Suicidal adolescents and ego defense mechanisms. Alan Apter; Doron Gothelf; Ronny Offer; Gidi Ratzoni; Israel Orbach; Sam Tyano; Cynthia R. Pfeffer.

Journal of the American Academy of Child and Adolescent Psychiatry, Nov 1997 v36 n11 p1520(8)

Author's Abstract: COPYRIGHT 1997 American Academy of Child and Adolescent Psychiatry

Objectives: To identify defense mechanisms that characterize adolescents with a range of suicidal behaviors and to differentiate them from nonsuicidal adolescents. Method: Fifty-five suicidal adolescent inpatients admitted for a definite suicide attempt were compared with 87 adolescent inpatients who had no history of suicide attempt or ideation and 81 nonpatients. Defense mechanisms were assessed by the Ego Defense Scale (EDS) which is part of a larger semistructured interview, the Child Suicide Potential Scale (CSPS), and by a self-report questionnaire, the Life Style Index (LSI). The CSPS was also used to quantify violent and suicidal behaviors. Results: On the LSI suicidal adolescent patients scored higher on denial, displacement, repression, and total defenses than the nonpatients. On the EDS they scored higher on regression, denial, projection, introjection, repression, and total defenses and lower on sublimation. LSI scores on displacement (higher) and on compensation (lower) distinguished suicidal from nonsuicidal inpatients. Denial and regression correlated positively and sublimation correlated negatively with both suicidal and violent behaviors. Introjection and repression correlated with suicidal behavior only. Conclusions: Overuse of displacement is connected with increased risk for suicidal and aggressive behaviors, while sublimation is probably a protective factor. In addition, several immature ego defenses possibly amplify aggression, which then is directed against the self by the maladaptive overuse of introjection, displacement, and repression.Key Words: suicide, violence, defense mechanism, adolescence.

Full Text: COPYRIGHT 1997 American Academy of Child and Adolescent Psychiatry

This article is an empirical examination of the defense mechanisms that characterize adolescent psychiatric patients who exhibit a range of suicidal and violent thoughts and behaviors.

Psychoanalysts have traditionally viewed the turning of aggression against the self as the basic defense of the suicidal individual (Abraham, 1957). Sandler and Joffe (1965) extended these formulations to children while adding the defenses of regression and fantasy formation to their explanations of suicidal behavior. More recently, Recklitis et al. (1992) reported empirical data that support these psychoanalytic notions in a sample of suicidal adolescents, and Spirito et al. (1989) reported withdrawal as a basic mechanism in a similar teenage population. Pfeffer et al. (1979, 1980, 1982, 1984) found that introjection is correlated with suicidal behavior in children. In a recent prospective study of children, Pfeffer et al. (1995) found that high use of repression was a protective factor against suicide attempts and high use of regression and compensation were risk factors for suicide.

A previous study (Apter et al., 1989b) of the relation between defenses and suicidal and violent behaviors was based on Plutchik and van Praag's (1989) two-stage vectorial model. According to this model, an aggressive drive or impulse, activated by a trigger, is acted upon by a variety of psychosocial, psychological, and biological variables that initially either augment or attenuate the drive. Then a new set of variables act on the drive and influence its direction, either inward or outward. Suicidal or violent behavior can thus be conceptualized as the vectorial result of competing sets of forces. This model makes explicit the widespread belief that outward aggression and suicide are interrelated both psychodynamically and socially. Using a self-report measure of defenses, Apter et al. (1989b) found that regression was more predominant in violent than in nonviolent patients. Repression appeared to have the effect of turning hostility inward (correlating with suicidal behavior), while projection and denial correlated with outward direction of hostility (violent behavior).

Some of this literature is contradictory and inconsistent because it is clearly difficult to empirically measure abstract psychoanalytic notions such as defense mechanisms and the measures used are different in nature and conception. However, on the basis of these psychoanalytic formulations and the empirical studies quoted, we decided that the following hypotheses could be formulated and tested:

1. There are specific defense mechanisms that are more typical of adolescents who display a range of suicidal thoughts and actions so that the frequency of their use by this group is higher than that observed in two control groups - nonpatient teenagers and nonsuicidal adolescent inpatients.

2. There is an overlap between suicidal behavior and violence. Some defense mechanisms correlate with both violent and suicidal behavior whereas other defenses [TABULAR DATA FOR TABLE 1 OMITTED] yield differential correlations with violence on the one hand and with suicidal behavior on the other.

