Frequency and Correlates of Suicidal Ideation in Pediatric Obsessive-Compulsive Disorder

Abstract

This study examined the frequency and sociodemographic and clinical correlates of suicidal ideation in a sample of children and adolescents with obsessive-compulsive disorder (OCD). Fifty-four youth with OCD and their parent(s) were administered the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime, Children’s Yale-Brown Obsessive Compulsive Scale, and Children’s Depression Rating Scale-Revised. Children completed the Suicidal Ideation Questionnaire - Junior, Child Obsessive Compulsive Impact Scale–Child, and Multidimensional Anxiety Scale for Children; parents completed the Child Obsessive Compulsive Impact Scale–Parent, Swanson, Nolan, and Pelham–IV Parent Scale, and Young Mania Rating Scale–Parent Version. Seven youth endorsed clinically significant levels of suicidal ideation on the Suicidal Ideation Questionnaire - Junior. Suicidal ideation was significantly related to clinician-rated depressive symptoms, age, child-rated impairment and anxiety symptoms, and symmetry, sexuality/religiosity and miscellaneous symptom dimensions. There was no significant association between suicidal ideation and obsessive-compulsive symptom severity, comorbidity patterns, or several parent-rated indices (e.g., impairment, impulsivity). These results provide initial information regarding the frequency and correlates of suicidal ideation in treatment-seeking youth with OCD. Clinical implications are discussed, as well as directions for future research.

Obsessive-compulsive disorder (OCD) occurs in approximately 1% of children and adolescents and is associated with significant impairment [1], reduced quality-of-life [2], and high rates of psychiatric comorbidity [3, 4]. Although recent attention has been given to the role of depression in pediatric OCD [5, 6], there has been limited research on the frequency and correlates of suicidal ideation in this population. Accordingly, the present study investigated the frequency, and sociodemographic and clinical correlates of suicidal ideation among children and adolescents with OCD. Implications of this research are significant with regards to assessment and intervention. Understanding correlates of suicidal ideation can guide the clinician to understand factors that are associated with higher risk of suicidal ideation. With regards to treatment, the presence of significant suicidal ideation and/or behaviors may guide personalized intervention planning (e.g., multimodal treatment; integrating treatment components tailored for suicidal ideation/depressive symptoms with exposure-based cognitive-behavioral therapy) to yield better patient outcomes.

To date, few data have been reported on suicidal ideation in youth with OCD. Among 348 adolescents consecutively admitted to an inpatient unit, rates of suicide attempts were much lower in adolescents with OCD (n = 40) compared to adolescents in other patient groups (e.g., inpatient adolescents with schizophrenia [n = 118], conduct disorder [n = 81], eating disorder [n = 50], or affective disorder [n = 59]) [7]. A non-significant inverse association between suicidal behavior and depressive symptoms was found (r = −.22), which was contrary to positive associations demonstrated in the other patient groups studied. In the 40 inpatient adolescents with OCD, there were weak, non-significant associations between suicidal behavior and trait anxiety, violence, aggression, and anger (rs = .06–.28); moderate associations were present between suicidal behavior and state anxiety and impulsivity (r = .37 and .38) [7].

Considerably more research on suicidality has been conducted in adults with OCD. Early observations initially suggested low incidence of suicidal ideation and behaviors [8–10]. However, more recent reports among adults with OCD have contradicted this notion, suggesting higher frequency of suicidal thoughts and behaviors that may be linked to the burden of the illness and its associated comorbidities [11–14]. Between 36%–63% of adults with OCD experience suicidal ideation [11–14], with as many as 25% reporting at least one prior suicide attempt [13]. Presence of major depressive disorder, posttraumatic stress disorder, substance use disorders, and impulse control disorders have been associated with suicidal behavior [11, 14]. With regards to clinical correlates, suicidal ideation has been associated with more severe obsessive-compulsive [15] and depressive symptoms [11, 16], hopelessness [11, 15], symmetry/ordering obsessions [16], and sexual, religious, and aggressive obsessions [14, 15].

