MENTAL HEALTH OF CHILDREN AND ADOLESCENTS

Evaluation and Disposition of Medicaid-Insured

Children and Adolescents With Suicide Attempts

Candice L. Williams, MD; William O. Cooper, MD, MPH; Leanne S. Balmer, RN;Judith A. Dudley, BS; Patricia S. Gideon, RN; Michelle M. DeRanieri, RN;Shannon M. Stratton, BS; S. Todd Callahan, MD, MPH

From the Divisions of General Pediatrics (Dr Williams and Dr Cooper), Adolescent and Young Adult Health (Dr Callahan), Department ofPediatrics, andDepartment of PreventiveMedicine (MsBalmer,MsDudley,MsGideon,MsDeRanieri, andMsStratton), Vanderbilt UniversityMedical Center, Nashville, TennThe authors declare that they have no conflict of interest.Address correspondence to S. ToddCallahan,MD,MPH, Division of Adolescent and YoungAdult Health, MonroeCarell Jr Children’s Hospitalat Vanderbilt, One Hundred Oaks, 719 Thompson Ln, Suite 36300, Nashville, TN 37204 (e-mail: todd.callahan@vanderbilt.edu).Received for publication January 21, 2014; accepted April 26, 2014.

ABSTRACT

AC

OBJECTIVE: Guidelines and quality of care measures for theevaluation of adolescent suicidal behavior recommend promptmental health evaluation, hospitalization of high-risk youth,and specific follow-up plans—all of which may be influencedby sociodemographic factors. The aim of this study was to iden-tify sociodemographic characteristics associated with variationsin the evaluation of youth with suicidal behavior.METHODS:We conducted a large cohort study of youth, aged 7to 18, enrolled in Tennessee Medicaid from 1995 to 2006, whofilled prescriptions for antidepressants and who presented forevaluation of injuries that were determined to be suicidal onthe basis of external cause-of-injury codes (E codes) and ICD-9-CM codes and review of individual medical records. Chi-square tests and logistic regression were performed to assessthe relationship between sociodemographic characteristics anddocumentation of mental health evaluation, hospitalization,and discharge instructions.

CADEMIC PEDIATRICSopyright ª 2015 by Academic Pediatric Association 36

RESULTS: Of 929 episodes of suicidal behavior evaluated in anacute setting, rural-residing youth were less likely to beadmitted to a psychiatric hospital (adjusted odds ratio [AOR]0.72; 95% confidence interval [CI] 0.55–0.95) and more likelyto be medically hospitalized only (AOR 1.92; 95% CI 1.39–2.65). Female subjects were less likely to be admitted to a psy-chiatric hospital (AOR 0.55; 95%CI 0.41–0.74) and more likelyto be discharged home (AOR 1.44; 95% CI 1.01–2.04). Only40% of those discharged to home had documentation ofdischarge instructions with both follow-up provider and date.CONCLUSIONS: In this statewide cohort of youth with suicidalbehavior, there were significant differences in disposition asso-ciated with sociodemographic characteristics.

KEYWORDS: adolescents; Medicaid; suicide

ACADEMIC PEDIATRICS 2015;15:36–40

WHAT’S NEW

In this statewide cohort of Medicaid-insured youth withsuicidal behavior, those from rural areas were morelikely to bemedically hospitalized. Of youth dischargedhome after acute evaluation of suicidal behavior, a mi-nority had documentation of a specific follow-up plan.

SUICIDE IS THE third leading cause of death in youngpeople aged 10 to 24 years in the United States, accountingfor approximately 4500 deaths every year.1 Nonfatal sui-cidal behavior (suicidal ideation or attempted suicide) iseven more common among youth and is an important pre-dictor of future suicide attempts and completed suicide.2

Annually, approximately 150,000 youth seek treatment inan emergency department after a self-harm event.3,4

Because mental health care after suicidal behavior iscritical to reduce the risk of future suicide,5 several

regional and national efforts have sought to improveand standardize the evaluation and treatment of adolescentself-injurious behavior.2,6–9 Guidelines from the AmericanAcademy of Child and Adolescent Psychiatry state thatyouth presenting to emergency departments with suicidalbehavior should undergo mental health evaluation forassessment and triage.10 Hospitalization in a medical orpsychiatric unit is recommended unless there is certaintyabout medical and psychiatric stability. In cases appro-priate for outpatient follow-up, patient discharge tohome/community is permissible if oversight from support-ive adults in a safe/secured environment can be assured andif specific follow-up, including provider, date, and time,has been arranged.Several nonclinical factors, including health insurance

coverage and sociodemographic factors (age, gender, andgeographic residence), are likely to be influential in theevaluation and disposition of youth with suicidal behavior.As the enactment of the Affordable Care Act increases the

Volume 15, Number 1January–February 2015

ACADEMIC PEDIATRICS MEDICAID AND SUICIDE ATTEMPTS 37

availability and parity of mental health insurance coverage,the influence of sociodemographic characteristics on theprovision of health care is of increasing interest. New ini-tiatives, including accountable care organizations, whichseek to integrate mental and primary health care in an effi-cient, cost-effective manner, require an understanding ofhow sociodemographic characteristics influence healthcare delivery. Small studies have documented the influenceof health insurance coverage; however, there are fewpopulation-based data on the relationship of sociodemo-graphic characteristics and the assessment and dispositionof adolescent suicidal behavior.11–14

The aims of this study were to identify and characterizevariations in the evaluation and disposition of youth pre-senting for medical evaluation of suicidal behavior associ-ated with sociodemographic characteristics. To addressthese aims, we utilized data from an ongoing study of youthin Tennessee’s Medicaid population who were recentlyprescribed antidepressants and who presented for evalua-tion of injury episodes that were confirmed to be suicidalin nature. This cohort provided a unique opportunity toexamine these factors in an insured population with well-characterized suicidal behavior that was relatively homo-geneous in terms of psychiatric risk.

METHODS

The study was performed as part of a larger retrospectivecohort study of antidepressant use and suicidal behavior inchildren and adolescents that included 80,183 youth (aged6 to 18 years) who were enrolled in Tennessee’s MedicaidProgram (TennCare) between 1995 and 2006 and whowereprescribed an antidepressant medication.15 The methodsfor identifying and confirming suicidal behavior in thiscohort have been previously described.16 Briefly, TennCareclaims data and linked death certificates were queried forInternational Classification of Diseases, 9th Revision,Clinical Modification (ICD-9-CM), codes and externalcause-of-injury codes (E codes) corresponding to potentialepisodes of suicidal behavior, as outlined previously.16 Foreach episode identified by these claims, medical recordswere sought and adjudicated to determine whether aninjury was deliberately self-inflicted and whether intentto die was explicitly stated or could be inferred using def-initions from the Columbia Classification Algorithm ofSuicide Assessment.17 For the larger study, 87% of epi-sodes identified by medical claims were obtained andhad adequate documentation for adjudication.

For this study examining the evaluation and dispositionof youth presenting with suicidal behavior, we excludedepisodes of completed suicide. We reviewed 965 suicidalepisodes (all episodes adjudicated as confirmed suicide at-tempts, preparatory actions toward suicidal behavior, andsuicidal ideation only). We excluded 36 records (4%)that were incomplete or that indicated that the subjectleft against medical advice. For the remaining 929 suicidalepisodes, medical records were reviewed by the principalinvestigator and trained research nurses to record key char-acteristics of the history, evaluation, and disposition. The

method of injury was characterized as cutting/stabbing,gunshot, hanging/asphyxiation, jump from height, inges-tion, or other, or suicidal ideation only. Clinical documen-tation of prior psychiatric history and prior suicide attemptswere recorded. Disposition was characterized as hospitali-zation (psychiatric [including youth transferred to a psy-chiatric hospital after admission to a medical hospital],medical only) or discharge to home or the community.We reviewed records for documentation of a mental healthevaluation by a psychiatric physician or nurse, socialworker, psychologist, counselor, or representative of amental health organization at any time before dischargefrom the acute care or medical hospital setting. For youthwho were discharged to home, we reviewed discharge in-structions for documentation of a specific follow-up pro-vider and date. Subjects were characterized as havingreceived complete discharge instructions (both specificprovider and date provided), partial discharge instructions(either specific provider or date provided), or no dischargeinstructions (neither specific provider nor date provided).We recorded sociodemographic data including gender,

age, race, and geographic residence of the parent/guardian.Gender and age at the time of the event were determined bymedical record documentation. Because of differences inthe epidemiology of suicide attempts between younger ad-olescents and older adolescents, age was subdivided into 2groups (7 to 14 years and 15 to 18 years).18 Race was cate-gorized as black, white, or other as self-reported by theparent/guardian and documented in administrative claims.Geographic residence using the parent/guardian’s addresswas defined as rural, suburban, or urban using StandardMetropolitan Statistical Area definitions.19 We chose tofocus on the geographic residence of the youth ratherthan where they sought care. Although youth may bemore likely to seek care in a setting close to them, youthfrom rural areas may seek care from facilities in nonruralareas. When youth seek care outside of their community,disposition and follow-up are likely to be influenced bythe availability of resources both at the treating facilityand in the child’s community.In the larger study, we found that less than 1% of the sui-

cidal episodes were repeat events; therefore, we performedchi-square analyses to assess the relationship between soci-odemographic characteristics and receipt of mental healthevaluation and disposition. Multivariate logistic regressionmodels were created to estimate the odds of receiving amental health evaluation, being hospitalized in a medicalor psychiatric facility, or being discharged home, after ad-justing for the other sociodemographic variables (race, age,gender, geographic residence). For youth who were dis-charged to home, models assessed the relationship betweensociodemographic characteristics and the receipt of mentalhealth evaluation and characteristics of follow-up instruc-tions. Statistical calculations were performed by Statav12.1 software (StataCorp, College Station, Tex).Permission to use the study data was obtained from the

Tennessee Department of Health and the TennCare Bureau.The study was reviewed and approved by the VanderbiltUniversity institutional review board.

Table 1. Evaluation and Disposition of Adolescents With Suicidal Behavior by Sociodemographic Characteristics

Characteristic Total

Mental Health

Evaluation

Disposition

Complete Follow-up

Instructions Documented†

Any Psychiatric

Hospitalization

Medical

Hospitalization Only

Discharged

Home

n 929 749 (81%) 473 (51%) 222 (24%) 231 (25%) 94 (41%)Gender

Male 257 (28%) 206 (80%) 157 (61%)* 48 (19%)* 52 (20%)* 19 (37%)Female 672 (72%) 543 (81%) 316 (47%) 174 (26%) 179 (27%) 75 (41%)

RaceWhite 758 (82%) 623 (82%) 383 (51%) 184 (24%) 189 (25%) 76 (40%)Black 139 (15%) 107 (77%) 75 (54%) 29 (21%) 34 (24%) 15 (44%)

Age group7–14 y 406 (44%) 315 (76%)* 214 (52%) 82 (20%)* 108 (27%) 47 (43%)15–18 y 523 (56%) 434 (83%) 259 (50%) 140 (27%) 123 (24%) 47 (38%)

Geographic residenceNonrural 542 (58%) 433 (80%) 295 (54%)* 102 (19%)* 143 (26%) 62 (43%)Rural 387 (42%) 316 (82%) 178 (46%) 120 (31%) 88 (23%) 32 (36%)

*P < .05.

†For those patients discharged to home only; complete discharge instructions include both a specific follow-up provider and specific

follow-up date.

38 WILLIAMS ET AL ACADEMIC PEDIATRICS

RESULTS

Of the 929 youth with confirmed suicidal events, 72%were girls, 82% were white, and 42% resided in a ruralarea. Fifty-six percent of the cohort was aged 15 to 18years. Ingestion was the most common mechanism ofinjury, accounting for 81% of the suicidal episodes, andcutting was the second most common (9%). The presenceof a psychiatric history was documented in 91% of thecases, and a previous suicide attempt was documented in41% of the cases. Inconsistencies in clinical documenta-tion of psychiatric history limited our ability to determinefrequencies of specific psychiatric diagnoses (ie, suicideattempt, bipolar mood disorder, or substance abuse) or theirrelationships to evaluation and disposition.

Table 1 displays the elements of the evaluation anddisposition of adolescents with suicidal behavior by socio-demographic characteristics. A mental health evaluationwas documented for 81% of episodes of suicidal behavior.Older adolescents were significantly more likely thanyounger adolescents to have a mental health evaluation(83% vs 76%; P ¼ .04); however, there were no significantdifferences in receipt of a mental health evaluation associ-ated with gender, race, or area of geographic residence.Youth who were hospitalized in a medical facility weremore likely to have a mental health evaluation than thosewho were discharged home (83% vs 69%, P ¼ .001).

After evaluation of suicidal episodes, 51% of youth werehospitalized in a psychiatric facility, 24% in a medical facil-ity only, and 25% were discharged to home from the acutesetting. Relative to youth from nonrural residences, youthfrom rural areas were significantly less likely to be admittedto a psychiatric hospital (54% vs 46%; P < .05) and signif-icantly more likely to be admitted to a medical hospital only(31% vs 19%; P < .05). There was no significant differencein the proportion discharged from the emergency depart-ment for rural versus nonrural residence. Girls were signif-icantly less likely than boys to be admitted to a psychiatrichospital (47% vs 61%; P< .05) but significantly more likelyto be admitted to a medical hospital only (26% vs 19%;

P< .05) and to be discharged to home from the acute setting(27% vs 20%; P < .05). Youth aged 15 to 18 years weremore likely to be admitted to a medical hospital than youthaged 7 to 14 years (27% vs 20%; P < .05).

Of the 25% of confirmed suicidal episodes that resultedin being discharged to home from the acute setting, 41%had documentation of specific discharge instructions withboth a specific follow-up provider and time, an additional42% had discharge instructions to follow-up with a specificprovider but without a specific time, and 13% lacked in-structions for follow-up. There were no differences inreceipt of discharge instructions by geographic residence,race, age, or gender.

Table 2 displays the adjusted odds of receipt of mentalhealth evaluation, disposition, and receipt of discharge in-structions after controlling for other sociodemographiccharacteristics. In the multivariate analysis, there were nosignificant differences in receipt of mental health evalua-tion associated with sociodemographic characteristics.Relative to boys, girls had significantly lower adjustedodds of being admitted to a psychiatric hospital (adjustedodds ratio [AOR] 0.55; 95% confidence interval [CI]0.41–0.74), and significantly higher adjusted odds of beingadmitted to a medical hospital only (AOR 1.57; 95% CI1.09–2.26) and being discharged from the acute setting tohome or the community (AOR 1.44; 95% CI 1.01–2.04).Youth from rural areas had significantly lower adjustedodds of being admitted to a psychiatric hospital (AOR0.72; 95% CI 0.55–0.95) and nearly twice the adjustedodds of being admitted to a medical hospital only (AOR1.92; 95% CI 1.39–2.65). There were no significant differ-ences in the adjusted odds of psychiatric hospitalization,medical hospitalization, or discharge from emergencydepartment associated with age or race.

DISCUSSION

In this large statewide cohort study of Medicaid-insuredadolescents who presented with suicidal behavior, we

Table 2. Multivariate Analysis of Evaluation and Disposition of Adolescents With Suicidal Behavior by Sociodemographic Characteristics

AOR* (95% CI) for:

Mental Health

Evaluation

Disposition

Specific Follow-up

Instructions Documented†

Any Psychiatric

Hospitalization

Medical

Hospitalization Only

Discharged

Home

GenderMale 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref)Female 1.04 (0.72–1.51) 0.55 (0.41–0.74)‡ 1.57 (1.09–2.26)‡ 1.44 (1.01–2.04)‡ 1.21 (0.47–3.07)

RaceWhite 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref)Black 0.75 (0.48–1.19) 1.04 (0.71–1.53) 1.07 (0.67–1.72) 0.87 (0.56–1.35) 2.90 (0.63–13.15)

Age group7–14 y 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref)15–18 y 1.36 (0.98–1.90) 0.90 (0.69–1.17) 1.36 (0.99–1.87) 0.87 (0.64–1.18) 1.03 (0.46–2.30)

Geographic residenceNonrural 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref) 1.00 (Ref)Rural 1.02 (0.72–1.45) 0.72 (0.55–0.95)‡ 1.92 (1.39–2.65)‡ 0.81 (0.59–1.12) 1.40 (0.60–3.29)

AOR ¼ adjusted odds ratio; CI ¼ confidence interval.

*Odds are adjusted for each of the other sociodemographic characteristics listed.

†For those patients discharged to home only; complete discharge instructions include both a specific follow-up provider and specific follow-

up date.

‡P < .05.

ACADEMIC PEDIATRICS MEDICAID AND SUICIDE ATTEMPTS 39

identified differences in the evaluation and disposition ofyouth that were associated with sociodemographic charac-teristics. Some of these differences can be attributed atleast in part to the epidemiology of suicide risk. Other dif-ferences are more likely to reflect differences in availableresources. We also identified potential missed opportu-nities for mental health evaluation and the documentationof discharge follow-up.

Youth from rural areas weremore likely to be admitted toa medical hospital and less likely to be admitted to a psychi-atric hospital than youth residing in nonrural areas. Thesefindings persisted even after controlling for gender, race,and age. Studies show that youth from rural areas are dispro-portionately more likely to attempt and complete suicidethan youth from suburban/urban regions. Among the rea-sons posited for this disparity are that rural youth are likelyto have lower socioeconomic status and poorer access tohealth care.20,21 In this study, the mechanisms for suicidalbehavior utilized by youth from rural and nonruralsettings were similar. In a policy statement on treatingsuicidal behavior, the American Academy of Pediatricsnoted the influence of availability of facilities, insurancecoverage, and managed care policies on the choicebetween medical and psychiatric hospitalization after asuicide attempt.22 The youth in our study were enrolled inthe state’s Medicaid program, so it seems likely that thedifferences in medical and psychiatric hospitalization forthese youth are a reflection of differential resources. In addi-tion to assessing and treating the medical needs of youthwith suicidal behavior, medical hospitalization ensuressafety and may facilitate mental health evaluation for youthin rural settings, where mental health resources may be lessaccessible.22,23 Differences in hospitalization and theireffects on hospital resource utilization and repeat suicidalbehavior represent important areas for future inquiry.

Girls with suicidal behavior were less likely than boys tobe admitted to a psychiatric facility and were more likely to

be discharged home from the acute setting. Epidemiologicdata show that male adolescents are less likely to attemptsuicide but are more likely to complete suicide than femaleadolescents.8,24 Recognition that boys are at higher risk forsuicide completion is reflected in clinical guidelines, whichshould lead to more intensive evaluation of adolescentmale subjects’ suicidal behavior.2

We found no significant disparities in receipt of mentalhealth evaluation by youth from rural and nonrural areas.In addition to the potential use of hospitalization to facili-tate evaluation of youth in rural areas, this parity may alsobe attributable to the widespread presence of mobile crisisteams (MCTs) in Tennessee. MCTs were developed in the1970s in response to the deinstitutionalization of patientswith mental illness and have been used to complement orreplace emergency department psychiatric care.25 Studiesshow that MCTs improve access to care for patients, pro-vide support for families, and reduce the number of inpa-tient psychiatric hospitalizations.25–27 With a presence ineach of Tennessee’s 95 counties, it is likely that MCTshave helped to reduce potential disparities in access tomental health evaluations after suicidal behavior.Although there were no significant differences based on

sociodemographic characteristics, nearly 20% of the 929confirmed episodes of suicidal behavior in this studyoccurred in subjects who did not have documentation ofa mental health evaluation before discharge. We also foundthat less than half of youth who were discharged homefrom the acute care setting had documentation of dischargeinstructions that included follow-up with specific providerand time, and 13% had no follow-up instructions. Thesefindings are notable given that the study populationincluded high-risk youth who had recently been prescribedantidepressants. The lack of documentation of follow-upinstructions is particularly striking because it is reasonableto expect that many of these youth had a relationship withan outpatient provider who prescribed the antidepressant.

40 WILLIAMS ET AL ACADEMIC PEDIATRICS

Documentation of mental health evaluation and follow-upappointments have been proposed as potential quality-of-care measures in the emergency department managementof suicidal behavior.28 These findings suggest importantopportunities to improve the care of youth with suicidalbehavior.

This study has several limitations. It included a cohort ofyouth from a single state who were prescribed antidepres-sants and whowere insured through theMedicaid program.Tennessee is a geographically diverse state, and at the timeof the study, approximately 30% of Tennessee’s youth werecovered by TennCare. However, the findings from thisstudy may not be generalizable to youth with other medicalor mental health conditions or to youth who were not pre-scribed antidepressants. Although youth in the cohort hadbeen prescribed an antidepressant, the specialty of theoutpatient prescriber (mental health, primary care, orother) was not available. The strengths of this study includethe large number of confirmed suicidal episodes includedin analysis. Although potential episodes of suicidalbehavior were identified using administrative claims data,these claims were augmented by adjudication of the med-ical records for each episode to delineate suicidal fromnonsuicidal self-injurious behavior.

CONCLUSIONS

In this statewide cohort of Medicaid-insured youth withconfirmed suicidal behavior, those from rural residenceswere less likely to be hospitalized in a psychiatric facilityand more likely to be medically hospitalized. Of youthdischarged home after acute evaluation of suicidalbehavior, less than half had documentation of a specificfollow-up plan. With increasing focus on accountablecare, these findings suggest barriers and opportunities toimplementing evidence-based care for youth with suicidalbehavior.

ACKNOWLEDGMENTS

Funding was provided by the National Institute of Mental Health (grant

5R01MH079903-03, Suicidality Associated With Antidepressants in

TennCare Children and Adolescents) to Dr Cooper. Presented in part at

the 2013 annual meeting of the Pediatric Academic Societies, Washing-

ton, DC.

REFERENCES

1. Centers for Disease Control and Prevention; National Center for

Injury Prevention and Control. Web-based Injury Statistics Query

and Reporting System (WISQARS). Available at: http://www.cdc.

gov/injury/wisqars/index.html. Accessed October 8, 2013.

2. American Psychiatric Association. American Psychiatric Association

Practice Guidelines for the Treatment of Psychiatric Disorders. Com-

pendium 2006. Arlington, Va: American Psychiatric Association;

2006.

3. Eaton DK, Kann L, Kinchen S, et al. Youth risk behavior surveil-

lance—United States, 2011. MMWR Surveill Summ. 2012;61:1–162.

4. Ting SA, Sullivan AF, Boudreaux ED, et al. Trends in US emergency

department visits for attempted suicide and self-inflicted injury,

1993–2008. Gen Hosp Psychiatry. 2012;34:557–565.

5. Newton AS, HammMP, Bethell J, et al. Pediatric suicide-related pre-

sentations: a systematic review ofmental health care in the emergency

department. Ann Emerg Med. 2010;56:649–659.

6. Jacobs D. Screening for mental health: a resource guide for imple-

menting the Joint Commission 2007 patient safety goals on suicide.

Available at: http://www.sprc.org/library/jcsafetygoals.pdf. Accessed

October 9, 2013.

7. Kennedy SP, Baraff LJ, Suddath RL, et al. Emergency department man-

agement of suicidal adolescents. Ann Emerg Med. 2004;43:452–460.

8. Moscicki EK. Gender differences in completed and attempted sui-

cides. Ann Epidemiol. 1994;4:152–158.

9. Stewart SE, Manion IG, Davidson S. Emergency management of the

adolescent suicide attempter: a review of the literature. J Adolesc

Health. 2002;30:312–325.

10. American Academy of Child and Adolescent Psychiatry Work Group

on Quality Issues. Practice parameter for the assessment and treat-

ment of children and adolescents with suicidal behavior. J Am Acad

Child Adolesc Psychiatry. 2001;40:24S–51S.

11. Donaldson D, Spirito A, Esposito-Smythers C. Treatment for adoles-

cents following a suicide attempt: results of a pilot trial. J Am Acad

Child Adolesc Psychiatry. 2005;44:113–120.

12. Greenfield B, Larson C, Hechtman L, et al. A rapid-response outpa-

tient model for reducing hospitalization rates among suicidal adoles-

cents. Psychiatr Serv. 2002;53:1574–1579.

13. Huffman LC,WangNE, SayninaO, et al. Predictors of hospitalization

after an emergency department visit for California youths with psy-

chiatric disorders. Psychiatr Serv. 2012;63:896–905.

14. Spirito A, Boergers J, Donaldson D, et al. An intervention trial to

improve adherence to community treatment by adolescents after a sui-

cide attempt. J Am Acad Child Adolesc Psychiatry. 2002;41:435–442.

15. Cooper WO, Callahan ST, Shintani A, et al. Antidepressants and sui-

cide attempts in children. Pediatrics. 2014;133:204–210.

16. Callahan ST, Fuchs DC, Shelton RC, et al. Identifying suicidal

behavior among adolescents using administrative claims data. Phar-

macoepidemiol Drug Saf. 2013;22:769–775.

17. Posner K, Oquendo MA, Gould M, et al. Columbia Classification Al-

gorithm of Suicide Assessment (C-CASA): classification of suicidal

events in the FDA’s pediatric suicidal risk analysis of antidepressants.

Am J Psychiatry. 2007;164:1035–1043.

18. Hysinger EB, Callahan ST, Caples TL, et al. Suicidal behavior differs

among early and late adolescents treated with antidepressant agents.

Pediatrics. 2011;128:447–454.

19. United States Census Bureau. Metropolitan and micropolitan statisti-

cal areas, main. Available at: http://www.census.gov/population/

metro/. Accessed June 19, 2013.

20. Kim K, Ozegovic D, Voaklander DC. Differences in incidence of

injury between rural and urban children in Canada and the USA: a

systematic review. Inj Prev. 2012;18:264–271.

21. Xiang Y, ZhaoW, Xiang H, et al. ED visits for drug-related poisoning

in the United States, 2007. Am J Emerg Med. 2012;30:293–301.

22. Shain BN. Suicide and suicide attempts in adolescents. Pediatrics.

2007;120:669–676.

23. Gasquet I, Choquet M. Hospitalization in a pediatric ward of adoles-

cent suicide attempters admitted to general hospitals. J Adolesc

Health. 1994;15:416–422.

24. Hall RC, Platt DE. Suicide risk assessment: a review of risk factors for

suicide in 100 patients who made severe suicide attempts. Evaluation

of suicide risk in a time of managed care. Psychosomatics. 1999;40:

18–27.

25. Geller JL, Fisher WH, McDermeit M. A national survey of mobile

crisis services and their evaluation. Psychiatr Serv. 1995;46:

893–897.

26. Currier GW, Fisher SG, Caine ED. Mobile crisis team intervention to

enhance linkage of discharged suicidal emergency department pa-

tients to outpatient psychiatric services: a randomized controlled trial.

Acad Emerg Med. 2010;17:36–43.

27. Scott RL. Evaluation of a mobile crisis program: effectiveness, effi-

ciency, and consumer satisfaction. Psychiatr Serv. 2000;51:

1153–1156.

28. Rhodes AE, Bethell J, Newton AS, et al. Developing measures of

quality for the emergency department management of pediatric

suicide-related behaviors. Pediatr Emerg Care. 2012;28:1124–

1128.