Self-Injurious Thoughts and Behaviors Among Youth in an ...

3 downloads 0 Views 250KB Size Report
Self-Injurious Thoughts and Behaviors Among Youth in an Underserved Area of the Southern United States: Exploring the Moderating Roles of Gender, ...
J Youth Adolescence (2010) 39:270–280 DOI 10.1007/s10964-009-9462-4

EMPIRICAL RESEARCH

Self-Injurious Thoughts and Behaviors Among Youth in an Underserved Area of the Southern United States: Exploring the Moderating Roles of Gender, Racial/Ethnic Background, and School-Level Robert D. Latzman • Kim L. Gratz • John Young • Laurie J. Heiden • John D. Damon • Terry L. Hight

Received: 30 July 2009 / Accepted: 1 October 2009 / Published online: 16 October 2009 Ó Springer Science+Business Media, LLC 2009

Abstract Despite the clinical relevance of self-injurious thoughts and behaviors (SITB) among youth, little is known about the subset of youth most at-risk for SITB. This study examined the moderating roles of gender, racial/ethnic background, and school-level (and their interactions) on rates of SITB within a large (N = 2638, 52.2% female), ethnically-diverse sample of middle- and high-school youth in a relatively poor and underserved area of the Southern United States. Extending extant research in this area, findings indicated a significant interaction between gender and race for self-injurious behaviors, with African–American boys reporting higher rates than all other groups. Findings also indicated significant interactions between school-level and both gender and race for self-injurious thoughts. Whereas comparable levels of self-injurious thoughts were reported across middle- and high-schools for girls and African–American youth, the frequency of these thoughts was higher among both boys and White students in highschool (vs. middle-school). Results highlight the need for further research on SITB among diverse youth in underserved areas.

R. D. Latzman (&)  K. L. Gratz Department of Psychiatry and Human Behavior, University of Mississippi Medical Center, 2500 North State St., Jackson, MS 39216, USA e-mail: [email protected] J. Young Department of Psychology, University of Mississippi, University, MS, USA L. J. Heiden  J. D. Damon  T. L. Hight Mississippi Children’s Home Services—CARES School, Jackson, MS, USA

123

Keywords Self-injury  Self-harm  Adolescents  Gender  Race  School-level

Introduction Prevalence estimates indicate that 13–26% of youth aged 12 to 20 years have engaged in self-injurious behaviors (Evans et al. 2005), defined as the deliberate, direct destruction of body tissue regardless of suicidal intent. As such, this overarching category includes both non-suicidal and suicidal self-injurious behaviors (SIB). SIB in youth have been shown to be associated with a range of clinical difficulties and negative outcomes, including internalizing and externalizing psychopathology (Sourander et al. 2006), interpersonal problems and distress (Hilt et al. 2008), drug and alcohol use (Hilt et al. 2008), and both previous suicide attempts and current suicidal ideation (Lloyd-Richardson et al. 2007). The clinical and public health significance of this phenomenon among youth is clear. Surprisingly, though, little is known about the subset of youth most at-risk for SIB, with little research examining the specific demographic factors (and combination of factors) associated with SIB among children and adolescents. Specifically, given the relative infancy of this area of research, studies that have examined differences in SIB across gender, race, and age (or grade) have typically focused on only one of these demographic characteristics, with little empirical attention directed toward the interrelationships among demographic factors and SIB. A more nuanced examination of the demographic characteristics associated with heightened risk for SIB among children and adolescents is critical in designing more targeted prevention and intervention efforts. In particular, for areas in which there are limited resources available, especially in rural

J Youth Adolescence (2010) 39:270–280

and/or underserved areas (e.g., Angold et al. 2002), identification of youth at highest risk will help determine the most appropriate allocation of these scarce resources. The current study aims to respond to this need by examining the moderating role of gender, racial/ethnic background, and school-level, as well as their interactions, on rates of SIB within a large, representative sample of youth living in Mississippi, a relatively poor and underserved area of the United States. Below, we present a brief overview of the extant literature regarding relations between each of these demographic factors and SIB among youth. Gender and Self-Injurious Behavior Research is equivocal with regard to the relationship between gender and SIB. Meta-analytic results (examining published studies through December, 2000) of studies examining suicide-related phenomenon and SIB among adolescents in the general population (Evans et al. 2005), as well as more recent empirical school-based investigations of self-reported SIB among youth in school settings (Hawton et al. 2002; De Leo and Heller 2004; Madge et al. 2008; Matsumoto et al. 2008; O’Connor et al. 2009), suggest that girls may be more likely than boys to engage in SIB. For example, Madge et al. (2008) found that among their large community sample (N = 30,000) of predominately 15- and 16-year-old adolescents across six European countries and Australia, girls were almost 3.5 times more likely than boys to engage in SIB. Consistent findings were reported in samples of Japanese youth (N = 1,726) in junior and senior high school (Matsumoto et al. 2008) and Canadian high-school students (Laye-Gindhu and SchonertReichl 2005; Ross and Heath 2002), which found higher rates of SIB in general and non-suicidal SIB, respectively, among girls versus boys. Taken together, results of examinations of gender differences in SIB outside of the United States suggest that girls are more likely than boys to exhibit SIB. Studies utilizing samples of adolescents from the United States, however, have found somewhat conflicting results. For example, in studies of non-suicidal SIB in particular, neither Hilt et al. (2008) nor Lloyd-Richardson et al. (2007) found gender differences in rates of these behaviors in samples of sixth to eighth graders (N = 508) and ninth to twelfth graders (N = 633), respectively. Likewise, Muehlenkamp and Gutierrez (2004) found no significant differences in non-suicidal SIB across gender within their sample (N = 390) of Midwestern high-school students. Thus, despite some evidence that girls may be at higher risk for SIB than boys, findings suggest that gender may be a less relevant demographic factor with regard to risk for SIB among youth in the United States (vs. other countries).