METHOD

Subjects

The suicidal adolescent inpatient group was composed of 55 consecutive admissions to a university-affiliated inpatient unit over a period of 12 months for a definite suicide attempt.

The nonsuicidal adolescent patients consisted of 87 consecutive admissions, for reasons other than suicidal behavior, to the same unit during the same period of time. This group of patients also did not have a history of suicide attempts or clinically significant ideation. Three patients (one with affective disorder and two with schizophrenia) were excluded from the study as they had a past history of suicidal behavior which had no relevance for their present admission. Reasons for admission included severe mental distress, cessation of normal functioning, or drastically impaired functioning accompanied by psychiatric symptomatology.

No attempt was made to select any particular diagnostic category of patients because clinical experience indicates that patients of all types, both children and adolescents, are at risk for suicide or violence or both (Apter et al., 1995; Pfeifer et al., 1983; Runeson, 1989).

The nonpatient group consisted of 81 adolescents drawn from the high school population in the same catchment area as that served by the hospital. All subjects in all three groups were Jewish. All the study subjects, both patients and nonpatients, and their parents signed informed consent forms.

The demographic and clinical characteristics of the population are shown in Table 1.

Measures

Life Style Index. The Life Style Index (LSI) was designed to assess ego defense mechanisms as conceptualized by Plutchik et al. (1979). It is a self-report questionnaire with 97 items. The eight defenses measured are compensation, displacement, projection, reaction formation, denial, intellectualization, regression, and repression. The items in the questionnaire are sentences describing feelings that arise in different situations (for example: "People who boss other people around make me furious") or certain behaviors (for example: "I have trouble getting rid of anything that belongs to me"). The subject is required to state whether a given answer applies to him or her (answer "yes") or not (answer "no"). Scores are also given as percentiles based on scores of a normative group of college students. Several reports have demonstrated the validity of the scale. The scale successfully discriminates between schizophrenic inpatients and normal controls (Plutchik et al., 1979) and between suicidal versus nonsuicidal and violent versus nonviolent adult inpatients (Apter et al., 1989b). The internal consistency for each of the LSI's ego defense mechanisms obtained from a data sample of psychiatric inpatients and a sample of college students as well as the test-retest correlations for an interval of 5 months are, respectively, as follows (Conte and Apter, 1995): compensation (.59, .43, .61); denial (.54, .52, .55); displacement (.69, .62, .76); intellectualization (.58, .30, .61); projection (.86, .75, .75); reaction formation (.73, .63, .76); regression (.65, .56, .38); and repression (.55, .38, .48).

Sign In Sign-Up Welcome! Close Would you like to make this site your homepage? It's fast and easy... Yes, Please make this my home page! No Thanks   Don't show this to me again. Close

