Community Mental Health Journal

, Volume 48, Issue 1, pp 71–78

Mental Health Screening of African American Adolescents and Facilitated Access to Care

Authors

    • Department of Child and Adolescent PsychiatryNew York State Psychiatric Institute, Columbia University
    • TeenScreen National Center for Mental Health CheckupsColumbia University
  • Deborah A. Kanter
    • Department of PsychologyFaculty of Health, York University
  • Leslie McGuire
    • Department of Child and Adolescent PsychiatryNew York State Psychiatric Institute, Columbia University
    • TeenScreen National Center for Mental Health CheckupsColumbia University
  • Mark Olfson
    • Department of Child and Adolescent PsychiatryNew York State Psychiatric Institute, Columbia University
    • TeenScreen National Center for Mental Health CheckupsColumbia University
Original Paper

DOI: 10.1007/s10597-011-9413-x

Cite this article as:
Husky, M.M., Kanter, D.A., McGuire, L. et al. Community Ment Health J (2012) 48: 71. doi:10.1007/s10597-011-9413-x

Abstract

This study retrospectively reviews de-identified records from school-based mental health screening in a predominantly African American community. We compare participation rates, screening results, referrals to services and access to care of white and African American adolescents. Among those offered screening, 20.1% of white students (n = 297), and 28.8% of African American students (n = 499) were screened (χ2 = 32.47, df = 1, P < .001). African American students (45.1%) were significantly more likely than white students (33.0%), (AOR = 1.59; P = .003) to be identified as being at risk. In both racial groups, most youth accessed the school-based services (89.02%, 95% CI 82.25–95.79) and community services (86.57%, 95% CI 78.41–94.73) to which they were referred. The groups did not differ in the odds of accessing community-based services (AOR = .58; P = .49). African American students were, however, more likely than white students to access school-based services (AOR = 10.08; P = .022). The findings support the effectiveness of screening in school settings in predominantly African American communities.

Keywords

AdolescenceAfrican American youthMental health screeningRace/ethnicity

Introduction

One in every 4–5 adolescents meets criteria for a mental disorder with severe impairment in the US (Merikangas et al. 2010). However, many adolescents with treatable mental health conditions do not access specialized services. Estimates of the percentage of youth in need who do not receive appropriate mental health care are generally quite high and vary from 28 (Garland et al. 2005) to 64% (Angold et al. 2002). Poor access to services also occurs among adolescents at risk for suicide (Cheung et al. 2009).

Substantial racial disparities between white youth and racial/ethnic minority groups exist in access to specialized mental health services (Kessler et al. 2005). African American individuals in particular are consistently less likely than their white counterparts to receive inpatient or outpatient mental health treatment (Alexandre et al. 2010; Angold et al. 2002; Garland et al. 2005; Sturm et al. 2003; Wu et al. 2010), and when they do, the quality of the care they receive is likely to be lower than that provided to white individuals (Wang et al. 2000).

Actively identifying adolescents with treatable conditions through systematic screening and linking them to appropriate services is one approach that has been developed to palliate unmet need for mental health services in youth (US Preventive Services Task Force 2009). Screening efforts have been carried out in school settings where they have been shown to improve upon routine in-school mechanisms in identifying high-risk students. School-based screening has also been shown to be safe (Gould et al. 2005) and to lead to a majority of the high risk adolescents receiving specialized services within a year of the screen (Gould et al. 2009). Schools may also provide a convenient setting for mental health screening with less racial disparity in access to care in school-based services than in community based services (Wu et al. 2010).

Little is known, however, about the extent to which African American parents and students are agreeable to consent to systematic voluntary school-based mental health screening and about the performance of screening in this population. To our knowledge, only one study conducted in a sample of African American high school students living in an urban area has documented the acceptability and performance of school-based screening in African American youth (Brown and Goldstein Grumet 2009). This study reported very low screen acceptance rates (14%), high positive screen rates (45%) and high rates of connection with school-based mental health services (86%). This issue is in need of replication beyond urban settings and to racially/ethnically mixed samples that permit comparisons across racial/ethnic groups. Much remains to be learned concerning the viability of school-based screening as a strategy for reducing unmet mental health care needs in adolescents at risk and racial/ethnic disparities in access to care. The purpose of the present study is to report the experiences of a program offering routine voluntary school-based mental health screening to adolescents in a primarily African American rural southern community. The objective is to compare African American students to white students with respect to (1) screening consent rates, (2) screening results and identification of youth at risk, (3) mental health referrals, and (4) subsequent connection with mental health services.

