Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jun 28.
Published in final edited form as: J Pediatr. 2015 Sep 26;167(6):1415–1421. doi: 10.1016/j.jpeds.2015.08.043

CHAOS and Health Risk Behaviors in Adolescents in Los Angeles

Avik Chatterjee 1, Matthew W Gillman 1, Mitchell D Wong 2
PMCID: PMC6023620  NIHMSID: NIHMS977033  PMID: 26394824

Abstract

Objective

Household chaos—physical and social disorder at home—is associated with poor self-regulatory behavior in children. We sought to determine the relationship between household chaos and substance use, sexual activity and violence-related risk behaviors in adolescents.

Study Design

We analyzed cross-sectional data among 929 high-school students in Los Angeles who completed a 90-minute interview assessing health behaviors and household chaos using the 14-question Chaos, Hubbub and Order Scale (CHAOS). Using the generalized estimating equation and adjusting for personal, parental, and family covariates, we examined associations of CHAOS score with substance use, sexual activity, and violent behavior outcome variables. We also examined the role of depression and school engagement as mediators.

Results

Mean (SD) age of the 929 students was 16.4 (1.3) years, 516 (55%) were female and 780 (84%) were Latino. After adjustment, compared to students with CHAOS score 0, those with the highest scores (5–14) had odds ratios of 3.1 (1.1,8.7) for smoking, 2.6 (1.6,4.4) for drinking, 6.1(1.8,21) for substance use at school, and 1.9(1.1,3.3) for fighting in the past 12 months. Associations between CHAOS score and sexual risk and other violent behaviors were not significant. Depression and school engagement attenuated the associations.

Conclusions

In this group of adolescents, highest CHAOS score was associated with increased odds of risky health behaviors, with depression and school engagement as potential mediators. In the future, CHAOS score could be measured to assess risk for such behaviors, or be a target for intervention to reduce chances of engaging in these behaviors

Keywords: adolescent health, substance abuse, chaos, health behaviors, violence

Introduction

Chaos has been defined as the overall physical, social and environmental disorder in a person’s life1 and may be an important but under-studied determinant of health. The Chaos, Hubbub and Order Scale (CHAOS) assesses the amount of disorder in the home environment,1 and has been linked longitudinally in children and adolescents to increased psychological distress,2 learned helplessness,3 and poor self-regulatory ability.4 A chaotic home environment has also been shown to be longitudinally associated with more conduct problems and hyperactivity-inattention in childhood and early adolescence.5 Despite the evidence that household chaos is associated with the above poor psychosocial outcomes in children, we are unaware of studies examining the relationship between household chaos and health behaviors in adolescents.

Risky health behaviors—such as substance use, sexual risk behaviors, and violent behaviors—are often acquired during adolescence6 and contribute significantly to morbidity and mortality.7,8 According to Youth Risk Behavior Surveillance System (YRBSS) data from 2013, 16% of teens have smoked cigarettes in the past 30 days, 34% report having been sexually active, and 18% have carried weapons in the past 30 days.6 Household chaos, through effects on self-regulatory ability, disruptive behaviors, or other mechanisms, could be a predisposing factor for risky health behaviors in adolescents, or could be an important downstream effect of risky behaviors, but has not been studied in this context. Understanding the association of chaos and adolescent health risk behaviors is an important first step in exploring the relationship.

The objective of our study was to determine whether household chaos is associated with self-reported engagement in health risk behaviors in a sample of low-income, minority high school students. We hypothesized that higher CHAOS score, because it is associated with increased psychological distress and decreased self-regulatory ability in adolescents, would ultimately be associated with higher risk of substance use, risky sexual behaviors, and violent behaviors.

Methods

The RISE (Reducing Health Inequalities through Social and Educational Change) study was a cross-sectional natural experimental study originally designed to examine the impact of exposure to high performing school environments on risky behaviors among low-income adolescents.9 In 2010 and 2011, the study sampled students who had applied to enter one of three high-performing charter high schools in Los Angeles between 2007–2010. Admission into these schools was based on a lottery. The RISE study compared those who were and were not accepted into these charter schools, and thus included some students who attended traditional public schools. We examined data from all participants who completed the face-to-face survey.

Population/Participants

Investigators accessed lists of applicants to the 9th grade lottery for the three charter schools in 2007–2010. Investigators randomly selected 2384 students from the list of applicants, 952 of whom had been accepted to attend charter schools by lottery, and 1432 of whom had not. Of the potential participants, 410 could not be contacted, and 736 were ineligible because they attended a private school, a charter school outside of the lottery, or received preference in the charter school lottery because they had a sibling at the charter school. Of the remaining students, 308 refused to participate (24.9% refusal rate). We excluded 1 subject who only answered the demographic questions, leaving 929 students who completed the interviews and whose data we analyzed.9

Measures

After receiving written consent, research assistants administered a 90-minute interview with students assessing their demographics, school information, depression (the 20-item Center for Epidemiologic Studies Depression Scale),10 school engagement,11 and parenting style.12 Investigators measured household chaos using the Chaos, Hubbub and Order Scale (CHAOS) questionnaire, 14 yes-no items such as “we almost always seem to be rushed,” and “it’s a real zoo in our home” (Table 1).1 The CHAOS questionnaire was derived from observations and parent interviews from hundreds of home visits during the Louisville Twin Study in the 1980’s1,13 and subsequently validated and used widely in children and adolescents.1,5,14 in a sample of 123 families in the Louisville study, the CHAOS questions had a Cronbach’s alpha of 0.79, and when validated against direct observation in a sample of 52 families, had a high correlation with observer-reported measures of household chaos.1 The CHAOS questionnaire has also been shown to be reliable and valid in Latino families. In a study of the CHAOS questionnaire in the families of 68 Latino children, the instrument was found to have similar reliability (Cronbach’s alpha of 0.79) and high convergent validity against other measures of parental stress and child behavior.15

Table 1.