Child Suicide Potential Scale. The Child Suicide Potential Scale (CSPS), a semistructured interview, was developed by Pfeffer et al. (1979, 1980, 1982, 1984) with the goal of investigating the potential for suicidal behavior in children. The interview includes several scales designed to identify factors related to suicidal behavior in the 6- to 12-year-old age group. The interview is given to the child and his or her parents, and the information compiled is used to score nine different scales. Reliability and validity of the CSPS have been the subject of several studies by Pfeifer et al. (1979, 1980, 1982, 1984). The interview has nine parts. Each part consists of questions designed to elicit information about the child, the child's behavior, feelings, family history, ego functions, and concept of death. Three sections of the interview - Spectrum of Suicidal Behavior, Spectrum of Assaultive Behavior, and Ego Defense Scale - are relevant to this study. The Spectrum of Suicidal Behavior section identifies the spectrum of suicidal behavior in the previous 6 months. This section is made up of six graded questions, ranging from an absence of any suicidal thoughts or actions (graded 1) to a serious suicide attempt resulting in death (graded 6). The intermediate grades of the index are as follows: suicidal ideation (2); suicide threats (3); a mild suicide attempt (4); and a severe suicide attempt (5). The score is determined according to the highest grade. The Spectrum of Assaultive Behavior section is also composed of six graded questions which deal with a spectrum of behaviors indicative of the child's level of violence, ranging from an absence of any assaultive behavior (score of 1) at one end to homicide (score of 6) at the other end. The intermediate assaultive behaviors include the following: assaultive ideation (2); assaultive threats (3); mild assaultive action (4); and serious assaultive action (5). Here, too, the score is determined on the basis of the highest level of violence. In the Ego Defense Scale (EDS), use of the following 11 defense mechanisms is examined: regression, denial, projection, introjection, reaction formation, undoing, displacement, intellectualization, compensation, sublimation, and repression. The degree of use of each mechanism ranges from 1 ("never") to 3 ("often") and is determined for each mechanism by the child's answers to a number of questions relevant to that mechanism. The total score for this section represents the sum of all the mechanisms ranging from minimum use (11) to maximal use (33). The CSPS was translated into Hebrew and back-translated into English. This procedure was repeated until complete concordance between translations was achieved. Certain items were modified for use in an Israeli and in an adolescent population, without changing the significance of the items for the particular defense mechanism as judged by three senior clinicians not connected with the project. In a study of a subset of 50 adolescent inpatients, two child psychiatry fellows attained the following interjudge reliabilities for the 11 defenses measured (the parentheses indicate the reliabilities reported by Pfeifer et al., 1984): intellectualization r = .93 (.96); regression r = .90 (.87); undoing r = .88 (.85); reaction formation r = .85 (.82); compensation r = .78 (.74); repression r = .79 (.74); denial r = .80 (.72); projection r = .72 (.69); introjection r = .67 (.57); sublimation r = .70 (.57); and displacement r = .57 (.52). The discriminative validity of the scale was proven in several clinical studies showing that the EDS distinguishes between various groups of adolescent inpatients (Gothelf et al., 1995) and between suicidal children and various control groups (Pfeffer et al., 1984, 1995). In addition, Pfeifer et al. (1995) showed that defense mechanism profiles are relatively stable along a 6- to 10-year follow-up period. Two different instruments were used to measure defense mechanisms because it has been shown that ego defenses are less reliable to evaluate (Vaillant, 1992). Thus using two different scales provides validation of discrimination. Procedure Patients. The patients were interviewed between the first and second week of admission. DSM-III-R diagnoses were made by a senior child psychiatrist, using the Schedule for Affective Disorders and Schizophrenia for School-Age Children (K-SADS), which has been shown to be reliable and valid in this population (Apter et al., 1989a). Personality disorders were diagnosed on the basis of the K-SADS as well as extensive open evaluation as there is no good valid instrument in Hebrew for diagnosing adolescent personality disorders. The CSPS was measured by research clinicians (child psychiatry fellows) who had been trained to achieve reliabilities for all items with the senior author at a level of at least [Kappa] = .8. The LSI was administered by a research assistant. Average examination time was 2 hours. Healthy Controls. Volunteers corresponding to the socioeconomic characteristics of the unit were located by a polling company. A senior clinical psychologist (R.O.) introduced himself in the following fashion: "You have agreed to participate in a psychological study. Please fill in the questionnaires anonymously." The CSPS interview and the LSI report form were administered by R.O. The administration of the questionnaires to the control group took approximately 6 months. All volunteers and their parents signed informed consent forms. No attempt was made to diagnose psychopathology in the control group because there is evidence to show that "super-normal" controls may give misleading results (Gothelf et al., 1995; Schwartz and Link, 1989). Data Analysis One-way analysis of variance (ANOVA) with multiple comparisons (t tests) with a Bonferroni correction for multiple [TABULAR DATA FOR TABLE 2 OMITTED] comparisons was used to analyze the data. When this was used an actual p value of .05 divided by 3, i.e., .0167 was used but the quoted p value was .05. The predictor variables were the suicidal and nonsuicidal groups of patients and the nonpatient adolescent group. The outcome variables were mean scores of each defense mechanism on the LSI and EDS. Pearson Product-Moment Correlation Coefficients were used to correlate the defense mechanism scores on each of the two scales (LSI and EDS) with one another and with the scores of two measures, the Spectrum of Assaultive Behavior and the Spectrum of Suicidal Behavior. Two-way ANOVA was performed to establish the effect of sex and diagnosis on the differences between defenses in suicidal and nonsuicidal adolescents. RESULTS Seven identical defense mechanism and total defense scores are common to the LSI and the EDS. Correlation coefficients between the defense scores on the LSI and EDS were calculated for the entire sample population. Significant (p [less than] .001) positive correlations were found only for repression (r = .25), regression (r = .29), and total defense scores (r = .27). Tables 2 and 3 show the comparison of the means and standard deviations on the LSI and EDS, respectively, for all research groups. On ANOVA, the LSI's compensation, denial, displacement, regression, repression, and total scores and CSPS's regression, denial, projection, introjection, reaction formation, sublimation, repression, and total scores significantly (p [less than] .05) distinguished between the study groups.