Methods

Participants

The current study is a retrospective record review. Since 2004, and independently from any research initiative, the Office of Mental Health of the State of Louisiana has been offering systematic screening using the Columbia TeenScreen program in thirteen public schools in the two school districts in a small city in Louisiana. The city has a population of approximately 50,000 persons, with 37% white and 61% African American individuals (US Census Bureau 2009). The median household income is approximately $26,000 and one-third of the population lives below the poverty level. The local staff estimated that a total of 3,210 students in grades 6 through 8 were offered screening between September 2008 and April 2009. Due to the racial/ethnic distribution in this sample, we excluded the records of 56 students who endorsed a race/ethnicity other than white or African American. We reviewed the records of all African American or white students who were screened. The final study sample included 796 white and African American students. The authors report no conflicts of interest.

Procedure

The screening procedures used by the local staff follow the guidelines provided by the TeenScreen National Center for Mental Health Checkups at Columbia University (Schools and Community Office at the TSNCMHC at Columbia University 2010). The local screening staff comprises 4 staff members, three of whom are masters’ level professionals and who receive weekly supervision from a licensed clinical social worker. All have received additional training from the national center to implement the following procedures. First, a member of the screening staff provides students attending physical education classes a ten min oral presentation that describes the screening process. At the end of the presentation, students are handed a consent form that they are asked to give to their parents and return the next day. Students who return forms with active written parent consent and who provide written assent are then invited to complete the Columbia Health Screen (CHS). This self-report questionnaire is used to identify suicidality and risk for mental health problems. Based on the results of the self-report screen, students who are deemed to be in need of further evaluation are administered a second-stage clinical interview by a master’s level clinician. If, after the interview, a student is determined to be in need of a mental health referral, parents are contacted, and the results of the screen and the recommendation for further attention are shared with them. The family’s insurance status, transportation situation, and personal preference for a type of provider are discussed. Parents are then offered the option of having the screening staff directly make an appointment with a provider of their choice, and they are informed that they would be contacted in approximately 3 months to follow-up and determine whether the student is still receiving services. This screening program has relations with 8 community mental health providers, 2 of whom accept Medicaid and 5 of whom have a sliding fee schedule. In this county, 1 of the 2 school districts is a recipient of a federal Safe Schools/Healthy Students Grant. It provides funding to enhance school and community-based mental health services. Through this mechanism, students have the option to be referred at no cost to community-based services which they may access within the school or in their home.

Following approval of the Institutional Review Board of the New York State Psychiatric Institute, we have asked the local screening program to allow for an external and retrospective review of their de-identified screening records. The screening staff made de-identified photocopies of the screening records, as well as of the records documenting whether or not students accessed services for which they were recommended. Screening records were for students screened between September 2008 and April 2009 and were sent to the authors for analysis.

Measures

Acceptability

The acceptability of screening was determined by examining the proportion of those who were screened among all of those who were offered to be screened, as screening implied obtaining both parent consent and student assent.

Columbia Health Screen

The CHS is an abridged version of the Columbia Suicide Screen (CSS) (Shaffer et al. 2004). The CHS and CSS share 14 identical mental health questions. These questions concern suicidal ideation, prior suicide attempts, and emotional problems. The assessment of emotional problems includes five domains: depression, anxiety, irritability, social withdrawal, and substance use. Respondents are asked to indicate the severity of the problem on a five-point Likert scale ranging from 1 (No problem) to 5 (Very bad problem). Most questions refer to the previous 3 months. A student is considered “positive” by endorsing one of the following criteria: (1) presence of suicidal ideation in the last 3 months, (2) any history of a suicide attempt, (3) 3 or more of the 5 emotional problem questions reported as representing a “bad” or “very bad” problem, (4) indication of being “more upset” after completing the CHS, (5) refusal to answer question(s), or 6) reporting a need for help with an emotional problem. The CSS which uses the same symptom criteria has a sensitivity of .75, a specificity of .83, a positive predictive value of .16, and a negative predictive value of .99 for identifying confirmed cases (Shaffer et al. 2004). Unlike the CSS, the CHS also scores refusal to answer questions and requests for mental health care as constituting a positive screen.

Second-Stage Clinical Interview

This structured interview addresses depression, suicidal ideation and behavior, anxiety, substance use, and psychosocial stressors over the past 3 months, as well as current and past mental health treatment. All interviews are conducted by master’s level screening staff members who make the final decision of whether or not a student has difficulties that warrant further attention. For the purpose of this study, the types of service referrals received were classified into school-based services and community-based services and extracted from these interviews.

Access to Services

Services were categorized into two groups: school-based services (which include services such as meetings with school social workers, guidance counselors and social skills groups), and community-based services (which include any contact with a local mental health care provider outside of school). As part of routine procedures, members of screening staff contacted families by telephone approximately 3 months after the referral to assess whether the connection with referral to a school-based or community-based provider was completed.

Data Analysis

Among those offered screening, we first compare those who were screened to those who were not with regard to race/ethnicity, sex, grade, school district, and free lunch status using χ2 tests (Table 1). Only aggregate information (school-level) regarding race/ethnicity, sex and free lunch status was available for those who were not screened (Louisiana Department of Education 2008). We then compare white and African American students who completed the screen with regard to the background characteristics (Table 2). A series of backward stepwise logistic regressions were used to assess the strength of associations between student race/ethnicity and status on the screens. Race/ethnicity was forced into all models. Sex, age, school district and school free lunch status were then entered with alpha set at .05 to retain and .10 to be removed from the models (Table 3). Similar regressions were used to assess associations between student race/ethnicity and access to care (Table 4). Because only one of the districts benefited from the federal Safe Schools/Healthy Students grant, school district and student/race ethnicity were forced into the model of access to services. Analyses regarding referrals to mental health services and access to services (Table 4) excluded adolescents who were seeing a mental health provider at the time of the screen.
Table 1

Estimates of student participation in a voluntary mental health screening program, overall and stratified by student race/ethnicity, sex, grade, school district, and free lunch status

 

Offered participation

Rates of students who were screened per 100 offered

95% CI

χ2 (df)

P value

Overall

3,210

24.80

   

Race/ethnicity

 White

1,478

20.09

18.05–22.13

32.47 (1)

<.001

 African American

1,732

28.81

26.68–30.94

Sex

 Female

1,573

29.62

27.36–31.88

39.79 (1)

<.001

 Male

1,637

19.98

18.04–21.92

Grade

 6th

274

39.42

33.63–45.21

45.29 (2)

<.001

 7th

1,534

25.88

23.69–28.07

 8th

1,402

20.76

18.64–22.88

School district

 District 1*

1,119

33.3%

30.57–36.09

67.2 (1)

<.001

 District 2

2,091

20.2%

18.51–21.95

Free lunch status

 80% and up

1,268

26.2%

23.76–28.6

70.04 (2)

<.001

 40–79%

1,003

31.7%

28.82–34.58

 39% and under

939

15.5%

13.23–17.87

* This school district received a federal safe schools/healthy students grant that provides no cost in school and home-based mental health services to students

Table 2

Demographic characteristics of white and African American students who received voluntary mental health screening

 

White n = 297 (%)

African-American n = 499 (%)

χ2 (df) or t (df)

P value

Sex

 Female

176 (59.3)

290 (58.5)

.48 (1)

.827

 Male

121 (40.7)

206 (41.5)

Age

12.84 (.79)

13.04 (1.04)

−2.85 (794)

.004

Grade

 6th

30 (10.1)

79 (15.8)

5.56 (2)

.062

 7th

150 (50.5)

246 (49.3)

 8th

117 (39.4)

174 (34.9)

School district

 District 1*

42 (11.3)

331 (88.7)

203.66 (1)

<.001

 District 2

255 (60.3)

168 (39.7)

* This school district received a federal safe schools/healthy students grant that provides no cost in school and home-based mental health services to students

Table 3

Positive screen criteria of white and African American students who received voluntary mental health screening

Screen criteria

White n = 297 (%)

African-American n = 499 (%)

AORa

95% CI

P value

Any positive screen criteriab

98 (33.0)

224 (44.9)

1.59

1.17–2.18

.003

 Suicidality

43 (14.5)

62 (12.4)

.99

.59–1.68

.988

 ≥3 Emotional problemsc

18 (6.1)

33 (6.6)

.96

.52–1.77

.890

 Asked for help for an emotional problem

73 (24.6)

192 (38.5)

1.75

1.18–2.61

.005

 More upset after the screen

14 (4.7)

11 (2.2)

.65

.26–1.65

.365

 Left questions unanswered

4 (1.3)

25 (5.0)

2.49

.76–8.11

.130

aWhite is the reference. Backward stepwise logistic regressions included sex, age, school district, and school free lunch status

bReference is negative screen

cReported as “bad” or “very bad” or likert response scale

Table 4

Post-screening referrals and referral completion of white and African American students who received voluntary mental health screening

 

White n1 (%)

African-American n2 (%)

AORab

95% CI

P value

Received a clinical interview (n1 = 297, n2 = 499)b

67 (22.6)

144 (28.9)

1.34

.95–1.91

.098

Referred to any services (n1 = 59, n2 = 120)c

48 (81.4)

97 (80.8)

.99

.44–2.21

.980

Referred to school-based services (n1 = 48, n2 = 97)d

35 (72.9)

69 (71.1)

.82

.32–2.12

.686

Accessed school-based services (n1 = 21, n2 = 61)e

16 (76.2)

57 (93.4)

10.08

1.39–73.28

.022

Referred to community-based services (n1 = 48, n2 = 97)d

30 (62.5)

48 (49.5)

1.56

.59–4.12

.367

Accessed community-based services (n1 = 27, n2 = 40)f

24 (88.9)

34 (85.0)

.58

.12–2.76

.491

aWhite is the reference. Backward stepwise logistic regressions included sex, age, school district, and school free lunch status

bReference is those screened who did not receive an interview. Those currently receiving treatment were excluded from analyses

cReference is those interviewed but not referred. Those currently receiving treatment were excluded from analyses

dReference is those referred to any services. Those currently receiving treatment were excluded from analyses

eReference is those referred to school services. School district was forced into the model. Those currently receiving treatment were excluded from analyses

fReference is those referred to community services. School district was forced into the model. Those currently receiving treatment were excluded from analyses

Results

Acceptability of Screening

Of the 3,210 students offered screening, 796 (24.8%) were screened. The proportion of African American students, females, and 6th grade students who consented to screening was significantly greater than that of white students, males, and seventh or eighth grade students, respectively. In addition, screen acceptance rates varied by school district and by free lunch status. Schools with a lower proportion of free lunch recipients had a lower screening participation rate.

Among those who were screened, there were no significant differences in the sex and grade distribution of African American and white students (Table 2). Screened African American students were slightly yet significantly older than white students.

Identification of Youth at Risk

Among the 796 students screened, 322 (40.5%) screened positive. Controlling for the potentially confounding effects of sex, age, school district and free lunch status, the proportion of African American students who screened positive was significantly greater than that of white students (Table 3). African American students had 1.75 times the odds of their white counterparts of asking for professional help for an emotional problem. The endorsement of lifetime suicide attempts, suicidal ideation in the past 3 months, leaving questions unanswered or the endorsement of at least 3 bad emotional problems did not significantly vary by race/ethnicity.

In total, 211 (26.5%) of those who were screened received a second-stage clinical interview. While 322 students scored positive on the screening instrument, 111 were not given a clinical interview, either because they were deemed to be false positives (n = 98) during the first minute of the interview or they were positive and mis-scored by the screening staff as negative screens (n = 13).

Connection with School-Based and Community-Based Services

Among the 211 interviewed students, 32 (15.2%) were currently seeing a mental health professional. There was not a significant difference in the proportion of African American (16.7%) and white (11.9%) students currently seeing a mental health professional (χ2 = .79, df = 1, P = .37). Of those who were interviewed and were not currently in treatment, 145 students (18.2% of the total sample) were deemed to be in need of a referral for further evaluation (Table 4).

Referrals were made to school-based services for 104 students (71.7%) and to community-based services for 78 students (53.8%). No significant racial/ethnic group differences in referral rates were found. A great majority of those referred to mental health services accessed these services within 3 months of the screen, including 89.0% of students who were referred to school-based services and 86.5% of students who were referred to community services. Bivariate analyses indicated that there were no significant differences between the school districts in the proportion of youth who accessed either school-based (χ2 = .001, df = 1, P = .97) or community-based (χ2 = .083, df = 1, P = .77) services (data not shown). While there were no between-group differences in rates of access to community-based services, the proportion of African American students who accessed school-based services was significantly higher than the proportion of white students (Table 4).

Discussion

The present study supports the general acceptability and feasibility of mental health screening among low income African American students. In contrast with prior studies which have reported that African American youth are either less or no more likely than white youth to participate, we found that African American students were more likely than their white counterparts to participate in voluntary mental health screening. The participation rate of African American students is nearly twice that obtained in a study offering mental health screening to urban African American students (Brown and Goldstein Grumet 2009), although the older age of this sample may have diminished screen acceptance. Furthermore, because the current study involves a program that had been in place for several years prior to the record review, the screen acceptance rate may reflect a relatively high level of comfort with the program in the school culture.

Roughly four in ten African American youth screened positive for a mental health concern. This is comparable with findings from a sample of urban African American youth with the same instrument (Brown and Goldstein Grumet 2009) and underscores the extent of mental health stress among African American young people.

The screen positive rate was significantly higher in African American than in white adolescents. This racial/ethnic difference was largely due to proportionately more African American than white youth asking for professional help for an emotional or behavioral problem. A propensity to ask for help might be related to relatively optimistic attitudes towards mental help-seeking among African American youth (Mojtabai 2007; Shim et al. 2009) in spite of relatively low rates of mental health service use (Schnittker et al. 2005). Additional research is needed in this area to offer a better understanding of the paucity of psychological or psychosocial resources available to low income African American youth.

Consistent with results from the Youth Risk Behavior Survey (National Center for Chronic Disease Prevention and Health Promotion 2009), there were no marked racial differences in the endorsement of suicidal ideation or prior suicide attempts. These results are in line with national data indicating that historically higher rates of suicidal ideation and behaviors among white than African American youth have diminished over the last several years (Ialongo et al. 2002; Joe et al. 2006).

In contrast with studies that have consistently reported marked racial differences in adolescents’ access to specialized mental health care (Angold et al. 2002; Elster et al. 2003; Garland et al. 2005), our results found little evidence of a racial/ethnic difference in access to community-based care. While there remain important disparities in socio-economic status between white and African American families, these disparities alone have been shown not to be the sole contributor to differences in access to care. Several studies reported that racial/ethnic disparities persisted after controlling for socio-economic indicators including parental income, insurance status and education (Elster et al. 2003) for review). Another factor that has been linked to differences in access to care is the lack of trust of African Americans for the health care system (Halbert et al. 2006) stemming from historical experiences of segregation, racism, and discrimination. It is possible that the consistent involvement of the school-based screening staff may have helped reduce ethnic differences in access to care by enhancing trust.

In accord with a recent school-based study (Brown and Goldstein Grumet 2009), a great majority of students in the current study who were referred to school-based mental health services successfully accessed these services. This pattern supports the notion that school-based mental health services facilitate access to care by overcoming transportation, cost, insurance, or other barriers that tend to hinder access to community health care providers and may contribute to racial/ethnic disparities in service use (Angold et al. 2002; Burns et al. 1995; Kataoka et al. 2007). In the current study, African American students were more likely than white students to access school-based mental health services. Developing school-based mental health services appears to represent a promising strategy for facilitating access to specialized care for youth in need and overcoming ethnic/racial disparities in access to community-based services. Future research might investigate whether stigmatization processes or peer deprecation of care received outside of school by African American students contributes to disparities in access to specialized community-based mental health services and contributes to a preference for school-based mental health services.

In interpreting the findings, several limitations should be acknowledged. First, limited person-level information was available for those who refused screen participation. Perhaps most importantly, we do not have access to information that permits comparison of the economic status of the two student groups. While we gathered aggregate data, person-level data would permit more precise estimates of the determinants of participation. Second, we did not have information on all of the students referred for school-based or community-based services opening potential selection bias. Third, though it is an important issue, our study was not designed to address the preferences of African Americans in our sample for school-based over community services. Finally, information on successful connection with services was based on parent report rather than administrative records and may therefore, be vulnerable to social desirability bias.

In conclusion, a two-stage school-based mental health screening process in a predominantly African American rural low income community achieved a high rate of successful connection of distressed youth to mental health services. While minorities are typically less likely to participate in screening efforts and access mental health services, African American youth enjoyed greater post-screening access to school-based care than their white counterparts and the two groups had similar access to community mental health care. These results are encouraging and particularly compelling in the context of the numerous known obstacles to spontaneous youth mental health seeking behavior (Ciguralov et al. 2008). At the same time, important barriers persist, which hinder the widespread implementation of school-based mental health screening. These barriers include relatively low acceptability of screening by superintendents and school professionals (Scherff et al. 2005), and dependence of screening on school personnel enthusiasm and willingness to participate (Hallfors et al. 2006). Future clinical challenges include assessing the effects of school-based screening on mental health and educational outcomes across racial/ethnic adolescent groups. At the policy level, challenges include building public and institutional trust and acceptance of voluntary mental health screening, identifying funding streams to support implementation and maintenance of screening initiatives, and integrating screening efforts into the existing array of school-based and community-based local mental health services.

Acknowledgments

We gratefully acknowledge the work of Margie Godwin and Jan Daniels in implementing mental health screening programs and granting permission to perform a review of their screening records.

Copyright information

© Springer Science+Business Media, LLC 2011