The Chaos, Hubbub and Order Scale (CHAOS) Questionnaire (Evans, Gonnella, Marcynyszyn, Gentile, & Salpekar, 2005).

For each of the following statements, please tell me whether you think it is true or false for you.

  1. There are very few disturbances in our home.

  2. We can usually find things when we need them.

  3. We almost always seem to be rushed.

  4. We are usually able to stay on top of things.

  5. No matter how hard we try, we always seem to be running late.

  6. It’s a real zoo in our home.

  7. At home we can talk to each other without being interrupted.

  8. There is a lot of needless worrying going on in our home.

  9. No matter what our family plans, it usually doesn’t seem to work out.

  10. It’s so noisy, you can’t hear yourself think in our home.

  11. I often get drawn into other people’s arguments at home.

  12. Our home is a good place to relax.

  13. The telephone takes up a lot of our time at home.

  14. The atmosphere in our home is calm.

Open in a new tab

Research assistants conducted interviews in as private a setting as possible, at the student’s home, school, or in a public place without parents present. All participants were able to complete the interview in English. Subjects completed sensitive portions of the interview, including report of the health risk behaviors via audio-enhanced computer-assisted self-interview (A-CASI). A-CASI has been shown to be more sensitive than in-person interview for sensitive topics.16 RISE Investigators drew questions about health risk behaviors from the Youth Risk Behavior Survey.17 Using an iPad, students could themselves read and then respond to the questions, or click a button to have the computer read the question to them, and then respond.

Primary outcomes included questions about substance use (tobacco use, alcohol use or marijuana use in the past 30 days, or any substance use at school in the past 30 days), sexual activity (whether students had had sexual activity in the past 90 days, whether they had used contraception at their last sexual encounter, or whether they had used substances during sex in the past 90 days), and violent behaviors (whether students had carried a weapon in the past 30 days, whether the student had been involved in a gang in the past 12 months, or whether the student had been in a fight in the past 12 months).

Data Analysis

We summed the fourteen individual CHAOS items to create a composite score on a 0–14 scale. In preliminary bivariate analysis, we found the relationship between CHAOS and risk behavior outcomes to be non-dichotomous and non-linear, and given the sample size and prevalence of primary outcomes, we divided the subjects into four roughly equal groups by CHAOS score—score 0, 1–2, 3–4, and 5–14). We dichotomized responses to questions about risky health behaviors.

We used the generalized estimating equation to examine the relationship between chaos and risky health behaviors, accounting for clustering at the school level by adding terms for fixed effects of each school. We sequentially added sets of covariates to control for pre-specified individual, family history, parental socio-economic factors, and school type (traditional public vs. charter). We pre-specified covariates based on prior evidence of association with CHAOS score, health risk behavior or both—it is known, for example, that age and race/ethnicity are associated with substance use in adolescents,6 as are socioeconomic status,18 parenting style,19 and family history.20 Model 0 is unadjusted for any covariates. In Model 1, we added the individual-level covariates grade in school, sex, race/ethnicity, and school type. In Model 2, we included the parental-level covariates parenting style and student-report of whether at least one parent works. In Model 3, we included student report of whether at least one parent smoked or whether a family member had an alcohol or drug problem. In Model 4, we introduced the potential mediators of depression and school engagement to look for attenuation in the relationship between CHAOS score and risk behaviors.

Sixteen participants (1.7%) left one or more CHAOS questions unanswered, and outcome variables had 12 or fewer missing values (<2%). We imputed missing data by using multiple imputation methods developed by Rubin21 and Schafer.22 We used Stata version 11 for all analyses (StataCorp, College Station, TX).23 Conducting the analyses with complete cases only did not change our findings.

The RISE study received approval from the Medical Institutional Review Board of the University of California at Los Angeles, and this study was granted exemption from review by the Institutional Review Board of the Harvard Pilgrim Health Care Institute.

Results

The 929 participants had a mean (SD) age of 16.4 (1.3) years, and 44% were male. Eighty-four percent of students were Latino, 10% African-American, and 3% of other race. On the CHAOS scale, 165 (17.7%) had a score of 0, 303 (32.6%) had a score of 1–2, 206 (22.2%) had a score of 3–4, and 255 (27.4%) had a score of 5–14. Cronbach’s alpha for the CHAOS scale in our sample is 0.78 compared to 0.79 in the original sample from which the measure was developed.9 Twenty-seven percent of students reported having at least one parent who smoked, and 92% reported having at least one parent who worked. On the CES-D questionnaire, 21% of students had a score consistent with minor depression (CES-D 16–26) and 7% had a score consistent with major depression. Nine percent reported smoking a cigarette in the past 30 days, 34% reported drinking alcohol in the past 30 days, 21% reported smoking marijuana in the past 30 days, and 11% reported using any substance at school in the past 30 days. Twenty-seven percent reported being sexually active in the past 90 days, 13% reported not using any contraception at their prior sexual encounter, and 7% reported using substances with sexual activity in the past 3 months. Twenty-four percent reported having been in a fight in the previous 12 months, 3% endorsed gang activity in the past 12 months, and 6% reported having carried a weapon in the past 30 days (Table 2).