There seems to be a great deal of support for the first part of our first hypothesis. Suicidal adolescent patients had significantly higher scores on several defense mechanisms than the nonpatient adolescents. On the LSI, suicidal patients used denial, displacement, repression, and total defenses significantly more frequently (p [less than] .05) than the nonpatients. Similarly, on the EDS, suicidal patients had significantly higher scores (p [less than] .05) on regression, denial, projection, introjection, repression, and total defenses. Sublimation scores on the EDS were significantly lower (p [less than] .05) in the suicidal group of patients than in the nonpatient group. There was only partial support regarding the second part of our first hypothesis. On the LSI, scores on two defense mechanisms (compensation lower and displacement higher) significantly (p [less than] .05) distinguished the suicidal from the nonsuicidal groups of adolescent patients. However, no defense mechanisms distinguishing suicidal from nonsuicidal patients were identified on the EDS. To test the second hypothesis, that defenses correlated with suicidal behavior are also correlated with violence toward others, we made use of the Spectrum of [TABULAR DATA FOR TABLE 3 OMITTED] Assaultive Behavior scores and the Spectrum of Suicidal Behavior scores from the CSPS. Since these are general measures which quantify the suicidal and assaultive behavior of each subject, including the control group subjects (nonpatients and nonsuicidal inpatients), the research hypothesis was tested for the entire sample population. The correlation between the violence and suicide indices was as expected significant and positive (r = .28, n = 214, p [less than] .001). Correlations between suicidal behavior, violent behavior, and the various defense mechanism scores on both measures are presented in Table 4. Several defense mechanisms showed significant correlations with one of the parameters (suicide or violence) but not with the other: compensation (on the LSI) and intellectualization (on the EDS) correlated with violent behavior but not with suicidal behavior. Repression (LSI) and introjection (EDS) correlated with suicidal behavior but not with violent behavior. A number of defense mechanisms yielded significant correlations with both parameters: denial (on both measures), displacement (LSI), sublimation (inversely), regression, projection, and repression. There were significantly more affective disorder and borderline personality disorder patients in the suicidal group and significantly more adolescents with schizophrenia, eating disorders, and conduct disorders in the nonsuicidal group (p [less than] .05). Using two-way ANOVA (suicidal patients versus nonsuicidal patients and diagnosis), diagnosis only contributed to the difference as regards introjection (p [less than] .05). Since even on this defense the difference between suicidal and nonsuicidal groups reached a significance level of p [less than] .001, we can conclude that diagnosis was not a confounding factor in our results. Repression was more common in nonsuicidal females than in nonsuicidal males and sublimation and total defenses were higher for females than males. However, two-way ANOVA showed no effect of gender on the differences in defenses between the suicidal and nonsuicidal youngsters. In addition, gender did not have any significant effect on the correlation between defense mechanisms and suicidal behavior. TABLE 4 Pearson Product-Moment Correlation Coefficients Between Suicidal Behavior, Violent Behavior, and Defense Mechanism Measures Suicidal Violent Defense Mechanism Behavior Behavior LSI's defense mechanisms Compensation -.07 .18(**) Denial .24(**) .17(**) Displacement .30(**) .18(**) Repression .17(**) .09 EDS's defense mechanisms Regression .16(*) .25(**) Denial .24(**) .30(**) Projection .20(**) .23(**) Introjection .26(**) .11 Intellectualization .03 .16(*) Sublimation -.16(*) -. 18(*) Repression .23(**) .18(**) Note: LSI = Life Style Index; EDS = Ego Defense Scale. * p [less than] .05; (**)p [less than] .01. There were, however, interesting effects of gender on the correlation between defense mechanisms and violence. Girls showed significant positive correlations between violent behavior and denial (LSI), regression, projection, and repression (EDS) and a significant negative correlation with sublimation (p [less than] .05). In boys the significant correlations were with compensation and displacement of the LSI and intellectualization of the EDS (negative correlation).