271

Racial/Ethnic Background and Self-Injurious Behavior The extant literature concerning relations between racial/ ethnic background and SIB is extremely scarce, and the few studies that have examined the influence of racial or ethnic background on rates of SIB have yielded mixed results. Some studies suggest that non-suicidal SIB may be more common among White, versus non-White, adolescents (Muelenkamp and Gutierrez 2004; Nock et al. 2006; LloydRichardson et al. 2007), while other research indicates similar rates of non-suicidal SIB across racial/ethnic groups (Hilt et al. 2008). For example, Hilt et al. (2008) found similar rates of non-suicidal SIB in White and non-White students living in the Northeastern United States. It is important to note, however, that 87% of the sample was White, making the detection of differences as a function of racial background difficult. Conversely, studies including a larger percentage of minority adolescents have found racial differences in rates of non-suicidal SIB, with one such study of 633 ninth to twelfth grade high school students (50% African American) in the Midwestern and Southern United States finding that White students were significantly more likely to engage in moderate/severe non-suicidal SIB than were African–American students (Lloyd-Richardson et al. 2007). No study to date has examined differences in SIB including both suicidal and non-suicidal behaviors. Age/Grade and Self-Injurious Behavior The extant literature is also equivocal with regard to the association between age (or grade in school) and SIB. Some studies suggest that older adolescents are more likely to engage in these behaviors than younger adolescents (Hilt et al. 2008; Sourander et al. 2006). For example, in a longitudinal study of Finnish children aged 12 to 15 (N = 839), Sourander et al. (2006) found an increase in SIB from age 12 to 15 years. Similarly, in their study of 10 to 14 year old girls (N = 94), Hilt et al. (2008) found that those who engaged in non-suicidal SIB were significantly older than those who did not. Conversely, other studies have failed to find significant differences in rates of non-suicidal SIB across age or grade in samples of middle-school (Hilt et al. 2008) or high-school (Lloyd-Richardson et al. 2007) students. Despite some evidence that rates of SIB may not differ across age, however, it is important to note that none of the studies described above have included youth from both middle schools and high schools in the same sample; thus, important school level-related variations in rates of SIB may have been obscured. Likewise, the use of samples with a limited age-range precludes examination of age- or school-related differences in the associations between gender or race and SIB. Consistent with this possibility, in a

123

272

study of a large (N = 3,628) database of consecutively admitted hospital patients with SIB ranging in age from late-childhood to adulthood, Hawton and Harriss (2008) found that the associations between gender and SIB change across adolescence. Specifically, the gender ratio changed from 8:1 (girls to boys) in the 10 to 14 year old group to 3.1:1 in the 15 to 19 year old group. Thus, although findings indicated higher rates of medically-attended SIB among girls than boys, they also provide suggestive evidence that these gender differences decrease as youth get older. Altogether, the extant literature suggests that although there may be gender differences in SIB, with girls significantly more likely than boys to engage in these behaviors (Evans et al. 2005; Madge et al. 2008), these gender differences may be less apparent in samples of youth in the United States (Hilt et al. 2008; Lloyd-Richardson et al. 2007). Further, there is some evidence to suggest that gender differences in rates of SIB may be moderated by age group (Hawton and Harriss 2008). As such, findings suggest the importance of examining rates of SIB across both gender and age group. Likewise, although research on racial/ethnic differences in rates of SIB among youth have produced inconsistent results, no studies have examined the extent to which race interacts with gender and age or school-level to influence SIB.

Current Study The primary purpose of the current study was to examine the independent and interactive roles of gender, racial/ ethnic background, and school-level in the prediction of SIB among a large, ethnically-diverse, mixed-gender sample of both middle and high school students in a relatively poor and underserved area of the Southern United States. Given evidence that self-injurious thoughts may be associated with some of the same outcomes as SIB (LayeGindhu and Schonert-Reichl 2005), and may fall under the broader category of self-injurious thoughts and behaviors (SITB), we also explored the moderating role of gender, race, and school-level on self-injurious thoughts. Understanding the role of these demographic variables in SITB among youth is critical to the identification of youth at greatest risk for SITB, as well as to the development of specific, targeted prevention and early intervention efforts. Research on this topic is particularly needed among youth in relatively poor and underserved areas of the country, which have generally been understudied by researchers in this area. One such area is the state of Mississippi, which ranks last in the country on a combined index of longevity, knowledge, and income (the American Human Development Index; see Burd-Sharps et al. 2009), and has the lowest educational attainment (see the 2009 Mississippi Human