Table 2.

Characteristics of 929 High School Students in Los Angeles, Age 13–19 Years, in the RISE study.

Overall
N=929		 CHAOS Score
0
N=165		 1–2
N=303		 3–4
N=206		 5–14
N=255
Number (%)
Sex
 Male 413 (44) 74 (45) 143 (47) 96 (47) 100 (39)
 Female 516 (56) 91 (55) 160 (53) 110 (53) 155 (61)
Race
 Latino 780 (84) 127 (77) 244 (81) 182 (88) 228 (89)
 African-American 117 (13) 31 (19) 48 (16) 19 (9) 19 (7)
 Other 32 (3) 7 (4) 11 (3) 5 (3) 8 (4)
CES-D 20 Score
 <16 665 (72) 141 (85) 258 (85) 134 (65) 132 (52)
 16–26 191 (21) 21 (13) 35 (12) 54 (26) 81 (31)
 ≥ 27 73 (8) 3 (2) 10 (3) 18 (9) 42 (16)
Parenting Style
 Normal 584 (63) 113 (69) 206 (68) 134 (65) 131 (51)
 Neglectful 146 (16) 5 (3) 23 (8) 31 (15) 87 (34)
 Indulgent 55 (6) 12 (7) 22 (7) 12 (6) 9 (4)
 Authoritative 88 (9) 30 (18) 39 (13) 14 (7) 5 (2)
 Authoritarian 54 (6) 3 (2) 13 (4) 15 (7) 23 (9)
 ≥ 1 parent works full-time 852 (92) 152 (92) 287 (95) 184 (89) 229 (89)
Family history
 Parent smokes 247 (27) 41 (25) 74 (24) 62 (30) 70 (28)
 Alcohol 121 (13) 4 (3) 27 (9) 31 (15) 59 (24)
 Drugs 49 (5) 11 (4) 11 (4) 8 (4) 27 (11)
Smoking in past 30 days 87 (9) 5 (3) 21 (7) 23 (11) 38 (15)
Alcohol use in the past 30 days 308 (34) 32 (19) 93 (31) 68 (33) 115 (46)
Marijuana in the past 30 days 190 (21) 22 (13) 56 (19) 47 (23) 65 (26)
Sex in the past 90 days 247 (27) 41 (25) 67 (22) 54 (26) 85 (34)
No contraception at last encounter 109 (13) 18 (12) 28 (10) 25 (13) 38 (17)
Substance use with sex in the past 3 months 66 (7) 6 (4) 14 (5) 13 (6) 33 (13)
Substance use at school in past 30 days 100 (11) 4 (2) 27 (9) 24 (12) 45 (18)
Fight in the past 12 months 224 (24) 25 (15) 66 (22) 58 (28) 75 (29)
Gang involvement in past 12 months 32 (3) 2 (1) 8 (3) 8 (4) 14 (6)
Carrying weapon in past 30 days 59 (6) 5 (3) 16 (3) 12 (6) 26 (10)
Open in a new tab

RISE= Reducing Health Inequalities through Social and Educational Change

CHAOS=Chaos, Order and Hubbub scale

CES-D 20=Center for Epidemiologic Studies Depression 20-Question Scale

As shown in Table 3, in unadjusted analysis, greater household chaos was associated with higher risk of engaging in substance use, sexual, and violent risk behaviors. The relationship between chaos and risky health behaviors was strongest for substance use at school, with an unadjusted odds ratio of 8.2 (95% CI: 2.6–26) comparing those with the most versus least household chaos. The association was less strong for sexual risk behaviors—OR 3.7 (95% CI: 1.5–9.0) for substance use with sex in the past 3 months and OR 1.5 (95% CI: 1.0–2.3) for sexual activity in the past 90 days, for the highest vs the lowest CHAOS group—and positive though not statistically significant for no contraception use at last encounter. Furthermore, the relationship of chaos with risky behaviors was generally dose-dependent, with odds ratio for engaging in the behavior increasing monotonically with greater levels of chaos. For example, regarding cigarette smoking in the past 30 days, compared to those with CHAOS score 0, the unadjusted odds ratio of smoking was 2.2 for those with a CHAOS score 1–2, 3.6 for those with a score 3–4, and 4.9 for those with the highest CHAOS score.

Table 3.

Cross-Sectional Association of CHAOS With Health Risk Behaviors Among 929 Adolescents in Los Angeles, Age 13–19 Years.

