchers might focus on cognitive and behavioral processes such as appraisal, coping efforts, or coping efficacy. A second direction might be to examine biological mechanisms that account for the observed associations, such as inflammation (Miller, Chen, & Parker, 2011). Understanding the putative processes linking ECV and health is essential if community interventions for improving children’s health are to be designed.
Finally, future research should assess factors that moderate associations between ECV and physical health outcomes. These might include demographic moderators such as sex and age, but it would be more meaningful to assess assets such as emotion regulation or coping, as these would help researchers to quantify youth vulnerability to the negative consequences of ECV. There is some evidence, for example, that youth with good emotion regulation skills do not exhibit dysregulated cortisol responses to community violence victimization in contrast to their peers with poor emotion regulation skills (Kliewer, 2016b).