123

J Youth Adolescence (2010) 39:270–280

Development Report; Burd-Sharps et al. 2009) and highest rate of children living below the poverty line (37%; see US Census Bureau 2007) in the United States. Further, almost half the population of Mississippi (40.2%) lives in an area classified as a Mental Health Professional Shortage Area, placing it in the bottom fifth of the country with regard to access to mental health resources (US Department of Health and Human Services 2009). Thus, research on SITB among youth in this state has the potential to speak to both the rates and associated demographics of this behavior among an atrisk and underserved sample. Indeed, despite findings that youth in this state are at relatively high risk for numerous maladaptive behaviors (including lifetime cigarette use, risky sexual behavior, and poor health and dietary behaviors; Centers for Disease Control and Prevention 2009), no studies to date have examined SITB among this population. Given past findings that older youth may be more likely to engage in SIB than younger youth (Hilt et al. 2008; Sourander et al. 2006), as well as findings that the average age of onset of non-suicidal SIB is 13–14 years (Favazza and Conterio 1989; Klonsky and Muehlenkamp 2007), we predicted a main effect of school-level on rates of SIB, with high school students reporting higher rates of SIB than middle school students. Likewise, given some evidence of gender differences in the rates of SIB (Evans et al. 2005), we hypothesized that girls would report higher rates of SIB than boys, but that this association would be moderated by school-level, with gender differences in rates of SIB decreasing in high school (vs. middle school). Given the paucity of research on the impact of race on SIB, as well as the equivocal findings from the studies that have examined this association, no hypotheses were made with regard to the main or interactive effects of racial background on SIB. Finally, given that the one previous study that examined non-suicidal self-injurious thoughts found significant gender differences (Laye-Gindhu and SchonertReichl 2005), we hypothesized a main effect of gender on rates of self-injurious thoughts, with girls reporting more frequent self-injurious thoughts than boys. No predictions were made regarding the impact of race or school-level on self-injurious thoughts.

Method Participants Participants included 2,638 sixth through twelfth grade students (52.2% female) from eight public middle-schools and high-schools across four districts in Mississippi. Consistent with the racial/ethnic demographics of the areas from which participants were drawn (see United States Census Bureau 2007), approximately two-thirds of the sample self-identified as White and one-third as African–American. Based on

J Youth Adolescence (2010) 39:270–280

school and area records, the sample was diverse with regard to economic status, with the median household income of the areas from which participants were drawn ranging from $32,687 to $59,278 (M = $38,017.73 ± $8,235.16) and approximately half the sample living in areas falling under the median state income of $36,424 (United States Census Bureau 2007). Further, the percentage of students within each school living below or within the poverty marker (as indexed by the percentage of students receiving free or reduced-fee lunches) ranged from approximately 20–80% (M = 66.93; see Mississippi Office of Healthy Schools 2009).1 Notably, however (and consistent with evidence suggesting that Mississippi as a whole is a relatively poor area; see Burd-Sharps et al. 2009; United States Census Bureau 2007), despite the economic diversity of our sample, 88.4% of the participants lived in areas falling under the median US income of $50,740 (United States Census Bureau 2007). See Table 1 for complete demographic information on the sample.

273 Table 1 Sample demographics (N = 2638) Gender Female

52.2% (N = 1376)

Male

47.8% (N = 1262)

Race White

66.2% (N = 1746)

African–American

33.8% (N = 892)

Grade 6th

20.0% (N = 484)

7th

20.6% (N = 500)

8th

19.5% (N = 473)

9th

23.1% (N = 561)

10th

7.2% (N = 175)

11th

6.3% (N = 154)

12th

3.3% (N = 79)

School-level Middle-school

63.3% (N = 1669)

High school

36.7% (N = 969)

Median household income

Measures For the purposes of this study, and consistent with past research (Klonskyv et al. 2003), three items from the Inventory of Depression and Anxiety Symptoms (IDAS; Watson et al. 2007) were used to assess SITB. Cutting or Burning Oneself Purposely Item number 41 on the IDAS, ‘‘I cut or burned myself on purpose,’’ was used as one measure of SIB. Participants responded to this item on a Likert-type scale ranging from 1 (Not at all) to 5 (Extremely). Consistent with past studies on SIB in youth (Hawton et al. 2003; Madge et al. 2008; O’Connor et al. 2009), a dichotomous SIB item was created by assigning a score of ‘‘0’’ to participants who responded ‘‘Not at all’’ to this question, and a score of ‘‘1’’ to participants who reported having engaged in these behaviors. Hurting Oneself Purposely Item number 9 on the IDAS, ‘‘I hurt myself purposely,’’ was used as another measure of SIB. Participants 1

Given the possibility that children within the same community may be more similar than children from different communities (thereby yielding non-independent observations), multilevel modeling (Raudenbush and Bryk 2002) was used to determine whether analyses needed to account for the nesting of these data. For each of the three dependent variables, a model with no predictors was run to assess whether any variance in the dependent variable was accounted for by the grouping of children within districts. Intraclass correlation coefficients suggested that less than 3% of the variance in each dependent variable was accounted for by school district. Thus, observations can be considered independent and analyses proceeded in a non-nested format.

\$35,000

48.3% (N = 1273)

$35,000–$50,000

39.6% (N = 1044)

[$50,000

12.2% (N = 321)

responded to this item on a Likert-type scale ranging from 1 (Not at all) to 5 (Extremely), and a dichotomous SIB item was created was created by assigning a score of ‘‘0’’ to participants who responded ‘‘Not at all’’ to this question, and a score of ‘‘1’’ to participants who reported having hurt themselves purposely at all. Self-Injurious Thoughts Item number 15 on the IDAS, ‘‘I thought about hurting myself’’ was used to assess self-injurious thoughts. Participants responded to this item on a Likert-type scale ranging from 1 (Not at all) to 5 (Extremely). Scores on this item were examined continuously (except when rates of self-injurious thoughts are reported). Demographics As part of a larger school-based survey on psychopathology in middle- and high-school students, information on participants’ gender, grade, and racial/ethnic background was collected. Procedure The Mississippi Children’s Home Services Institutional Review Board approved all study procedures. Prior to data collection, principals from each participating school mailed

123

274

home an information letter concerning the study procedures. Parents were also given the option to have their children opt out of participating. Estimates provided by school administrators indicate that fewer than 5% of students in each school did not participate. Anonymous surveys were administered by school teachers during regularly scheduled classes (across 168 classrooms, approximately 15–20 students per classroom).