Substance Use Risk Behaviors Sexual Risk Behaviors Violence-Related Risk Behaviors
Smoking in the past 30 days Alcohol in the past 30 days Marijuana in the past 30 days Any substance use at school in the past 30 days Sex in the past 90 days No contraception at last sexual encounter Substance use with sex in the past 30 days Fight in the past 12 months Carrying a weapon in the past 30 days Gang involvement in the past 12 months
CHAOS Score Odds ratio (95% CI)
Model 0 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
1–2 2.2 (0.9,5.5) 1.9 (1.2,2.9) 1.5 (0.9,2.5) 3.3(1.0,11) 0.8(0.5,1.3) 0.9(0.4,1.6) 1.2(0.5,3.2) 1.5(0.9,2.4) 1.8(0.6,5.1) 2.2(0.5,9.9)
3–4 3.6 (1.4,9.1) 2.0 (1.2,3.3) 1.9 (1.1,3.3) 5.0(1.5,16) 1.1(0.7,1.7) 1.3(0.7,2.6) 1.7(0.5,4.6) 2.1(1.3,3.5) 2.0(0.7,6.0) 3.1(0.7,14)
5–14 4.9 (2.0,11.9) 3.4 (2.1,5.4) 2.1 (1.3,3.6) 8.2(2.6,26) 1.5(1.0,2.3) 1.7(0.9,3.1) 3.7(1.5,9.0) 2.2(1.4,3.6) 3.7(1.4,9.9) 4.5(1.1,19)
Model 1 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
1–2 2.4 (0.9,6.6) 1.8 (1.1,2.9) 1.5 (0.8,2.5) 3.5(1.0,12) 0.8(0.5,1.3) 0.9(0.4,1.7) 1.2(0.5,3.2) 1.5(0.9,2.6) 1.9(0.7,5.3) 2.4(0.5,12)
3–4 4.2 (1.5,11.4) 2.1 (1.2,3.4) 2.2 (1.2,3.8) 5.7(1.6,20) 1.2(0.7,2.0) 1.5(0.7,2.9) 1.8(0.7,4.9) 2.2(1.3,3.8) 2.0(0.7,6.0) 3.8(0.8,19)
5–14 5.8 (2.2,15.2) 3.5 (2.1,5.6) 2.5 (1.4,4.2) 9.7(2.9,32) 1.6(1.0,2.6) 1.8(0.9,3.5) 3.9(1.6,9.8) 2.5(1.5,4.2) 3.9(1.5,11) 5.5(1.2,25)
Model 2 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
1–2 2.2 (0.8,6.0) 1.8 (1.1,2.9) 1.3 (0.8,2.4) 3.3(1.0,11) 0.8(0.5,1.3) 0.8(0.4,1.6) 1.1(0.4,2.9) 1.5(0.9,2.5) 1.7(0.6,5.1) 2.4(0.5,12)
3–4 3.1 (1.1,8.7) 1.9 (1.1,3.1) 1.7 (1.0,3.1) 4.9(1.4,17) 1.0(0.6,1.8) 1.3(0.7,2.8) 1.3(0.5,3.7) 2.0(1.2,3.5) 1.6(0.5,5.0) 3.1(0.6,16)
5–14 3.2 (1.2,8.8) 2.9 (1.7,4.8) 1.5 (0.8,2.8) 7.3 (2.1,25) 1.3(0.8,2.1) 1.6(0.8,3.3) 2.4(0.9,6.2) 1.9(1.1,3.4) 2.3(0.8,6.8) 4.1(0.8,20)
Fully adjusted model 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
1–2 2.2 (0.8,6.0) 1.8 (1.1,2.8) 1.3 (0.7,2.3) 3.1(0.9,11) 0.8(0.5,1.2) 0.8(0.4,1.6) 1.0(0.4,2.8) 1.5(0.9,2.6) 1.7(0.6,5.0) 2.2(0.5,11)
3–4 3.1 (1.1,8.7) 1.8 (1.1,3.0) 1.6 (0.9,2.9) 4.5(1.3,16) 1.0(0.6,1.7) 1.3(0.6,2.7) 1.3(0.5,3.6) 2.0(1.1,3.3) 1.6(0.5,5.0) 2.7(0.5,14)
5–14 3.1 (1.1,8.7) 2.6 (1.6,4.4) 1.4 (0.8,2.5) 6.1(1.8,21) 1.2(0.7,2.0) 1.5(0.7,3.0) 2.1(0.8,5.5) 1.9(1.1,3.3) 2.1(0.7,6.2) 3.4(0.7,17)
Open in a new tab

Model 0 is unadjusted for any covariates.

Model 1 is adjusted for the individual-level covariates grade in school, school type (charter vs. traditional), sex, and race/ethnicity.

Model 2 is adjusted for the personal characteristics (grade in school, school type, sex, race/ethnicity) and the parental-level covariates of parenting style and student-report of whether at least one parent works.

The fully adjusted model is adjusted for personal characteristics (grade in school, school type, sex, race/ethnicity), parenting style and parental employment, and family substance abuse history (parent smoking, family history of drug or alcohol abuse).

Adjusting for personal and family demographics, parenting style and family history of substance use attenuated the relationship between chaos and risky health behaviors, but the relationship remained statistically significant for smoking and alcohol use in the past 30 days, any substance use at school, and having been in a fight in the past 12 months. In the final adjusted model, greater household chaos was most strongly associated with any substance use at school (OR 6.1, 95% CI: 1.8–21), followed by cigarette use (OR 3.1, 95% CI: 1.1–8.7), alcohol use (OR 2.6, 95% CI: 1.6–4.4), and fighting in the past 12 months (OR 1.9, 1.1–3.3). Additionally adjusting for potential mediators depression and school engagement further attenuated odds ratios (Table 4).