DISCUSSION The correlations between the two defense mechanism scale scores are generally quite poor (Table 2), which means that the concurrent validity of the scales is low. McCullough (1992) noted that the problem of achieving reliable and valid identification of defenses stems from the fact that they are abstract intrapsychic constructs. Accordingly, the specific behaviors that identify these mental dispositions are heterogeneous, and no single behavior is either necessary or sufficient to identify a defense. Furthermore, different raters view and code defenses from different perspectives so that the same behavior or statement could be interpreted as representing one type of defense or another or as not being defensive at all. Nonetheless it can be seen that our raters were able to achieve reasonable reliability. The two scales used, the LSI and the EDS, are based on different approaches of measuring. The LSI, being a self-report questionnaire, is reliable and relatively easy to administer and score. However, it depends on the subject's self-knowledge and truthfulness. The EDS, being based on observer assessment gathered from various sources of information (the adolescent, parents, and therapists), is thus more valid. Unlike the LSI, which is composed solely of items relating to feelings and attitudes, the EDS also relies on overt behaviors of the subject. Therefore, it seems that the two scales are complementary to each other, and both should be administered in order to achieve a more reliable and valid assessment of the use of defense mechanisms. Of the items that differentiated the study groups, it may be possible to assign more importance to those defenses derived from both instruments. Denial, repression, and total defenses were higher in suicidal patients compared with normal subjects on both instruments. The inpatients used several defenses more frequently than the nonpatients, replicating other studies using the LSI (Conte and Apter, 1995; Plutchik et al., 1979) and EDS (Gothelf et al., 1995; Pfeifer et al., 1995). Thus, total defense scores correlate with psychopathology in general, lending support to the notion of nonspecific psychological distress as conceptualized by Dohrenwend et al. (1980). There are numerous psychological, physiological, and biochemical measures that correlate with nonspecific psychological distress, i.e., they are associated with major psychiatric disorders in general but not with any specific one. If this is true, then defense mechanisms measurement may be useful for identifying individuals in psychological distress. This finding may also be conceptualized as seeing defense mechanisms being activated by any kind of emotional upheaval. The results of this study, together with the results of previous studies (e.g., Gothelf et al., 1995; Pfeifer et al., 1995), indicate that psychiatric inpatients have many common defense mechanisms, and the strength of this result may explain the less significant findings when different psychopathologies (Gothelf et al., 1995) or suicidal and nonsuicidal patients are compared. The defenses that differentiated suicidal inpatients from nonsuicidal inpatients were LSI's high displacement and low compensation scores. There are not enough data in our study to explain this finding; however, we would like to suggest possible hypotheses which could be tested in future research. Displacement can be defined in various ways, such as: ". . . the individual copes with an emotional conflict or with an internal stressor, by generalizing or redirecting his emotions to a different, usually less threatening object" (Vaillant, 1986, p. 104) or ". . . the purposeful, unconscious shifting from one object to another in the interest of solving a conflict" (p. 108) or ". . . the redirection of emotions towards a relatively less important object . . ." (p. 115). "Displacement means expression of an affect, impulse or action towards a person, object or body part, which was not the one that originally evoked that affect or emotion. The affect is completely known, but is now aimed at a less conflictual target" (Vaillant, 1986, p. 131). These similar definitions of displacement have several elements in common. First, there is no change in the unpleasant emotion. The emotion or impulse retains its crude original quality. Second, the emotion is shifted from the original object to a new object. Any object can be included in this category, including the self. The redirection of the unpleasant emotion toward the self represents a particular case of displacement. The importance of this finding lies in the identification of a mode of behavior characteristic of the potential suicidal patient. There seems to be an inability to cope with objects perceived as invulnerable or potentially dangerous. Although displacement is not usually considered an immature or maladaptive defense mechanism but rather "intermediate" or "neurotic" (Vaillant and Drake, 1985), the results of this study are in line with those of Conte et al. (1983), who found that overuse of displacement predicted rehospitalizations of psychiatric patients. The findings suggest that the empirically derived concept of displacement represents a more maladaptive ego defense than previously considered. Recklitis et al. (1992) emphasized the mechanism of turning against the self as related to suicide. If suicidal behavior is a particular case of displacement, then the mechanism of turning against the self could be conceptualized as being derived from the broader mechanism of displacement. This is a somewhat similar idea to the notions of Cantor (1983) and Maltsberger (1986), who describe suicide as an intrapunitive response to conflict with internalized aggression resulting in negative self-object representation. The second finding that differentiates suicidal patients from nonsuicidal patients is the significantly less frequent use of the mechanism of compensation by suicidal individuals. Pfeffer (1986) defines compensation as a mechanism that is activated when a child feels he or she lacks ability or does not fit in. This leads to unconscious overdevelopment in another area. The mechanism does not appear in the classic lists of mechanisms, and its main distinction is that it is a dynamic mechanism whose characteristics accompany the child from childhood to adulthood. It is adaptive and it is not pathological in essence. The relative lack of compensation in the suicidal inpatients may indicate a deficit in the developing ego's ability to integrate the stresses and strains of adolescent development. Nonetheless, since compensation is a self-report measure with relatively low internal reliability (.59; .43), this result should be regarded with caution. The second hypothesis of the study was confirmed: several defense mechanisms correlated with both violent and suicidal behavior, while other defenses correlated only with violence or only with suicidal behavior. In terms of the two-stage theory of Plutchik and van Praag (1989), repression as measured by the EDS, denial, displacement, regression and projection, which are correlated with both violent and suicidal behavior, may be regarded as general augmentors of aggression. Sublimation, which is correlated negatively and significantly with both violence and suicidality, may be regarded as an attenuator of aggression. Compensation and intellectualization correlate significantly and positively with violence only. Thus they may be regarded as turning aggression outward in the second stage of the model, while repression (as measured by the LSI) and introjection would appear to correlate with inwardly expressed hostility. It is not clear why intellectualization and compensation, which are relatively high-level defenses, should be related to turning outward of aggression, although in the classic psychoanalytic literature the ability to express both aggression and libido outward instead of inward is regarded as a sign of increased differentiation and of more advanced levels of object relations. In conclusion, it must be stated that it is extremely difficult to reduce complex concepts such as defense mechanisms to measurable and empirical concepts, and the defects (such as lack of concurrent validity) in the instruments we have used must be taken into account when assessing our findings. Nonetheless many clinicians use concepts of ego defense structure as a way of organizing clinical data and as a prelude to planning treatment. Thus empirical research in this field is of great importance and we hope that this study, along with others, will provide more impetus for the joining of forces between dynamically oriented clinicians and empirical researchers. Clinical Implications Despite its heuristic value in suicide prevention, not much empirical research has been conducted on the ego functions that characterize suicidal adolescents. In this study we focused on one major aspect of these ego functions, that is, ego mechanisms of defense. Many nonspecific ego defenses are overused by suicidal adolescents. In addition, we found that frequent use of displacement and underuse of compensation are relatively specific to suicidal adolescents. Several immature ego defenses (denial, projection, and regression) possibly amplify aggression, which then is directed inward, against the self, by the realadaptive overuse of introjection, displacement, and repression. The replication of these findings in further studies could assist clinicians in identifying adolescents at risk for suicide. In terms of psychotherapy, the study possibly indicates that suicidality may represent maladaptive use of ego defenses such as displacement. Conversely, the mature ego defenses, sublimation and compensation, are possible protective factors. Prevention and outcome may be enhanced by helping adolescents at risk for suicidal behavior to practice effective coping skills. Knowledge of defense profiles may help clinicians develop coping strategies with patients and may help predict outcome. Thus patients who overuse displacement and introjection and underuse compensation may be more likely to make suicide attempts - a hypothesis which needs to be tested prospectively. Whether psychotherapy could reduce the use of ego defenses that augment aggressive and suicidal impulses and foster ego defenses that attenuate these impulses is a question which requires further research. REFERENCES Abraham K (1957), Notes on manic-depressive insanity. In: Selected Papers on Psychoanalysis, Vol 13, Jones E, ed. London: Hogarth Press, pp 137-156 Apter A, Gothelf D, Orbach I et al. (1995), Correlation of suicidal and violent behavior in different diagnostic categories in hospitalized adolescent patients. J Am Acad Child Adolesc Psychiatry 34:912-918 Apter A, Orvaschel H, Laseg M, Moses T, Tyano S (1989a), Psychometric properties of the K-SADS-P in an Israeli adolescent inpatient population. J Am Acad Child Adolesc Psychiatry 28:61-65 Apter A, Plutchik R, Sevy S, Korn M, Brown S, van Praag HM (1989b), Defense mechanisms in risk of suicide and risk of violence. Am J Psychiatry 146:1027-1031 Cantor P (1983), Depression and suicide in children. In: Handbook of Clinical Child Psychology, Walker CE, Roberts MC, eds. New York: Wiley, pp 453-474 Conte HR, Apter A (1995), The Life Style Index: a self report measure of ego defenses. In: Ego Defenses: Theory and Measurement, Conte HR, Plutchik R, eds. New York: Wiley, pp 179-201 Conte HR, Plutchik R, Schwartz B, Wild K (1983), Psychodynamic variables related to outcome in hospitalized schizophrenics. Paper presented at the Convention of the American Psychological Association, Anaheim, CA Dohrenwend BP, Shrout PE, Egri G, Mendelsohn FS (1980), Nonspecific psychological distress and other dimensions of psychopathology measures for use in the general population. Arch Gen Psychiatry 37:1229-1236 Gothelf D, Apter A, Ratzoni Get al. (1995), Defense mechanisms in severe adolescent anorexia nervosa. J Am Acad Child Adolesc Psychiatry 34:1648-1654 Maltsberger JT (1986), Suicide Risk: The Formulation of Clinical Judgment. New York: University Press McCullough L (1992), Toward reliability in identifying ego defenses: clinical techniques. In: Ego Mechanisms of Defense: A Guide for Clinicians and Researchers, Vaillant GE, ed. Washington, DC: American Psychiatric Press, pp 171-179 Pfeifer CR (1986), The Suicidal Child. New York: Basic Books Pfeifer CR, Conte HR, Plutchik R, Jerett I (1979), Suicidal behavior in latency-age children: an empirical study. J Am Acad Child Psychiatry 18:679-692 Pfeffer CR, Conte HR, Plutchik R, Jerett I (1980), Suicidal behavior in latency-age children: an outpatient population. J Am Acad Child Psychiatry 19:703-710 Pfeffer CR, Hurt SW, Peskin JR, Siefker CA (1995), Suicidal children grow up: ego functions associated with suicide attempts. J Am Acad Child Adolesc Psychiatry 34:1318-1325 Pfeffer CR, Plutchik R, Mizruchi MS (1983), Suicidal and assaultive behavior in children: classification, measurement, and interrelations. Am J Psychiatry 140:154-157 Pfeffer CR, Solomon G, Plutchik R, Mizruchi MS, Weiner A (1982), Suicidal behavior in latency age psychiatric inpatients. J Am Acad Child Psychiatry 21:564-569 Pfeffer CR, Zuckerman S, Plutchik R, Mizruchi MS (1984), Suicidal behavior in normal school children: a comparison with child psychiatric inpatients. J Am Acad Child Psychiatry 23:416-423 Plutchik R, Kellerman H, Conte HR (1979), A structural theory of ego defenses and emotions. In: Emotions in Personality and Psychopathology, Izard CE, ed. New York: Plenum, pp 229-257 Plutchik R, van Praag HM (1989), Correlates of suicide and violence risk, III: a two-stage model of countervailing forces. Psychiatry Res 28:215-225 Recklitis CJ, Noam GG, Borst SR (1992), Adolescent suicide and defensive style. Suicide Life Threat Behav 22:374-387 Runeson B (1989), Mental disorders in youth suicide: DSM-III-R Axes I and II. Acta Psychiatr Scand 79:490-497 Sandler J, Joffe WG (1965), Notes on child depression. Int J Psychoanal 46:88-89 Schwartz S, Link BG (1989), The "well control" artefact in case/control studies of specific psychiatric disorders. Psychol Med 19:737-742 Spirito A, Brown L, Overholzer J, Fritz G (1989), Attempted suicide in adolescence: a review and critique of the literature. Clin Psychol Rev 9:335-363 Vaillant GE (1986), Appendix: six assessment schemes for defense mechanisms. In: Empirical Study of Ego Mechanisms of Defense, Vaillant GE, ed. Washington, DC: American Psychiatric Press, pp 101-152 Vaillant GE, ed (1992), Ego Mechanisms of Defense: A Guide for Clinicians and Researchers. Washington, DC: American Psychiatric Press Vaillant GE, Drake RE (1985), Maturity of ego defenses in relation to DSM-III, Axis II personality disorder. Arch Gen Psychiatry 42:597-601 Drs. Apter, Gothelf, Ratzoni, and Tyano are with the Department of Child and Adolescent Psychiatry, Geha Psychiatric Hospital and Sackler School of Medicine, University of Tel Aviv, Israel. Dr. Offer is a psychologist at the Shaphir Institute, Tel Aviv. Dr. Orbach is with the Department of Psychology, Bar Ilan University, Ramat Gan, Israel. Dr. Pfeffer is with the Department of Psychiatry, New York Hospital-Cornell Medical Center.