To address the significant dearth of information about suicidal phenomena in youth with OCD, the present study examined the frequency and sociodemographic and clinical correlates of suicidal ideation in a well-characterized sample of children and adolescents with OCD who had qualified to participate in a clinical trial study. We were interested in addressing the following research hypotheses. First, we aimed to examine the frequency of suicidal ideation in youth with OCD. Based on findings in adults, we expected that 20% of youth would exhibit suicidal ideation. Second, we examined the extent to which suicidal ideation was associated with clinical characteristics, including obsessive-compulsive symptom severity, obsessive-compulsive symptom dimensions, depressive symptoms, anxiety symptoms, OCD-related impairment, and impulsivity. Based on the limited available literature, we anticipated that, similar to findings in adults with OCD, suicidal ideation would be directly associated with obsessive-compulsive, depressive, and anxiety symptom severity, as well as with parent- and child-rated functional impairment. No hypotheses were generated regarding associations with symptom dimensions. Third, we examined if suicidal ideation was associated with sociodemographic characteristics and comorbidity patterns (i.e., presence of: any anxiety disorder, tic disorders, externalizing disorder, and any comorbid disorder). We expected that suicidal ideation would be associated with age given evidence that the frequency of suicidal ideation increases during adolescence with a moderate decline into young adulthood; no hypotheses were generated regarding comorbidity patterns.

Method

Participants and Procedures

Participants included 54 youth (33 males, 61.1%), aged 7–17 years (Mean age = 11.87, SD = 3.22), and a parent or guardian. Participants were recruited as part of a larger multimodal treatment study across two sites with expertise in pediatric OCD treatment. Table 1 details participant characteristics. In regard to inclusion criteria, all youth had a primary diagnosis of OCD, Children’s Yale-Brown Obsessive-Compulsive Scale (CY-BOCS) total score ≥18, and were able to speak and read English. Exclusion criteria included a prior adequate trial of or allergy to sertraline, generally poor physical health, presence of active suicidality or suicide attempt in the past 12 months determined through interview with a board certified child and adolescent psychiatrist, concomitant psychotropic medications other than medication for attention deficit hyperactivity disorder or sedative/hypnotics for insomnia, or presence of comorbid psychosis, bipolar disorder, autism, anorexia, or substance abuse/dependence determined through clinical interview and administration of the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime version (K-SADS-PL) [17]. Participating youth were predominantly Caucasian (96.3%), with other youth identifying as Black (1.9%) and Asian (1.9%).

Table 1.

Characteristics of Study Participants

Variable
Age (years)
Range	7–17 years
Mean, SD	11.87, 3.22
Gender, n (%)
Male	33 (61.1%)
Female	21 (38.9%)
Ethnicity, n (%)
White	52 (96.3%)
Black	1 (1.9%)
Asian	1 (1.9%)
Psychiatric comorbidity, n (%)
At least one DSM-IV-TR Axis I disorder	38 (70.4%)
Anxiety Disorder	28 (51.9%)
Tic Disorder	12 (22.2%)
Externalizing Disorder	9 (16.7%)
Depressive Disorder	3 (5.6%)
ADHD medication, n (%)	36 (66.7%)
Clinically significant suicidality, n (%)	7 (13.0%)

Note: Anxiety Disorder category encompassed Separation Anxiety Disorder, Generalized Anxiety Disorder, Social Phobia, and Specific Phobia; Tic Disorder category included Tourette Disorder and Chronic Motor or Vocal Tic Disorder; Externalizing Disorder category encompassed Attention-Deficit/Hyperactivity Disorder (ADHD) and Oppositional-Defiant Disorder (ODD); Depressive Disorder included Major Depressive Disorder and Dysthymic Disorder.

Procedure

All procedures met standards set by the respective site’s institutional review board, and written parental consent and child assent were obtained from all participants. At the screening visit, participants were administered the K-SADS-PL [17] by a trained clinician to establish a primary diagnosis of OCD, and to assess for comorbid conditions. After administration of several clinician-administered measures, child participants and parents/caregivers completed study questionnaires independently. Raters were master’s or doctoral level clinicians who were trained by the first, second, and last authors in the administration of clinician-rated measures at study onset (and as needed thereafter) through instructional meetings, in vivo observation, and direct supervision.

Measures

Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime (K-SADS-PL) [17]

The K-SADS-PL is a clinician-administered, semi-structured diagnostic interview for DSM-IV childhood disorders. The K-SADS-PL was administered to parents and children separately, and clinical judgment was used to determine diagnoses based upon the combined data.