Table 4.

Cross-Sectional Association of CHAOS With Health Risk Behaviors Among 929 Adolescents in Los Angeles, Age 13–19 Years, Adjusting for Potential Mediators

Substance Use Risk Behaviors Sexual Risk Behaviors Violence-Related Risk Behaviors
Smoking in the past 30 days Alcohol in the past 30 days Marijuana in the past 30 days Any substance use at school in the past 30 days Sex in the past 90 days No contraception at last sexual encounter Substance use with sex in the past 30 days Fight in the past 12 months Carrying a weapon in the past 30 days Gang involvement in the past 12 months
CHAOS Score Odds ratio (95% CI)
Fully adjusted 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
5–14 3.1 (1.1,8.7) 2.6 (1.6,4.4) 1.4 (0.8,2.5) 6.1(1.8,21) 1.2(0.7,2.0) 1.5(0.7,3.0) 2.1(0.8,5.5) 1.9(1.1,3.3) 2.1(0.7,6.2) 3.4(0.7,17)
+depression 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
5–14 2.7 (1.0, 7.5) 2.4 (1.4, 4.1) 1.1 (0.6,2.1) 5.4 (1.5,19) 1.0(0.6,1.7) 1.2(0.6,2.6) 1.8(0.7,4.1) 1.6(0.9,2.9) 2.0(0.7,6) 3.0(0.6,15)
+school engagement 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
5–14 2.6 (0.9, 7.2) 2.2 (1.3, 3.8) 1.0 (0.5, 1.9) 4.7 (1.3, 17) 1.1(0.7,1.8) 1.3(0.6,2.6) 1.8(0.7,4.9) 1.6(0.9,2.8) 1.9(0.6,5.6) 3.2(0.6,16)
+depression+school engagement 0 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
5–14 2.4 (0.8, 6.7) 2.2 (1.3, 3.7) 0.9 (0.5, 1.7) 4.5(1.3, 16) 1.0(0.6,1.7) 1.1(0.5,2.3) 1.6(0.6,4.4) 1.4(0.8,2.6) 1.9(0.6,5.6) 2.9(0.6,15)
Open in a new tab

The fully adjusted model is adjusted for personal characteristics (grade in school, school type, sex, race/ethnicity), parenting style and parental employment, and family substance abuse history (parent smoking, family history of drug or alcohol abuse).

Discussion

In this cross-sectional study of adolescents in Los Angeles from low socioeconomic backgrounds, we found greater household chaos to be associated with significantly greater odds of substance use and sexual and violent risk behaviors. This is the first study to provide insight into the relationship between this important and impactful psychosocial measure—household chaos—and risky health behaviors in adolescents.

Rates of health risk behaviors were comparable between our population and national data. According to Youth Risk Behavior Surveillance System (YRBSS) data from 2013, 16% of teens have smoked cigarettes in the past 30 days, 35% have used alcohol in the past 30 days, and 23% have smoked marijuana in the past 30 days.6 Comparably, in our population the percentages were 9%, 34% and 21%, respectively. About 27% of our students reported being sexually active, compared to 34% in the YRBSS nationally, with 6% of our students reporting carrying a weapon in the past 30 days, compared to 18% nationally.6 While our study population (89% Latino, in a large urban center) is otherwise different from the overall adolescent population in the US, comparable substance abuse rates suggest generalizability.

There was a strong relationship between CHAOS score and cigarette smoking and alcohol use among adolescents. There was also a strong relationship between CHAOS score and any substance use in school, though the confidence interval was wide (due to low numbers of students engaging in those behaviors across CHAOS score groupings). Interestingly, in our study, marijuana was the one substance use behavior not associated with higher household chaos (drug use for drugs other than marijuana was, however). It is possible that the relationship between marijuana and household chaos is fundamentally different than the relationship for other health risk behaviors, with marijuana use being common and accepted enough (21%) among adolescents that marijuana use is not associated with household chaos. Recent legislation legalizing recreational use of marijuana in Washington and Colorado has raised concerns about these laws’ impact on adolescent substance abuse behaviors,24 and further study is warranted.

We did not find an association between CHAOS score and sexual activity or alcohol with last sexual encounter. It is possible that the nature of the relationship between CHAOS and sexual behaviors is different than for other risky behaviors. There is less literature linking depression and sexual activity in adolescents, and some evidence suggests that adolescent sexual risk behaviors may precede depression.25 Furthermore, rates of sexual activity were lower in our population than in the general adolescent population, so there might be an unmeasured protective factor confounding the relationship between CHAOS and sexual behaviors in our population.

We did, on the other hand, find a relationship between CHAOS and behaviors related to violence. As CHAOS quartile increased, risk of having been in a fight, participated in a gang, or carried a gun increased, with the relationship for the latter two behaviors being particularly strong. A prior study linked household CHAOS scores to subsequent disruptive behaviors in 9–12 year old children, as measured by the Strengths and Difficulties Questionnaire,5 making the link between CHAOS score and violent behaviors in older adolescents seem plausible.