J Youth Adolescence (2010) 39:270–280 Table 2 Correlations among dependent variables and demographic characteristics (N = 2638) 1

2

3

Analysis Plan Analyses were conducted with purposely hurting oneself, cutting or burning oneself, and thinking about hurting oneself as the primary dependent variables, and gender (female vs. male), racial/ethnic background (African– American vs. White), and school-level (middle- vs. highschool) serving as the independent variables. The moderating roles of gender, racial/ethnic background, and school-level on the two SIB variables (i.e., purposely cutting or burning oneself, and purposely hurting oneself) were examined using two separate hierarchical logistic regression analyses. Independent variables were entered in three steps: Step 1 included all three main effects and the covariate of median household income; Step 2 included all two-way interactions among the three primary variables of interest; and Step 3 included the three-way interaction of gender, race, and school-level. The relationship between gender, race, and school-level and selfinjurious thoughts was examined using a 2 (female vs. male) 9 2 (African–American vs. White) 9 2 (middle- vs. high-school) analysis of covariance (ANCOVA), controlling for median household income.2 Preliminary Analyses Interrelations among SITB items and key demographic variables are presented in Table 2. The three SITB items were significantly correlated with one another and their correlations with the demographic factors were all small. Overall, 14.1% of participants (N = 370) endorsed cutting or burning themselves and 14.8% of participants (N = 390) endorsed hurting themselves on purpose. When examined dichotomously, 20.9% (N = 550) of participants endorsed thinking about hurting themselves. Rates of cutting or burning oneself and hurting oneself did not differ by gender (v2s = 1.06, .14, respectively; ps [ .30). However,

5

6

7

8

1. Cut or Burned Self 2. Hurt Self

.54

3. Thoughts of Hurting Self

.49

.53

4. Gender (Female) -.02 -.01

Results

4

5. Race (White) 6. School-Level (High-School) 7. Median household income 8. Poverty

.07

-.06 -.05 -.04 -.03 .08

.06

.08

-.12 -.08 -.11

.04

.03

.04 -.20 .02

.32

.03

.03 -.06 -.01 -.58 -.68

Significant correlations are shown in boldface. Correlations of C |.05| are significant p \ .01, C |.07|, p \ .001. Cut or burned self and hurt self are measured dichotomously. Thoughts of hurting self is measured continuously. Poverty = percentage of students within each school living below or within the poverty marker as described in the text

girls were significantly more likely than boys to endorse thinking about hurting themselves (v2 = 10.71, p \ .01). Rates of cutting or burning oneself, hurting oneself, and thinking about hurting oneself differed as a function of race (v2s = 8.84, 6.52, 7.46, respectively; all ps \ .01), with African–American youth being significantly more likely to endorse these items than White youth. Likewise, significant differences were found in rates of all SITB items as a function of school-level, with high-school students being more likely to report cutting or burning themselves (v2 = 16.79, p \ .001), hurting themselves (v2 = 10.84, p \ .001), and thinking about hurting themselves (v2 = 18.17, p \ .001). Table 3 shows the percentages of students endorsing each SITB item as a function of gender, race, and school-level. Consistent with the dichotomous data on rates of selfinjurious thoughts (yes vs. no), one-way analyses of variance (ANOVAs) conducted on the continuous measure of the frequency of self-injurious thoughts revealed that thoughts of hurting oneself differed significantly by gender (F [1, 2634] = 11.34, p \ .01) and school-level (F [1, 2634] = 11.95, p \ .01), with both girls (M = 1.51) and high-school students (M = 1.52) reporting more frequent thoughts of hurting themselves than boys (M = 1.39) or middle-school students (M = 1.38). Frequency of self-injurious thoughts did not differ as a function of race, F (1, 2634) = .96, p [ .30. Logistic Regression Analyses Predicting SIB

2

Of note, all findings remained the same when accounting for the percentage of students within each school living below or within the poverty marker.

123

Two hierarchical regression analyses were conducted to examine whether the demographic variables of interest

J Youth Adolescence (2010) 39:270–280

275

Table 3 Rates of self-injurious thoughts and behaviors among all youth, and as a function of gender, race, and school-level (N = 2638) Hurt self (%)

Cut or burned self (%)

Thoughts of hurting self (%)

14.8

14.1

20.9

15.1 14.6

14.8 13.4

18.2 23.4

White

13.5

12.7

19.3

African–American

17.3

16.9

23.9

Middle-school

13.1

12.0

18.3

High-school

17.8

17.8

25.3

Total sample By gender Boys Girls By race

By age

Table 4 Logistic regression analysis examining rates of cutting or burning oneself on purpose as a function of gender, race, and schoollevel (N = 2638) Predictor

B

SE Wald

OR

95% CI

Step 1 SES $35 k–$50 k vs. \$35 k [$50 k vs. \$35 k Gender: female Race: white School: high-school