Children’s Yale-Brown Obsessive Compulsive Scale (CY-BOCS) [18]

The CY-BOCS included both a Symptom Checklist and a Severity Scale. The Symptom Checklist consists of 62 items that assess the presence or absence of obsessions and compulsions. The items used in these analyses were those under the categories of obsessions (i.e., contamination, aggressive, sexual, hoarding/saving, magical, somatic, and religious) and compulsions (i.e., washing, checking, repeating, counting, ordering/arranging, hoarding/saving, magical/superstitious rituals, and rituals involving other persons), as well as the items from the miscellaneous category under each (e.g., “fear of saying certain things”, “the need to tell, ask, or confess”). The CY-BOCS Symptom Checklist was used to calculate five dimension scores derived from past factor analytic research [19, 20]: contamination/washing, symmetry/ordering, hoarding/saving, aggressive/checking, and sexual/religious. A miscellaneous symptom dimension was also calculated based on past precedence [21, 22]. Presence of a current obsession or compulsion was scored as 1 for the respective symptom dimension whereas its absence was scored as a 0. Symptom dimensions were derived from summing items comprising the specific dimension, allowing for more comprehensive representation of dimensional characteristics (i.e., a dimension’s score is a measure of magnitude, as opposed to a yes/no construct). The Severity Scale is a 10-item, clinician-rated, semi-structured interview assessing the severity of obsessions and compulsions across five items (time occupied by symptoms, interference, distress, resistance, and degree of control over symptoms); all 10 Severity Scale items are summed to derive a total severity score. The CY-BOCS is widely used and both the Symptom Checklist and Severity Scale are psychometrically sound [18, 23, 24].

Suicidal Ideation Questionnaire–JR (SIQ-JR) [25, 26]

The SIQ-JR is a 15-item child-report questionnaire assessing the presence and frequency of a wide range of thoughts related to death and dying, passive and active suicidal ideation, and suicidal intent. Items are rated on a 7-point Likert-scale ranging from 0 (“I never had this thought”) to 6 (“This thought was in my mind almost every day”), and participants indicated the frequency with which such thoughts were present over the past month. Three subscale scores (Ideation, Intent, Thoughts) and a Total (summative) score are provided. Each subscale score ranges from 0–30, and the Total score ranges from 0 to 90; higher Total scores indicate greater severity of suicidal ideation, with a raw score cutoff of 31 considered indicative of significant suicidal ideation (Reynolds, 1988). Although others have provided various cutoff methods for determining significant suicidal ideation [27], we considered a multiple-path scoring procedure for designating a youth as displaying clinically significant suicidal ideation to avoid missing children who exhibited significant suicidal ideation even though they did not meet cutoff criteria. As such, in addition to the previously stated Total score cutoff of 31, youth were classified as exhibiting clinically significant suicidal ideation if: (a) the child’s response to items 2 through 4 (“I thought about killing myself”, “I thought about how I would kill myself”, “I thought about when I would kill myself”) indicated such thoughts as occurring more frequently than once in the last month (item score 3 or higher); or (b) as proposed by [25] and utilized by others [26, 28, 29], if the child rated two or more of the six “critical items” 2 through 4, 7 through 9 (“I thought about what to write in a suicide note”, “I thought about writing a will”, “I thought about telling people I plan to kill myself”) as occurring at least weekly (item score 4 or higher). Good psychometric properties have been established for the use of the SIQ-JR [26].

Children’s Depression Rating Scale-Revised (CDRS-R) [30]

The CDRS-R is a semi-structured, psychometrically-sound clinician-administered interview, which is administered concurrently to a child and adult caregiver to diagnose depression and level of severity. The clinician rates 17 symptom areas including those that comprise DSM-IV criteria for a diagnosis of depression.

Multidimensional Anxiety Scale for Children (MASC) [31]

The MASC is a psychometrically sound 39-item child self-report questionnaire that assesses symptoms of general, social, and separation anxiety. A total anxiety score is computed by summing all items.

Child Obsessive Compulsive Impact Scale–Child/Parent Versions (COIS-C/P) [32]

The COIS-C/P are psychometrically-sound child self-report or parent-report measures which assess the extent to which pediatric OCD causes impairment in specific areas of child psychosocial functioning (school activities, social activities, home/family activities).