While the cross-sectional nature of the data makes determining the direction of the association difficult, we suspect that household chaos likely precedes and influences initiation of these health risk behaviors in adolescents. As noted above, in prior studies, higher CHAOS score preceded disruptive behaviors in older adolescents,5 and depression preceded sexual risk and substance use behaviors,25 indicating that CHAOS may similarly precede all three types of behaviors measured in our study. If risky behaviors were themselves causing household chaos, one might hypothesize that a riskier behavior such as marijuana use would be associated with more chaos than smoking, but the reverse is true.

On the other hand, risky health behaviors may cause increased household chaos—directly, or indirectly through decreased financial resources from using tobacco, alcohol, or drugs. Some of the downstream effects of household chaos are known and include significant negative psycho-social consequences in children.2–4 Given the adverse effects of household chaos, if the above risky health behaviors lead to greater chaos in a family, it would be reasonable to screen for high CHAOS score in adolescents with known substance use to try to mitigate downstream effects.

A third possible explanation for the link between chaos and substance use is that both household chaos and these behaviors are related to an underlying factor (either protective or risk-increasing) unmeasured in our study. While we adjusted for parental employment, parenting style, and parent history of substance use, other unmeasured aspects of parents, families and neighborhoods are likely to exist that could affect both household chaos in the home and a child’s likelihood to participate in risky behaviors. While investigators also measured self-efficacy26 and hopelessness,27 they were not initially part of our causal model, and in unadjusted bivariate analysis, were not related to either CHAOS score or outcome variables.

Prior research suggests that age, sex, race,6 socioeconomic status,18 parenting style,19 and family history of substance use20 are risk factors for substance use in adolescents, and we included each of these as confounders in our models. Age, sex and race are known to affect rates of sexual activity,28 so we included these factors in the final models, as well as socioeconomic status, which seemed reasonable to include. The association between CHAOS score and smoking, alcohol use, any substance use at school and fighting persisted after adjustment, indicating that there is an independent relationship between household chaos and these risky behaviors.

Depression has also been found to predispose toward high-risk sexual behavior29 and substance use disorder30 in adolescents. Previous studies indicate that higher CHAOS increases risk for psychosocial distress and learned helplessness14 which may make students more susceptible to depression, which in turn increases the likelihood of engaging in health risk behaviors. We thus thought of depression as a possible mediator between chaos and health risk behaviors, and included it in an additional model. Of note, as described by Van der Weele, formal mediation analysis is problematic for common dichotomous outcomes,31 such as those in our study, so we opted to introduce our potential mediators to the fully adjusted final model to look for attenuation of the relationship, which may suggest mediation. As expected, including these behaviors attenuated the relationship between chaos and risky behaviors, but the exact role of these behaviors—confounder or mediator—cannot be determined from this cross-sectional study (Table 4).

Additionally, in our study, students with greater household chaos were less likely to be engaged in school. Students who are less engaged in school may be more likely to have the time and inclination to begin smoking cigarettes, drinking alcohol, using drugs, having sex, or participating in violent activities. In support of this hypothesized mechanism, we found that adjusting for depression and school engagement attenuated the relationship between household chaos and risky behaviors (Table 4). The relationship between CHAOS and alcohol use was the least attenuated by the mediators, indicating that household chaos might affect alcohol use via a unique path.

If future studies indicate a causal relationship between household chaos and health risk behaviors, interventions that decrease social and physical disorder in the lives of adolescents may help to decrease the risk of engaging in health risk behaviors. While decreasing household chaos may seem like a challenging task, charter schools, which roughly half the participants in our study attended, have achieved success with academic achievement in part due to applications of strict routines and classroom management systems. Indeed, schools may be an important place where interventions to decrease health risk behaviors might be developed effectively, through understanding the roles of household chaos, depression, school engagement, and additional psychosocial factors. In addition, greater household chaos, at least in HIV-positive adults, is associated with less utilization of health care services and more missed doctor’s visits,32 so prevention and treatment interventions for risky health behaviors in adolescents should consider household chaos.

Also, the risky health behaviors we examined are important health disparities issues, with disparities in different directions based on the type of behaviors and group of interest. Black and Latino students are more likely to smoke marijuana than white students, and Latino students more likely to drink alcohol.6 White youth are more likely to smoke tobacco.6 For those black and Latino youth who do engage in substance abuse, morbidity and mortality from substance use is worse compared to whites.33,34 In the national YRBSS survey, black students were more likely to report being sexually active.6 White students were more likely to report having carried a weapon in the previous 30 days, while black and Latino students were more likely to report getting into fights.6 Because our study population is mostly Latino, we cannot compare the impact of household chaos on health risk behaviors across racial/ethnic groups. However, our study does increase our understanding of risky health behaviors in a particularly vulnerable population.

As mentioned above, limitations of our study include limitations on causal inference inherent to cross-sectional studies, and limitations to generalizability posed by the racial/ethnic make-up and location of our study population. But given comparable rates of substance use6 and a previous study confirming the reliability and validity of the CHAOS measure in Latino families,15 we believe that our findings are generalizable to the broader adolescent populations. More than half of potential participants could not be contacted, were ineligible, or refused to participate, so selection bias is possible. Our measures of health risk behaviors were also self-reported and, due to recall or reporting bias, may not accurately reflect risky behaviors among all participants, despite the use of A-CASI.