.93 .22 18.83** 2.54 1.67–3.88 .44 .26

2.79

1.55

.93–2.58

.12 .11

1.20

1.13

.91–1.41

.11

1.04

.81–1.34

8.11**

.66

.49–.88

.04 .13 -.42 .15

Gender 9 school-level

.76 .24 10.17** 2.13 1.34–3.39 .22 .25 .79 1.25 .76–2.05 -.21 .23

.81

.81

.51–1.28

.17 .48

.13

1.19

.47–3.01

Step 3 Gender 9 race 9 schoollevel * p \ .05; ** p \ .01

(and/or their interactions) reliably distinguish youth with SIB from those without SIB (when controlling for SES).3 Cutting or Burning Oneself on Purpose The full model for the item pertaining to cutting or burning oneself on purpose was significant (v2 = 65.28, p \ .001). As shown in Table 4, school-level reliably improved the prediction of SIB status (yes vs. no), as did the addition of the two-way interactions in the second step of the model (v2 = 13.12, p \ .01). However, only the interaction of gender and race emerged as a reliable predictor of SIB 3

status in the second step of the model (see Fig. 1), with rates of SIB differing as a function of race for boys (11.6% for Whites vs. 21.6% for African–Americans) but not girls (13.7% for Whites vs. 13.0% for African–Americans). The three-way interaction of gender, race, and school-level did not reliably improve the model (v2 = .13, p [ .70). Hurting Oneself Purposely

Step 2 Gender 9 race School 9 race

Fig. 1 Interaction between gender and racial/ethnic background predicting rates of cutting or burning oneself on purpose. * v2 = 11.62, p \ .01

All results remained the same when SIB items were examined continuously.

The same pattern of findings emerged for the SIB item focused on hurting oneself purposely (see Table 5). Specifically, the full model for this SIB item was significant (v2 = 36.78, p \ .001), and school-level reliably improved the prediction of SIB status (yes vs. no) in the first step of the model. Further, as with findings for the SIB item described above, the interaction of gender and race in the second step of the model emerged as a reliable predictor of SIB status (see Fig. 2). Yet, contrary to the findings for the item focused specifically on cutting and burning oneself, the addition of all two-way interaction variables in the second step of the model failed to reliably improve the model, with this step as a whole only approaching significance (v2 = 5.37, p \ .15). However, when only the gender by race interaction was entered alone in the second step of model, the step reliably improved the prediction of SIB status (v2 = 5.33, p \ .05). Specifically, whereas rates of SIB did not differ as a function of race for girls (14.4% for Whites vs. 14.8% for African–Americans), African– American boys were significantly more likely to report hurting themselves purposely than White boys (20.2 vs. 12.6%, respectively). Once again, the addition of the three-

123

276

J Youth Adolescence (2010) 39:270–280

Table 5 Logistic regression analysis examining rates of hurting oneself purposely as a function of gender, racial/ethnic background, and school-level (N = 2638) Predictor

B

SE Wald

OR

Table 6 Three-way analysis of covariance (controlling for SES) predicting thoughts of hurting oneself (N = 2638) df

Mean square

F

95% CI SES

2

16.35

17.72**

Gender

1

6.59

7.14**

Race

1

1.07

1.16

.80 .21 14.62** 2.22 1.48–3.34 .68 .25 7.12** 1.97 1.20–3.34

School-level

1

9.49

10.28**

Gender 9 race

1

1.14

1.24

Gender: female

.04 .11

.14

1.04

.84–1.30

School 9 race

1

4.30

4.66*

Race: white

.11 .13

.73

1.11

.87–1.42

.61

.46–.82

Gender 9 school-level Gender 9 race 9 school-level

1 1

1.97 .38

2.13 .41

2624

.92

Step 1 SES $35 k–$50 k vs. \$35 k [$50 k vs. \$35 k

School-level: high-school

-.49 .15 10.69**

Step 2

Error

Gender 9 race

.52 .23

4.86*

1.68 1.06–2.65

School 9 race

-.03 .25

.01

.97

.60–1.59

Gender 9 school-level

-.04 .23

.03

.96

.61–1.51

-.29 .47

.38

.75

.30–1.88

* p \ .05; ** p \ .01

Step 3 Gender 9 race 9 schoollevel * p \ .05; ** p \ .01

Fig. 3 Interaction between school-level and racial/ethnic background predicting frequency of thoughts of hurting oneself

Fig. 2 Interaction between gender and racial/ethnic background predicting rates of purposely hurting oneself. * v2 = 4.43, p \ .05

way interaction variable did not reliably improve the model (v2 = .38, p [ .50). Analysis of Variance Predicting Self-injurious Thoughts To examine the demographic factors associated with thoughts of hurting oneself, a three-way ANCOVA (controlling for SES) was conducted. As shown in Table 6,

123

findings revealed significant main effects for gender and school-level, as well as a significant interaction between race and school-level, F (1, 2633) = 4.66, p \ .05. Specifically, as shown in Fig. 3, whereas school-level was not significantly associated with self-injurious thoughts among African–American students, it was associated with the frequency of these thoughts for White students, with White students in high-school reporting more frequent thoughts of harming themselves than White students in middle-school.