Swanson, Nolan, and Pelham–IV Parent Scale (SNAP-IV) [33]

The SNAP-IV is a measure completed by a parent/caregiver that provides a dimensional scaling of the child’s exhibition of DSM-IV items for inattention, impulsivity, hyperactivity, and oppositionality. Adequate psychometric properties have been demonstrated, with high internal consistency [34] and good predictive validity in identifying youth with ADHD relative to non-clinical youth [35]. Only the Impulsivity subscale was used in the present study.

Young Mania Rating Scale–Parent Version (YMRS-P) [36]

The YMRS-P is an 11-item parent-rated multiple-choice measure of mania symptoms. Questions assess for elevated mood, motor activity/energy level, sexual interest, sleep, irritability, speech rate and amount, changes in thought patterns, changes in thought content, disruptive/aggressive behavior, changes in appearance, and insight about needing treatment. Items are scored from 0 to 8 with a total score calculated that ranges from 0 to 60. This measure demonstrates good internal consistency, construct validity [37], and diagnostic efficiency [36]. Although bipolar disorder was exclusionary, the YMRS-P was administered to provide a continuous assessment of manic symptoms.

Data Analysis

All statistical procedures were completed using SPSS 20. Descriptive statistics were calculated to examine suicidal ideation frequency and phenomenology, with independent samples t-tests utilized to assess for differences in severity of suicidal ideation by comorbidity status. Given the modest sample size and the frequency of certain comorbid conditions, we examined the following comorbidity patterns: presence/absence of an anxiety disorder (Separation Anxiety Disorder, Generalized Anxiety Disorder, Social Phobia, or Specific Phobia); presence/absence of a tic disorder (Tourette Disorder, or Chronic Motor or Vocal Disorder); and presence/absence of an externalizing disorder (Attention-Deficit/Hyperactivity Disorder or Oppositional-Defiant Disorder). The association between suicidal ideation and the presence/absence of a depressive disorder was not examined because the frequency of occurrence in the current sample (3 of 54 participants, or 5.6%) precluded meaningful analysis. Pearson correlation coefficients were calculated between metrics of suicidal ideation and clinical constructs. No significant violations of variable distribution assumptions were observed. Given the preliminary nature of this research, no correction for Type I error was used.

Results

Frequency of Suicidal Ideation in Youth with OCD

Table 2 details the frequency and intensity of SIQ-JR item endorsement by scale, as well as for the individual items. The majority of participants endorsed frequencies less than monthly on all scales and individual items, with the mean Total score falling well below the cutoff for clinical significance. Four of the 15 items had more than 5% of the sample endorse as occurring at least twice weekly, with 2 of these items (thinking about the death of others, and death in general) being endorsed by more than 5% of the sample as occurring daily. Seven of 54 children (13%) endorsed clinically significant suicidal ideation.

Table 2.

Frequency of SIQ-JR Item Endorsement by Subscale and Individual Items (N = 54)

Scale	Mean score (SD)
Total	10.59 (10.48)
Ideation	4.39 (4.33)
Intent	3.09 (3.20)
Thoughts	3.11 (4.55)

Specific Items n (%)	Never	Less than monthly	Monthly	Twice monthly	Weekly	Twice weekly	Almost every day
1. Advantages	24 (44.4)	19 (35.2)	4 (7.4)	2 (3.7)	1 (1.9)	4 (7.4)	0
2. Considering suicide	32 (59.3)	14 (25.9)	2 (3.6)	2 (3.6)	2 (3.6)	2 (3.6)	0
3. Planning method	39 (72.2)	10 (18.5)	1 (1.9)	2 (3.7)	2 (3.7)	0	0
4. Planning timeline	42 (77.8)	7 (13)	2 (3.7)	1 (1.9)	2 (3.7)	0	0
5. Death of others	18 (33.3)	16 (29.6)	5 (9.3)	7 (13.0)	2 (3.7)	1 (1.9)	5 (9.3)
6. Death in general	15 (27.8)	18 (33.3)	5 (9.3)	6 (11.1)	3 (5.6)	3 (5.6)	4 (7.4)
7. Planning note	45 (83.3)	8 (14.8)	1 (1.9)	0	0	0	0
8. Planning a will	46 (85.2)	8 (14.3)	0	0	0	0	0
9. Telling others	48 (88.9)	4 (7.4)	1 (1.8)	0	0	1 (1.8)	0
10. Impact on others	33 (61.1)	10 (18.5)	5 (9.3)	4 (7.4)	1 (1.9)	1 (1.9)	0
11. Desire to die	31 (57.4)	16 (29.6)	0	4 (7.4)	1 (1.9)	2 (3.7)	0
12. See as solution	36 (66.7)	10 (18.5)	1 (1.9)	4 (7.4)	0	3 (5.6)	0
13. Pleasing others	37 (68.5)	12 (22.2)	2 (3.7)	2 (3.7)	0	1 (1.9)	0
14. Regret existence	34 (63.0)	13 (24.1)	2 (3.7)	2 (3.7)	2 (3.7)	1 (1.9)	0
15. Apathy of others	38 (70.4)	10 (18.5)	5 (9.3)	0	1 (1.9)	0	0

Associations between Suicidal Ideation and Clinical Characteristics

Table 3 displays single-order correlations, internal consistency estimates, means and standard deviation for study variables. Suicidal ideation was strongly related to age, child-rated impairment, child-rated anxiety, and symmetry and sexuality/religiosity symptom dimensions. Moderate relationships were observed between suicidal ideation and clinician-rated depressive symptom severity, as well as the miscellaneous symptom dimension. Suicidal ideation was not significantly associated with gender, obsessive-compulsive symptom severity, parent-rated impairment, parent-rated impulsivity and mania symptoms, or hoarding, contamination and aggression/checking obsessive-compulsive symptom dimensions.

Table 3.

Pearson Correlations, Internal Consistency, and Descriptive Statistics for Study Variables (N = 54)

Measure	SIQ-JR				α	Mean	SD
	Total	Ideation	Intent	Thoughts
SIQ-JR Total	–				.86	10.59	10.48
Ideation	.903**	–			.64	4.39	4.33
Intent	.825**	.693**	–		.54	3.09	3.20
Thoughts	.863**	.642**	.537**	–	.86	3.11	4.55
CY-BOCS Total	−.017	−.079	−.130	.128	.82	24.69	4.83
Obsessions	.025	−.077	−.087	.192	.71	11.91	2.66
Compulsions	−.056	−.067	−.151	.040	.71	12.78	2.62
Miscellaneous	.270*	.234	.034	.376**	.75	1.52	2.12
Contamination	.054	−.023	−.062	.188	.69	1.24	1.55
Symmetry/Ordering	.424**	.429**	.345*	.325*	.59	1.56	1.52
Hoarding	−.064	−.107	−.011	−.037	.59	0.74	0.96
Sexual/Religious	.351**	.323*	.202	.358**	.36	1.17	1.11
Aggressive/Checking	.144	.059	.089	.212	.67	1.94	2.08
CDRS T-score	.313*	.150	.183	.385**	.78	55.06	9.90
MASC Total	.409**	.289*	.370**	.406**	.90	51.83	19.26
COIS-Parent	.068	.015	.067	.096	.91	22.98	14.11
COIS-Child	.439**	.351*	.273*	.466**	.87	18.29	11.34
SNAP-IV Impulsivity	−.064	.016	−.218	−.009	.89	6.94	6.12
YMRS-P	.230	.192	.133	.254	.69	7.38	6.85
Age (years)	.507**	.350**	.474**	.501**	–	11.87	3.22

Note: SIQ-JR = Suicidal Ideation Questionnaire-JR; CY-BOCS = Children’s Yale-Brown Obsessive-Compulsive Scale; CDRS = Children’s Depression Rating Scale; MASC = Multidimensional Anxiety Scale for Children; COIS-C/P = Child Obsessive Compulsive Impact Scale–Child/Parent; SNAP-IV = Swanson, Nolan, and Pelham–IV; YMRS-P = Young Mania Rating Scale–Parent.

^*ρ<.05; ρ<.01.

There were no significant relationship observed between suicidal ideation and status (presence/absence) of comorbid anxiety (t(52)=0.38, ρ>.05), tic (t(52)=1.26, ρ>.05), or externalizing disorders (t=1.31, ρ>.05), nor did the total number of comorbid conditions diagnosed significantly predict suicidal ideation (R²=.04, F(1,52)=2.25, ρ>.05).

Discussion

This study examined the frequency and sociodemographic and clinical correlates of suicidal ideation in youth with OCD. Overall, 13% of the current sample endorsed clinically significant levels of suicidal ideation, which is somewhat lower than estimates from adult studies (36–63%) [11–14], but may be due in part to study exclusion criteria, in which children with active suicidality, a recent attempt, and certain comorbidities (e.g., bipolar disorder) were not enrolled. Nonetheless, these data indicate that a sizable number of youngsters with OCD experience clinically significant suicidal ideation, and highlight the relevance of its assessment in the clinical care of youth with OCD. The positive association of age with suicidal ideation is consistent with higher rates found in adults, and may be understood in several potential ways. First, younger children may be less knowledgeable about death in general or have less accessibility to mechanisms of self-harm, precluding them from considering suicide as a potential option. Second, there are more domains that could be negatively impacted by OCD as children age (e.g., academic performance, social expectations). Older youth may experience more distress as impairment is compounded, thereby leading to increased suicidal ideation. Indeed, significant associations were found between child-rated impairment and suicidal ideation. As children’s perception of the potential for OCD to negatively impact their life increases, feelings of hopelessness and discouragement may also increase, resulting in heightened suicidal ideation.

Findings that suicidality was significantly associated with child-rated anxiety are consistent with Apter et al. [7] in which moderate relations were found between suicidal behavior and state anxiety. Highly anxious children may experience decreased self-efficacy in regard to their ability to manage obsessive-compulsive symptoms and, consequently, may limit symptom resistance, which will only serve to maintain their disorder and leave them feeling more discouraged. Child-rated depressive symptom severity was positively related to suicidal ideation, which is consistent with findings in adults with OCD [11, 14] and other studies that have shown children’s depression to be associated with increased suicidality [38, 39]. Further, since child-rated suicidal ideation was positively associated with child-rated impairment and anxiety, and depressive symptom severity, it is possible that children, compared to others, are most in tune with their feelings and how OCD is affecting them. Yet another possibility is that these results are explained in part by shared method variance.

Regarding symptom topology, suicidal ideation was positively related to symmetry/ordering, sexuality/religiosity, and miscellaneous obsessive-compulsive symptoms. Alonso et al. [16] found symmetry/ordering obsessions to be associated with suicidal acts in adults with OCD. Though the specific nature of this relationship is still uncertain, some data indicate that the symmetry/ordering dimension is associated with distinct neural substrates [40, 41] and impairment on some measures of cognitive functioning (e.g., set shifting tasks) [42]. We speculate that different obsessive-compulsive symptom dimensions, such as symmetry/ordering, may be associated with neuropsychological functioning in distinct ways which, consequently, might result in varying levels of distress and comorbid symptoms. These markers may translate into heightened levels of distress as reflected by suicidal ideation in trying to manage symptoms, although further research is required to determine whether neuropsychological differences associated with different symptom dimensions are in any way related to suicidal ideation. Findings that the sexuality/religiosity dimension was associated with suicidal ideation are consistent with previous research [14, 43]. The increased suicidal ideation in youth with sexual and/or religious obsessions, which is consistent with findings in adults [11, 14, 43], may be due to distress related to having thoughts that they deem impure or inappropriate; indeed, some evidence in adults suggests that sexual and/or religious obsessions are associated with greater levels of distress relative to other symptom dimensions [44]. Even further, children may keep such obsessions to themselves to prevent negative judgments from others, which may compound distress and isolation, contributing to suicidal ideation. Interestingly, the miscellaneous symptom dimension was also significantly associated with suicidal ideation. Review of particular items of the miscellaneous symptom scale may help explain our findings. Examples of such items on this scale include the need to know or remember, fear of saying certain things, fear of not saying just the right thing, intrusive images, sounds, words, music or numbers, mental rituals, need to tell, ask, or confess, measures to prevent harm to self, others, or to prevent terrible consequences, or self-damaging behavior. Accordingly, youth with a frequent need to tell, ask, or confess may feel distressed due to their frequent pursuit of reassurance or if they do not receive reassurance. Miscellaneous symptoms have also been associated with symmetry/ordering and sexuality/religiosity symptoms [21], which may account for this association.

Several non-significant findings are also of interest. First, OCD severity was not directly related to suicidal ideation. As discussed previously, this may be due to exclusion criteria for this study, OCD-related impairment or disability caused by comorbid conditions, or other social and/or family factors that are independent of degree of severity. On the other hand, rates of suicidal ideation found in this study were not significantly greater than rates in community [45] and lower than those found in youth with non-OCD anxiety disorders [46], perhaps suggesting that youth with OCD may not differ markedly from others in terms of their overall risk. However, this should not suggest that clinicians be less concerned about rates of suicidal ideation in this cohort. Indeed, suicidal ideation and behaviors should be comprehensively and thoroughly assessed at treatment initiation and during the course of intervention. Second, comorbidity status was not associated with severity of suicidal ideation. This is in contrast to findings that mood disorders, anxiety disorders, and disruptive behavior disorders increase the risk for suicidal thoughts and behaviors in youth [47, 48]. On balance, small percentages of youth with mood and disruptive behavior disorders, and our exclusion of bipolar disorder, limit our ability to examine associations. Third, impulsivity was not linked to suicidal ideation, as it has been in samples of adolescents without OCD [49]. In contrast to impulse control disorders, OCD is characterized by harm avoidance [50] which may reflect the lack of association.

Several limitations should be noted and results should be interpreted with caution. First, although well-characterized, the sample was modest in size, composed of predominantly Caucasian youth, and excluded youth with active suicidality or a recent attempt, thus limiting generalizability. The possibility of Type II error is elevated in the presence of modest statistical power and future studies with more representative and/or clinically different (i.e., naturalistic) samples may yield different findings. Second, other factors not examined in this study (e.g., illness onset and duration, social support) may be relevant to understanding suicidal ideation in youth with OCD. With a larger sample, moderators (e.g., does age moderate the relationship between obsessive-compulsive severity and suicidal ideation?) and mediators should be examined to further our understanding of which youth are at risk and the mechanisms in action. Third, suicidal ideation was assessed solely by child report; this study lacks a more qualitative account of suicidality as well as parental perspectives that may help understand various antecedents and functions. Fourth, we utilized a single respondent for some of our measures (e.g., anxiety), and thus future research should use multi-informant assessment of clinical characteristics. Finally, this study was cross-sectional in nature, therefore preventing statements of causality.

Within these limitations, this study presents some of the first data on suicidal ideation among youth with OCD, and therefore has important implications. With regards to assessment, these data provide some insight into which youngsters are at increased risk for exhibiting suicidal ideation. In particular, suicidal ideation was associated with depressive and anxiety symptoms, older age, functional impairment, and certain symptom domains. Although assessment of suicidal ideation and behavior should be conducted with every patient with information being drawn from multiple sources, clinicians should be aware of potential risk factors that may elevate risk. Should a child exhibit suicidal ideation, the clinician must consider the level of risk along with underlying diagnoses and other psychosocial factors. If suicidal ideation is present, this should be assessed at each visit and as appropriate otherwise until resolution. As well, should correlates found in this study and by others in non-OCD samples [51] arise during treatment, the clinician should systematically assess for the presence of suicidal thoughts and behavior.

With regards to treatment, psychosocial and pharmacological interventions for youth with OCD and youth with depression are robust in efficacy [52, 53]. Effective treatment of OCD is frequently associated with reductions in depressive symptoms [54, 55] as OCD symptoms improve and become less interfering. It remains an empirical question as to if suicidal ideation on average would also decrease with successful treatment. However, in the presence of significant suicidal thoughts and behaviors at treatment presentation, sequential treatment with antidepressant therapy followed by cognitive-behavioral therapy may be warranted in some cases given evidence that depression comorbidity can attenuate cognitive-behavioral therapy outcome [3]. Alternatively, integrating treatment components tailored for suicidal ideation and depressive symptoms (e.g., cognitive restructuring, affect regulation, problem solving, behavioral activation; [56]) within an exposure-based cognitive-behavioral therapy protocol may be appropriate for other youth. Ultimately, determining individualized approaches for specific patient subgroups is needed to maximize care and thus, this is highlighted as a primary area of future research.

Summary

This study reports some of the first data on the frequency and sociodemographic and clinical correlates of suicidal ideation in 54 youth with OCD. Although suicidal ideation was not significantly related to obsessive-compulsive symptom severity, comorbidity patterns, or parent-rated impairment, direct and modest associations were found with clinician-rated depressive symptoms, age, child-rated impairment and anxiety symptoms, and symmetry, sexuality/religiosity and miscellaneous symptom dimensions. Although study limitations must be considered, these data provide initial information regarding the frequency and correlates of suicidal ideation in treatment-seeking youth with OCD.