Because they are prevalent but preventable causes of future morbidity and mortality, understanding potential risk factors, such as household chaos, and psychosocial impacts of adolescent substance use, sexual activity and violence is important, and successful interventions to decrease participation early in adolescence will require that understanding.

Acknowledgments

Funding: Dr. Chatterjee received funding support from an Institutional National Research Service Award, (T32HP12706), the Ryoichi Sasakawa Fellowship Fund, and the Department of Population Medicine at Harvard Medical School/Harvard Pilgrim Health Care Institute. The funding source had no role in study design, data collection or analysis, or in the writing of the manuscript or decision to submit for publication.

Abbreviations

CHOAS

Chaos, Hubbub, and Order Scale

RISE

Reducing Health Inequalities Through Social and Educational Change

A-ACASI

audio-enhanced computer-assisted self-interview

CES-D

Center for Epidemiologic Studies Depression Scale

Footnotes

Conflict of Interest: The authors have no financial conflicts of interest to disclose.

Statement of Authorship: Avik Chatterjee wrote the first draft of the manuscript, and no one received an honorarium, grant, or other form of payment to anyone to produce the manuscript.

References

  • 1.Matheny APJ, Ludwig JL, Phillips KAY. Bringing Order Out of Chaos: Psychometric Characteristics of the Confusion, Hubbub, and Order Scale. 1995:429–444. [Google Scholar]
  • 2.Rutter M. A children’s behaviour questionnaire for completion by teachers: preliminary findings. [Accessed September 27, 2013];J Child Psychol Psychiatry. 1967 :8. doi: 10.1111/j.1469-7610.1967.tb02175.x. Available at: http://onlinelibrary.wiley.com/doi/10.1111/j.1469-7610.1967.tb02175.x/full. [DOI] [PubMed]
  • 3.Bullinger M, Hygge S, Evans G, Meis M, Mackensen Sv. The Psychological Cost of Aircraft Noise for Children. Zentralblatt für Hyg und Umweltmedizin. 1999;202(2–4):127–138. [PubMed] [Google Scholar]
  • 4.Humphrey LL. Children’s and teachers’ perspectives on children’s self-control: the development of two rating scales. J Consult Clin Psychol. 1982;50:624–633. doi: 10.1037//0022-006x.50.5.624. [DOI] [PubMed] [Google Scholar]
  • 5.Jaffee SR, Hanscombe KB, Haworth CMA, Davis OSP, Plomin R. Chaotic Homes and Children’s Disruptive Behavior: A Longitudinal Cross-Lagged Twin Study. Psychol Sci. 2012;23:643–650. doi: 10.1177/0956797611431693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kann L, Kinchen S, Shanklin S, et al. Youth Risk Behavior Surveillance — United States, 2013. Morb Mortal Wkly Rep. 2014;63(4) Available at: http://www.cdc.gov/mmwr/pdf/ss/ss6304.pdf. [PubMed] [Google Scholar]
  • 7.Wong MD, Shapiro MF, Boscardin WJ, Ettner SL. Contribution of major diseases to disparities in mortality. N Engl J Med. 2002;347:1585–1592. doi: 10.1056/NEJMsa012979. [DOI] [PubMed] [Google Scholar]
  • 8.Adolescent Mortality. Washington DC: 2009. Available at: http://mchb.hrsa.gov/chusa08/hstat/hsa/pages/225am.html. [Google Scholar]
  • 9.Wong MD, Coller K, Dudovitz R, et al. Successful schools and risky behaviors among low-income adolescents. Pediatrics. 2014;134(2):389–396. doi: 10.1542/peds.2013-3573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Garrison CZ, Addy CL, Jackson KL, McKeown RE, Waller JL. The CES-D as a screen for depression and other psychiatric disorders in adolescents. J Am Acad Child Adolesc Psychiatry. 1991;30:636–641. doi: 10.1097/00004583-199107000-00017. S0890-8567(09)64332-X [pii]\r. [DOI] [PubMed] [Google Scholar]
  • 11.Yazzie-Mintz E. Voices of Students on Engagement: A Report on the 2006 High School Survey of Student Engagement. 2007 Available at: http://www.eric.ed.gov/ERICWebPortal/recordDetail?accno=ED495758.
  • 12.Lamborn S, Mounts N, Steinberg L, Dornbusch S. Patterns of competence and adjustment among adolescents from authoritative, authoritarian, indulgent, and neglectful families. Child Dev. 1991;62(5):1049–1065. doi: 10.1111/j.1467-8624.1991.tb01588.x. [DOI] [PubMed] [Google Scholar]
  • 13.Wilson RS. The Louisville Twin Study: developmental synchronies in behavior. Child Dev. 1983;54:298–316. doi: 10.2307/1129693. [DOI] [PubMed] [Google Scholar]
  • 14.Evans GW, Gonnella C, Marcynyszyn La, Gentile L, Salpekar N. The role of chaos in poverty and children’s socioemotional adjustment. Psychol Sci. 2005;16(7):560–5. doi: 10.1111/j.0956-7976.2005.01575.x. [DOI] [PubMed] [Google Scholar]