Discussion The current investigation is the first to examine the interactive relationship among gender, racial/ethnic background, and school-level in the prediction of SITB among youth. This study builds upon an existing literature that is

J Youth Adolescence (2010) 39:270–280

mixed with regard to findings of the influence of these demographic factors. Furthermore, rarely have the interrelationships among these demographic factors and SITB been examined as they were in the current study. Our more comprehensive examination of demographic factors associated with SITB is critical in identifying those youth at greatest risk for SITB, as well as designing more targeted prevention and early intervention efforts. In our large (N = 2638), ethnically-diverse, mixedgender sample of sixth through twelfth grade youth in Mississippi, approximately 15% reported SIB and 21% reported self-injurious thoughts. These rates of SIB are similar to those reported in studies of nonsuicidal SIB among high-school students in both Canada (13.9%; Ross and Heath 2002) and the Midwestern United States (15.9%; Muelenkamp and Gutierrez 2004). Interestingly, though, the rates of self-injurious thoughts reported by youth in our sample were higher than those previously reported in the literature (10%; Laye-Gindhu and Schonert-Reichl 2005). This discrepancy may be due, in part, to our examination of a diverse underserved population, in contrast to the largely White Canadian sample examined by Laye-Gindhu and Schonert-Reichl (2005). Consistent with past research on youth in the United States (Hilt et al. 2008; Lloyd-Richardson et al. 2007), findings indicated no gender differences in rates of SIB in our Southern United States sample. However, findings did indicate that girls were more likely than boys to report thinking about hurting themselves, consistent with past findings of gender differences in selfinjurious thoughts in particular (Laye-Gindhu and Schonert-Reichl 2005). Results of our study indicated that African–American (versus White) youth and high-school (versus middleschool) students were more likely to endorse all SITB items. Contrary to previous findings suggesting higher (Muelenkamp and Gutierrez 2004; Nock et al. 2006; Lloyd-Richardson et al. 2007) or equal (Hilt et al. 2008) rates of SIB among White versus non-White youth, the African–American youth in our sample had significantly higher rates of SIB than their White peers. It warrants mention, however, that this study is the first to examine racial/ethnic differences in SITB among youth in a relatively poor and underserved area of the Southern United States, making direct comparisons with previous studies difficult. Indeed, it is possible that African–American youth in Mississippi (an area of the country with historically negative race relations; see Davis et al. 2009) may be a more at-risk group for a variety of clinical difficulties than minority youth in other parts of the country. With regard to school level-related differences in rates of SITB, we found higher rates of all SITB items among high-school (versus middle-school) students. Although research on age- or grade-related differences in SIB is

277

equivocal (with some studies finding higher rates among older youth (Hilt et al. 2008; Sourander et al. 2006) and other studies finding no evidence for age- (Lloyd-Richardson et al. 2007) or grade-related (Hilt et al. 2008) differences in rates of this behavior), none of these studies has included youth from both middle and high-schools in the same sample, potentially obscuring the school level-related differences that emerged in our findings. Indeed, in one of the only studies to examine SIB in youth of a wide range of ages (from 12 to 17 years), Matsumoto et al. (2008) reported significant differences in rates of SIB between only the 12- and 13-year old students. However, differences in rates of SIB between middle-school and highschool students were not examined, making the current study the first to examine differences in rates of SITB between youth in middle- and high-schools. Future research is needed to examine youth from late childhood through late adolescence, and from elementary school through high school. In addition to examining the role of demographic characteristics in isolation, we expanded upon the extant literature by examining interactions among these demographic factors. Results of the current study revealed a significant interaction between gender and racial/ethnic background in predicting SIB status. Specifically, although rates of SIB did not differ as a function of race among girls, there were racial/ethnic differences in rates of SIB among boys, with African–American boys reporting higher rates of SIB than White boys. Given past findings suggesting that White youth (vs. ethnic minority youth) and girls (vs. boys) may be at greater risk for SIB (see Lloyd-Richardson et al. 2007 and Madge et al. 2008, respectively), findings that African–American boys reported higher rates of SIB than all other groups were unexpected. Nonetheless, it is important to note that past studies have not examined the interaction of race and gender in the prediction of SIB and, as a result, findings of racial differences specific to boys may have been obscured by the focus on the main effects of these characteristics only. Then again, these findings may also reflect the at-risk status of African–American males in the state of Mississippi specifically (see Centers for Disease Control and Prevention 2009). With regard to the impact of demographic characteristics on self-injurious thoughts, findings indicated a significant interaction between race and school-level. Specifically, whereas White high-school students reported higher levels of self-injurious thoughts than White middleschool students, there were no significant school-related differences in self-injurious thoughts for African–American youth. Findings suggest that African–American youth may be more likely to have these types of thoughts at an earlier age than White youth. However, by the time youth reach high-school, these racial/ethnic differences seem to

123

278

disappear. Although rarely examined in the extant SITB literature (for an exception, see Laye-Gindhu and SchonertReichl 2005), self-injurious thoughts are of great clinical relevance. Indeed, the related area of research on suicideand life-threatening behaviors indicates a substantial overlap between thoughts and behaviors, as well as evidence that thoughts of this nature predict future behaviors (Haavisto et al. 2003; Pearce and Martin 1994). Strengths and Limitations One major strength of the current study was our large, ethnically-diverse, mixed-gender, community sample of middle- and high-school students. Additionally, our use of anonymous questionnaires administered within regular classes may have resulted in students being more comfortable openly endorsing items, as previous research has found higher endorsement rates using anonymous questionnaire methodologies as opposed to interviews (Evans et al. 2005). Furthermore, our examination of demographic factors (particularly race) associated with SITB among youth in a relatively poor and underserved area is a strength. A limitation of the current study was our use of singleitem questions to assess SITB. Although this approach is used commonly in the literature on SITB in both youth (Sourander et al. 2006; O’Connor et al. 2009) and adult populations (Klonsky et al. 2003), it may underestimate the occurrence of SIB, compared to approaches that use behaviorally-specific items. Nonetheless, it is important to note that the findings obtained using both SIB items (i.e., the broader item focused on hurting oneself purposely and the more specific item focused on cutting or burning oneself purposely) were comparable to one another and suggest that gender and race may interact in predicting SIB among youth. Another limitation concerns the absence of information on the suicidal intent of the SIB. As such, it is not possible to distinguish between SIB behaviors with suicidal intent (i.e., suicide attempts) and those without suicidal intent (i.e., non-suicidal self-injury or deliberate self-harm; Chapman et al. 2006; Gratz 2001). Given research suggesting that the factors associated with non-suicidal and suicidal SIB may differ (Muelenkamp and Gutierrez 2004), future research is needed to examine each of these types of SIB specifically, as well as the role of demographic factors, both independently and in interaction with one another, in predicting these specific forms of SIB. Given that data in this study were cross-sectional in nature, findings with regard to school level-related differences in SIB should be interpreted with caution. In particular, it is not clear whether findings of higher rates of SITB among high-school versus middle-school students reflect a developmentally-related change or simply a cohort effect. Further, it is not clear how these important

123

J Youth Adolescence (2010) 39:270–280

demographic factors interact with one another across childhood and adolescence. Future longitudinal research that includes large, ethnically diverse samples of youth from various geographic regions is needed to examine the role of these demographic factors over time, as well as the clinical correlates and putative underlying mechanisms of SITB among these various groups of youth. In addition to examining these potential etiological mechanisms longitudinally, further empirical attention should be given to populations that have generally not been represented in the literature (e.g., at-risk youth in relatively poor, underserved areas). In particular, results of the current study suggest that African–American boys (at least in the state of Mississippi) may represent a particularly highrisk group for SITB. This is particularly notable given that African–American boys are generally a neglected and understudied group in research on SITB, compared to other groups for whom SITB are considered more common and relevant (e.g., White girls). Further, it is unclear whether and to what extent the extant literature in this area, often based on predominantly White female samples, will generalize to more diverse populations. Research is needed to examine this question directly. If research finds that rates of SITB differ among groups and the etiological mechanisms and correlates also differ, new or tailored treatment approaches may be needed.

Conclusion Consistent with past research on SIB among youth, findings indicate that approximately 15% of youth in sixth through twelfth grade report engaging in SIB, and 21% endorse the presence of self-injurious thoughts. However, findings also indicate that the rates of (and risk for) SITB may vary as function of gender, racial/ethnic background, and schoollevel (as well as their interaction). In particular, findings suggest that African–American boys may be at heightened risk for SIB, reporting significantly higher rates of SIB than their White peers. Given that African–American boys have been a relatively understudied population with regard to SITB, the findings of this study highlight the importance of continuing to examine SIB among this subset of youth.

References Angold, A., Erkanli, A., Farmer, E. M. Z., Fairbank, J. A., Burns, B. J., Keeler, G., et al. (2002). Psychiatric disorder, impairment, and service use in rural African Amercian and White youth. Archives of General Psychiatry, 59, 893–901. Burd-Sharps, S., Lewis, K., & Martins, E. B. (2009). A portrait of Mississippi: Mississippi human development report 2009. New York: American Human Development Project.

J Youth Adolescence (2010) 39:270–280 Centers for Disease Control and Prevention. (2009). 2007 Youth risk behavior survey. Available at: www.cdc.gov/yrbss. Accessed on July 24. Chapman, A. L., Gratz, K. L., & Brown, M. Z. (2006). Solving the puzzle of deliberate self-harm: The experiential avoidance model. Behaviour Research and Therapy, 44, 371–394. Davis, A., Gardner, B. B., & Gardner, M. R. (2009). Deep South: A social anthropological study of caste and class. Columbia, South Carolina: University of South Carolina Press. De Leo, D., & Heller, T. S. (2004). Who are the kids who self-harm? An Australian self-report school survey. The Medical Journal of Australia, 181, 140–144. Evans, E., Hawton, K., Rodham, K., & Deeks, J. (2005). The prevalence of suicidal phenomena in adolescents: A systematic review of population-based studies. Suicide and Life-Threatening Behaviors, 35, 239–250. Favazza, A. R., & Conterio, K. (1989). Female habitual self mutilators. Acta Psychiatrica Scandinavica, 79, 283–289. Gratz, K. L. (2001). Measurement of deliberate self-harm: Preliminary data on the deliberate self-harm inventory. Journal of Psychopathology and Behavioral Assessment, 23, 253–263. Haavisto, A., Sourander, A., Ellila¨, H., Va¨lima¨ki, M., Santalahti, P., & Helenius, H. (2003). Suicidal ideation and suicide attempts among child and adolescent psychiatric inpatients in Finland. Journal of Affective Disorders, 76, 211–221. Hawton, K., Hall, S., Simkin, S., Bale, L., Bond, A., Codd, S., et al. (2003). Deliberate self-harm in adolescents: A study of characteristics and trends in Oxford, 1990-2000. Journal of Child Psychology and Psychiatry, 44, 1191–1198. Hawton, K., & Harriss, L. (2008). The changing gender ratio in occurrence of deliberate self-harm across the lifecycle. Crisis: The Journal of Crisis Intervention and Suicide Prevention, 29, 4–10. Hawton, K., Rodham, K., Evans, E., & Weatherall, R. (2002). Deliberate self harm in adolescents: Self report survey in schools in England. British Medical Journal, 325, 1207–1211. Hilt, L. M., Cha, C. B., & Nolen-Hoeksema, S. (2008a). Nonsuicidal self-injury in young adolescent girls: Moderators of the distressfunction relationship. Journal of Consulting and Clinical Psychology, 76, 63–71. Hilt, L. M., Nock, M. K., Lloyd-Richardson, E. E., & Prinstein, M. J. (2008b). Longitundinal study of nonsuicidal self-injury among young adolescents. Journal of Early Adolescence, 28, 455–469. Klonsky, E. D., & Muehlenkamp, J. J. (2007). Self-injury: A research review for the practitioner. Journal of Clinical Psychology, 63, 1045–1056. Klonsky, E. D., Oltmanns, T. F., & Turkheimer, E. (2003). Deliberate self-harm in a nonclinical population: Prevalence and psychological correlates. American Journal of Psychiatry, 160, 1501– 1508. Laye-Gindhu, A., & Schonert-Reichl, K. A. (2005). Nonsuicidal selfharm among community adolescents: Understanding the ‘‘whats’’ and ‘‘whys’’ of self-harm. Journal of Youth and Adolescence, 34, 447–457. Lloyd-Richardson, E. E., Perrine, N., Dierker, L., & Kelley, M. L. (2007). Characteristics and functions of non-suicidal self-injury in a community sample of adolescents. Psychological Medicine, 37, 1183–1192. Madge, N., Hewitt, A., Hawton, K., Jan de Wilde, E., Corcoran, P., et al. (2008). Deliberate self harm within an international community sample of young people: Comparative findings from the child & adolescent self-harm in Europe (CASE) study. Journal of Child Psychology and Psychiatry, 49, 667–677. Matsumoto, T., Imamura, F., Chiba, Y., Katsumata, Y., Kitani, M., & Takeshima, T. (2008). Prevalences of lifetime histories of selfcutting and suicidal ideation in Japanese adolescents:

279 Differences by age. Psychiatry and Clinical Neurosciences, 62, 362–364. Mississippi Office of Healthy Schools. (2009). Free and reduced lunches percentages report (2008–2009). Available at: www. healthyschoolsms.org/nutrition_services/reports.htm. Accessed on September 8, 2009. Muehlenkamp, J. J., & Gutierrez, P. M. (2004). An investigation of differences between self-injurious behavior and suicide attempts in a sample of adolescents. Suicide and Life-Threatening Behavior, 34, 12–23. Nock, M. K., Joiner, T. E., Gordon, K. H., Lloyd-Richardson, E., & Prinstein, M. J. (2006). Non-suicidal self-injury among adolescents: Diagnostic correlates and relations to suicide attempts. Psychiatry Research, 144, 65–72. O’Connor, K., Rasmussen, S., Miles, J., & Hawton, K. (2009). Selfharm in adolescents: Self-report survey in schools in Scotland. The British Journal of Psychiatry, 194, 68–72. Pearce, C. M., & Martin, G. (1994). Predicting suicide attempts among adolescents. Acta Psychiatrica Scandinavica, 90, 324– 328. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage Publications, Inc. Ross, S., & Heath, N. (2002). A study of the frequency of selfmutilation in a community sample of adolescents. Journal of Youth and Adolescence, 31, 67–77. Sourander, A., Aromaa, M., Pihlakoski, L., Haavisto, A., Rautava, P., Helenius, H., et al. (2006). Early predictors of deliberate selfharm among adolescents. A prospective follow-up study from age 3 to age 15. Journal of Affective Disorders, 93, 87–96. United States Census Bureau. (2007). American Factfinder. Available at: www.factfinder.census.gov. Accessed on September 8, 2009. United States Department of Health and Human Services. (2009). Health professional shortage areas (HPSAs). Available at www.hpsafind.hrsa.gov. Accessed on September 8, 2009. Watson, D., O’Hara, M. W., Simms, L. J., Kotov, R., Chmielewski, M., McDade-Montez, E., et al. (2007). Development and validation of the inventory of depression and anxiety symptoms (IDAS). Psychological Assessment, 19, 253–268.

Author Biographies Robert D. Latzman is a Postdoctoral Research Fellow in the Department of Psychiatry and Human Behavior at the University of Mississippi Medical Center. He received his Ph.D. in Clinical Psychology from the University of Iowa. His main research interests are in developmental psychopathology, (dis)inhibitory processes, temperament/personality, neuropsychology, and externalizing and related behaviors. Kim L. Gratz is an Assistant Professor and Director of Personality Disorders Research in the Department of Psychiatry and Human Behavior at the University of Mississippi Medical Center. She received her PhD in Clinical Psychology from the University of Massachusetts Boston. Her clinical and research interests focus on the role of emotion dysregulation and experiential avoidance in borderline personality disorder (BPD) and deliberate self-harm among both adolescents and adults, as well as the intergenerational transmission of borderline personality and related pathology. John Young is an Assistant Professor of Psychology at the University of Mississippi. He received his Ph.D. in Clinical Psychology from the University of Hawaii. His major research interests are the

123

280 dissemination of evidence-based services for children and adolescents and mental health care in schools. Laurie J. Heiden is the Division Director for Education Services at Mississippi Children’s Home Services in Jackson, Mississippi. She received her M.Ed. in Special Education and Education Leadership from the University of Southern Mississippi and Grand Canyon University. Her main research interests include relations between mental health and educational attainment in educational settings. John D. Damon is the Chief Operating Officer of Mississippi Children’s Home Services in Jackson, Mississippi. He received his

123

J Youth Adolescence (2010) 39:270–280 Ph.D. in Clinical Psychology from Jackson State University. He has research interests in residential treatment of children and adolescents and systems of care. Terry L. Hight is the Chief Administrative Officer of Mississippi Children’s Home Services in Jackson, Mississippi. He received his Ph.D. in Counseling Psychology from Virginia Commonwealth University. His research interests include residential treatment of children and adolescents and marital and family therapies.