  • 15.Haack LM, Gerdes AC, Schneider BW, Hurtado GD. Advancing our knowledge of ADHD in Latino children: Psychometric and cultural properties of Spanish-versions of parental/family functioning measures. J Abnorm Child Psychol. 2011;39(1):33–43. doi: 10.1007/s10802-010-9441-y. [DOI] [PubMed] [Google Scholar]
  • 16.Kurth AE, Martin DP, Golden MR, et al. A comparison between audio computer-assisted self-interviews and clinician interviews for obtaining the sexual history. Sex Transm Dis. 2004;31:719–726. doi: 10.1097/01.olq.0000145855.36181.13. [DOI] [PubMed] [Google Scholar]
  • 17.Brener ND, Collins JL, Kann L, Warren CW, Williams BI. Reliability of the Youth Risk Behavior Survey Questionnaire. Am J Epidemiol. 1995;141:575–580. doi: 10.1093/oxfordjournals.aje.a117473. Available at: http://aje.oxfordjournals.org/content/141/6/575.short. [DOI] [PubMed] [Google Scholar]
  • 18.Hanson MD, Chen E. Socioeconomic status and substance use behaviors in adolescents: the role of family resources versus family social status. J Health Psychol. 2007;12:32–35. doi: 10.1177/1359105306069073. [DOI] [PubMed] [Google Scholar]
  • 19.Story S. The Effects of Parenting Style on Adolescent Substance Use. Perspect (University New Hampshire) 2011:68–74. Available at: http://ezproxy.library.capella.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=sih&AN=80029476&site=ehost-live&scope=site.
  • 20.Biederman J, Faraone SV, Monuteaux MC, Feighner JA. Patterns of alcohol and drug use in adolescents can be predicted by parental substance use disorders. Pediatrics. 2000;106:792–797. doi: 10.1542/peds.106.4.792. [DOI] [PubMed] [Google Scholar]
  • 21.Rubin DB. Statistical Matching Using File Concatenation with Adjusted Weights and Multiple Imputations. J Bus Econ Stat. 1986;4:87–94. doi: 10.2307/1391390. [DOI] [Google Scholar]
  • 22.Schafer J. Analysis of Incomplete Multivariate Data. Boca Raton, FL: Chapman and Hall/CRC; 1997. [Google Scholar]
  • 23.StataCorp. Stata. 2009. [Google Scholar]
  • 24.Volkow N, Baler R, Compton W, Weiss S. Adverse Health Effects of Marijuana Use. N Engl J Med. 2014;(370):2219–2227. doi: 10.1056/NEJMra1402309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Hallfors DD, Waller MW, Bauer D, Ford CA, Halpern CT. Which comes first in adolescence--sex and drugs or depression? Am J Prev Med. 2005;29:163–170. doi: 10.1016/j.amepre.2005.06.002. [DOI] [PubMed] [Google Scholar]
  • 26.Schwarzer R, Jerusalem M. General Self-Efficacy Scale. 1995 [Google Scholar]
  • 27.Bolland JM. Hopelessness and risk behaviour among adolescents living in high-poverty inner-city neighbourhoods. J Adolesc. 2003;26:145–158. doi: 10.1016/S0140-1971(02)00136-7. [DOI] [PubMed] [Google Scholar]
  • 28.Abma J, Martinez G, Copen C. Teenagers in the United States: Sexual Activity, Contraceptive Use,and Childbearing, National Survey of Family Growth 2006–2008. 2010 Available at: http://www.cdc.gov/nchs/data/series/sr_23/sr23_030.pdf. [PubMed]
  • 29.Lehrer JA, Shrier LA, Gortmaker S, Buka S. Depressive symptoms as a longitudinal predictor of sexual risk behaviors among US middle and high school students. Pediatrics. 2006;118:189–200. doi: 10.1542/peds.2005-1320. [DOI] [PubMed] [Google Scholar]
  • 30.Rao U, Ryan ND, Dahl RE, et al. Factors associated with the development of substance use disorder in depressed adolescents. J Am Acad Child Adolesc Psychiatry. 1999;38:1109–1117. doi: 10.1097/00004583-199909000-00014. [DOI] [PubMed] [Google Scholar]
  • 31.VanderWeele TJ, Vansteelandt S. Odds ratios for mediation analysis for a dichotomous outcome. Am J Epidemiol. 2010;172:1339–1348. doi: 10.1093/aje/kwq332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Wong MD, Sarkisian CA, Davis C, Kinsler J, Cunningham WE. The Association Between Life Chaos, Health Care Use, and Health Status Among HIV-Infected Persons. 2007:1286–1291. doi: 10.1007/s11606-007-0265-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Shield KD, Gmel G, Kehoe-Chan T, Dawson DA, Grant BF, Rehm J. Mortality and Potential Years of Life Lost Attributable to Alcohol Consumption by Race and Sex in the United States in 2005. PLoS One. 2013:8. doi: 10.1371/journal.pone.0051923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Galea S, Ahern J, Tardiff K, et al. Racial/ethnic disparities in overdose mortality trends in New York City, 1990–1998. J Urban Health. 2003;80:201–211. doi: 10.1093/jurban/jtg023. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES