Articles: Contagion of deliberate self-harm among adolescent inpatients.
By Tero J. Taiminen
Issue: February 1998, Volume 37, Issue 2
Page Number: 211
Authors: Tero J. Taiminen, MD, PhD; Kristiina Kallio-Soukainen, MD; Hannele Nokso-Koivisto, MD; Anne Kaljonen, MSc; Hans Helenius, MSc
Self-mutilation is usually defined as the deliberate destruction of body tissue without a conscious intent to die (Herpetz, 1995). However, it is often difficult to determine whether some forms of deliberate self-harm (DSH), such as wrist cutting, are better understood as self-mutilation or parasuicidal behavior. Therefore, in the present article, we use the term DSH to refer all self-injurious behavior, both with and without suicidal tendencies.
Major self-mutilation includes eye enucleation and amputation of limbs or genitals, and it is usually associated with psychotic states or severe gender identity disturbances (Favazza, 1989; Pies and Popli, 1995). Minor self-mutilation includes skin cutting and burning, venipuncture and blood letting, self-hitting and scratching, hair pulling, and bone breaking (Favazza, 1989; Pies and Popli, 1995). Minor self-mutilation is relatively common, with a prevalence ranging from 400 to 1,400 per 100,000 (Favazza and Rosenthal, 1993). It is encountered predominantly in four groups of patients: in those with borderline personality disorder, in those with mental retardation and other organic conditions, in those with eating disorders, and in incarcerated men with antisocial personality disorder (Favazza, 1989; Herpetz, 1995; Hillbrand et al., 1994; Pies and Popli, 1995; Simeon et al., 1992).
In patients with borderline personality disorder, self-mutilation typically begins in adolescence and may persist for decades (Favazza and Conterio, 1989; Simeon et al., 1992). Although the psychological motivations for the first acts of self-mutilation in adolescence are heterogeneous, self-mutilation and the act of letting of blood may quickly become a regular method for relieving anxiety or anger (Favazza, 1989; Solomon and Farrand, 1996). Favazza and Conterio (1989) even found that the majority of female habitual self-mutilators considered their self-injurious behavior to be an addiction. It has been suggested that self-injurious behavior stimulates the production and release of endorphins, which may eventually lead to addiction (Coid et al., 1983; Pies and Popli, 1995).
Davidson et al. (1989) discussed two types of exposure to suicide contagion among teenagers: direct and indirect. They called the exposure direct if the object of imitation was actually known, and indirect if the suicide was known to the person only through news accounts or word of mouth. We suggest that this division into direct and indirect exposure is applicable to contagion of all types of DSH.
Contagion of both completed suicide and parasuicidal behavior has been documented among adolescents both in macro-level surveys concerning indirect exposure (Davidson et al., 1989; Gould and Shaffer, 1986; Phillips and Carstensen, 1986; Schmidtke and Hafner, 1988) and in micro-level studies, which have considered both direct and indirect exposure (Brent et al., 1989; Kaminer, 1986; Robbins and Conroy, 1983; Vaisanen and Hagglund, 1981). However, negative results have also been reported (Brent et al., 1996; Simkin et al., 1995). Psychiatric inpatients with personality disorders and schizophrenia seem to be particularly susceptible to suicide contagion (Rissmiller and Rissmiller, 1990; Taiminen et al., 1992). It is probably weak ego and diffuse identity that make both adolescents and psychiatric inpatients susceptible to various forms of identification (Sacks and Eth, 1981; Taiminen et al., 1992). Taiminen (1992) suggested that as a part of the empathic process, an adolescent may project his or her best qualities onto the internal representation of the suicide victim. Thereafter the adolescent may identify himself or herself with this representation, and with suicidality as well. Against this background it is quite surprising that suicide contagion seems not to be a frequent problem in psychiatric units for adolescents (King et al., 1995).
[pagebreak]
Compared with contagion of completed suicides, contagion of self-mutilation among adolescents has been much less studied. To our knowledge, there are no macro-level studies on contagion of self-mutilation. At the micro-level, there are some descriptions of contagious parasuicidal behavior and self-mutilation due to direct exposure within adolescent treatment programs (Crabtree and Grossman, 1974; Ghaziuddin et al., 1992; Matthews, 1968; Offer and Barglow, 1960; Rosen and Walsh, 1989; Walsh and Rosen, 1985). Offer and Barglow (1960) noted that self-mutilators were often group leaders whose behavior was widely imitated, and self-mutilation became a learned form of acting-out behavior with a ritualistic exhibitionistic character. Ross and McKay (1979) reported that some self-mutilating acts in a training school for delinquent girls occurred in a context of an initiation rite. Rosen and Walsh (1989) came to the conclusion that contagious self-mutilation may be viewed as a concrete display of affinity between two people. Adolescents appeared to use self-mutilation to communicate feelings and to ensure a tight bond within a relationship (Rosen and Walsh, 1989).
Studies on the quantitative importance of contagious self-mutilation among adolescent inpatients have yielded contradictory results. King et al. (1995) did not find any clustering of self-mutilation among 57 adolescent inpatients. However, two studies have found that self-mutilation occurred in statistically significant clusters in a long-term treatment facility for seriously disturbed adolescents (Rosen and Walsh, 1989; Walsh and Rosen, 1985). This is an interesting finding, because though some individual contagion incidents of completed adult inpatient suicides have been described, systematic and statistically significant temporal clustering of either completed inpatient suicides or inpatient parasuicides has not been demonstrated (King et al., 1995; Modestin and Wurmle, 1989; Taiminen and Helenius, 1994). This indicates that self-mutilation may be even more contagious than suicidal behavior among adolescent inpatients. This hypothesis is directly supported by the finding of Walsh and Rosen (1985) that, during their 1-year observation period, suicidal talks were not clustered in time across subjects though self-mutilative acts were.
The aims of the present study were (1) to determine the quantitative importance of contagion of DSH among adolescent psychiatric inpatients with statistical methods, and (2) to clarify the psychological mechanisms behind contagion by interviewing those subjects who had been involved in DSH contagion incidents.
METHOD
Setting
The setting for this study was an adolescent psychiatric ward of Turku University Central Hospital in southwestern Finland. This acute closed ward with 10 beds takes care of both voluntarily and involuntarily admitted adolescents between 12 and 18 years of age.
Subjects
For 12 months we collected data on all incidents of DSH. During this period 51 adolescents, 35 female and 16 male, were treated on the ward. During the 1-year data collection period, 64 acts of DSH were performed by 12 adolescents who are the subjects in the present study. The number of acts per subject ranged from 1 to 19 (mean 5.4, SD 5.4).
Of the 64 acts of DSH, 58 were skin cutting, 2 were drug overdoses, and 4 were self-starvation. With self-starvation episodes, only the first day of the episode was recorded as a day of DSH. All the acts of skin cutting were clearly minor self-mutilation in nature, but the drug overdoses and episodes of self-starvation can probably be better classified as parasuicides. There were no completed suicides, acts of major self-mutilation, or acts of minor self-mutilation other than skin cutting (e.g., hair pulling or self-inflicted burns) during the study period.
Clinical Interviews
Seven subjects were involved in four or more contagion incidents. We attempted to have these seven subjects interviewed by three clinicians (T.J.T., K.K.-.S, and H.N.-K.) in the peak month of DSH contagion. One subject refused to discuss any issues related to DSH, and the interviews were carried out for six subjects. In semistructured interviews we asked for information particularly on the following subjects: patient's personal history of DSH, patient's subjective affective states during the acts of DSH, patient's estimation of her own motivation for DSH, patient's impression why other adolescents on the ward had been involved in DSH, patient's relationships with other adolescents involved in DSH, and patient's judgment of the role of imitation or identification on acts of DSH on the ward.
Statistical Analyses
To analyze whether the acts of DSH occurred in a nonrandom fashion, we used a method previously described by Rosen and Walsh (1989). DSH contagion was defined as two or more acts of DSH that involved two or more adolescents and occurred on the same day or consecutive days.
The data were analyzed with 11 independent sequential [[Chi.].sup.2] analyses (Rosen and Walsh, 1989). The first [[Chi].sup.2] test examined the distribution of contagion events between the subject who was involved in the most contagion incidents and the other 11 subjects (1 x 11 contingency table). The next [[Chi].sup.2] test was conducted for the subject who was involved in the second highest number of contagion incidents, but the data for the first subject were omitted (1 x 10 contingency table). This procedure was followed for each succeeding subject, which resulted in 11 independent [[Chi].sup.2] tests. The obtained [[Chi].sup.2] values were then summed to obtain an overall value.
RESULTS
Demographic and Diagnostic Characteristics of Subjects With DSH
The 12 subjects involved in acts of DSH during the 12-month period were all female. Their mean age was 15.5 years (SD 1.7, range 12.3 to 17.9); 11 were of Caucasian and 1 of African race. Their most common DSM-IV diagnoses on Axis I were major depression (seven subjects) and impulse control disorders (three subjects). Nine subjects fulfilled the criteria for one or more personality disorders, borderline personality disorder being the most common (nine subjects). Their mean length of hospitalization during the study period was 91.9 days (SD 73.2, range 7 to 200), and their mean length of hospitalization before the first DSH act was 56.2 days (SD 56.3, range 1 to 179). Age, complete DSM-IV diagnoses, length of hospitalization during the 12-month period, length of hospitalization before the first DSH act, type and number of DSH events, and number of DSH contagion incidents of the 12 subjects are presented in Table 1. The subjects in Table 1 and Figure 1 are numbered according to the temporal order of their first act of DSH during the 12-month period, e.g., subject 1 committed the first act of DSH during the study period.
Contagion of DSH
There were 37 instances of contagion during the 12-month period. The number of contagion incidents per subject ranged from 0 to 17 (mean 6.2, SD 5.8), and 10 of 12 patients with DSH were involved in contagion episodes. Their mean length of hospitalization before the first DSH act was 55.1 days (SD 55.2, range 10 to 179). An overall value obtained from the 11 independent [[Chi].sup.2] tests was significant ([[Chi].sup.2] = 9.1, df = 55, p [less than] .05). Thus, DSH occurred in nonrandom fashion and incidents of DSH were not independent of each other during the 12-month period.
Instances of contagion are illustrated with a sociogram, presented in Figure 1, in which each line connecting two subjects represents an incident of contagion. Two subjects, 4 and 5, were not involved in episodes of contagion at all. Subjects 1, 2, and 3 were involved in the first episode of contagion. Subjects 6 and 8, subjects 7 and 8, and subjects 6 and 7 were then involved in dyadic interactions of contagion. At the time of these dyadic episodes, subjects 1, 2, and 3 were no longer on the ward. Finally, there were two larger clusters of DSH, the first one involving subjects 7, 9, and 10, and the second one subjects 7, 8, 9, 11, and 12.
In most episodes of contagion the type of DSH was self-mutilation by skin cutting of wrists or arm and blood letting. Three subjects, 7, 8, and 12, also stored up blood in bottles kept in their wardrobes. However, there were also a few episodes in which skin cutting triggered other types of DSH in other subjects such as parasuicidal behavior by drug overdose or self-starving.
[TABULAR DATA FOR TABLE 1 OMITTED]
Findings From the Clinical Interviews
Of the six subjects interviewed (7, 8, 9, 10, 11, and 12), two subjects, 8 and 11, were involved in DSH for the first time, whereas four others had committed self-mutilation before hospitalization. It is of interest that subjects 8 and 11 were also the youngest of the 12 who had been involved in acts of DSH.
Two subjects, 11 and 12, said that self-mutilation did not relieve their anxiety or anger at all. According to their own judgment, they were involved in DSH mainly because they did not want to feel themselves outsiders. Subjects 6, 7, 8, 9, and 10 found that self-mutilation instantly relieved anxiety and/or anger. Subjects 8 and 9 did not experience pain while injuring themselves.
Only in subject 10 was self-injurious behavior probably not influenced by others, which is also supported by the fact that her two skin-cutting incidents preceded self-mutilation events by subjects 7 and 9. In five subjects DSH was clearly influenced by interaction with other inpatients. In subject 9 a dyadic idealizing interaction with subject 7 become evident. Subject 9 projected esteemed qualities onto subject 7, and thereafter she identified herself with them and with self-mutilative behavior as well.
Cardasis et al. (1997) found that patients with borderline personality disorder more often brought transitional objects into the psychiatric hospital than other patients. On the other hand, it has been suggested that the blood produced by self-cutting may serve as a comforting transitional object (Favazza, 1989). Subjects 7, 8, and 12, who had a peculiar habit of storing blood in their wardrobes, were all aware of the same custom in the other two subjects. We got the impression that only for subject 7 did blood storage serve as a transitional object, while for subjects 8 and 12, the main incentive for this habit was imitation of subject 7.
With the exception of subject 10, the contagion incidents were associated with group dynamics. Skin cutting and blood letting had become an initiation rite that strengthened group cohesion. However, it seems that a prerequisite for the feeling of togetherness was a shared emotional experience of relief after an event of DSH. Subject 11, who did not feel any relief with self-mutilation, was openly treated with contempt by subjects 6, 7, 8, and 9 and was labeled by them as a fake or a pretender.
DISCUSSION
Contagion of DSH
The present study replicates the findings from two previous studies carried out by Walsh and Rosen (Rosen and Walsh, 1989; Walsh and Rosen, 1985), that microlevel contagion of DSH occurs in psychiatric units for adolescents. Moreover, our results suggest that a majority of DSH incidents on an adolescent psychiatric ward may have been influenced by contagion. The present study also demonstrates that one type of DSH behavior, e.g., self-mutilation by skin cutting, may trigger another type of DSH behavior, e.g., parasuicide by drug overdose, in another individual.
Characteristics of Adolescents Who Are at Risk for DSH Contagion
Nine of the 12 female subjects with DSH incidents fulfilled the diagnostic criteria for borderline personality disorder. This is in line with previous studies, which have found that female patients with borderline personality disorder are predisposed both to self-mutilation and to imitative behavior (Favazza, 1989; Herpetz, 1995; Phillips and Carstensen, 1986).
Two of the six subjects interviewed started their DSH behavior on the ward. They were the two youngest patients on the ward, both with borderline personality disorder. The finding that DSH can spread on a psychiatric ward to previously DSH-naive adolescents is particularly alarming when we keep in mind the assumption that self-injurious behavior may advance to an addiction (Cold et al., 1983; Pies and Popli, 1995).
Social Factors on the Closed Psychiatric Ward
It has been proposed that insecurity in the organization and the high number of inexperienced staff prepare the ground for suicide contagion in psychiatric hospitals (Taiminen et al., 1992), and this assumption may hold true also concerning contagion of other types of DSH. However, during the two larger clusters of DSH within the study period, the number of inexperienced staff on the ward was at a minimum and during the study there were no organizational changes threatening.
During the two larger clusters of DSH, the ward was slightly overcrowded and the majority of the inpatients were depressed female adolescents with borderline personality disorder. Jones and Daniels (1996) have noted that stressful events such as changes in room population often provoke self-injury in socially isolated primates. It is possible that social isolation and social stress created by the overpopulation of the closed ward, together with a patient population exceptionally vulnerable to imitation, exposed the ward to contagion incidents.
Offer and Barglow (1960) noted that most of the self-mutilation incidents of adolescent inpatients occurred in the evening and during weekends, when structured activities were at a minimum. Ross and McKay (1979) found that incidence of self-mutilation of adolescent girls increased when the frequency of acting-out behaviors decreased. In the present study, the mean length of hospitalization before the first DSH contagion incident was almost 2 months. However, the ward described in the study was developed mainly for acute treatment and psychiatric evaluation, and therefore its resources for structured group activities are limited. Temporary overpopulation of the ward probably contributed to the lack of structured activities for the patients.
Clinical Implications: What Can Be Done to Prevent DSH Contagion in Psychiatric Units for Adolescents?
First, we suggest that the concentration of many female adolescents with borderline personality disorder on the same ward should be avoided. If this is not possible, the length of hospitalization should be limited to 2 weeks. These patients are vulnerable to a self-injurious subculture, and contagious self-mutilation can later become a socially shameful addiction promoting suicidal behavior (Favazza and Conterio, 1989).
Second, it seems that the method of DSH is also prone to imitation, because the majority of the subjects with DSH used the same method, skin cutting of wrist or arm, sometimes with blood letting. Therefore, access to the method of DSH chosen by an inpatient should be made as difficult as possible for other patients.
Third, in one subject of the present study the model of projective identification, previously described in the context of suicide contagion (Taiminen, 1992), seemed helpful. Projective identification with an object with DSH can often be detected by staff members if they actively look for signs of the blurring of self and object representations. After detection, the staff can discuss the differences between the subject and the adolescent with DSH, and the process of projective identification can be smothered (Taiminen, 1992).
Fourth, our results suggest that acts of DSH functioned as rites of togetherness for small groups of female adolescents. It should be evaluated whether these harmful rites of DSH could be replaced by some other group activity with a rite-like character, e.g., bungee-jumping or mountaineering.
Fifth, we believe that the issue of contagion should be openly discussed with the adolescents involved in DSH contagion. Ross and McKay (1979) noted that when they discussed in a neutral way self-mutilative behavior of adolescent girls with them, the frequency of self-mutilation decreased. We also found that the structured interviews described in the present study functioned as a therapeutic intervention. Most adolescents avoid behavior that they themselves perceive as imitative. Therefore, direct clarification of the possible contagious aspects of DSH behavior may prevent an adolescent from being involved in DSH incidents in the future.
REFERENCES
Brent DA, Kerr MM, Goldstein C, Bozigar J, Wartella M, Allan MJ (1989), An outbreak of suicide and suicidal behavior in a high school. J Am Acad Child Adolesc Psychiatry 28:918-924
Brent DA, Moritz G, Bridge J, Perper J, Canobbio R (1996), Long-term impact of exposure to suicide: a three-year controlled follow-up. J Am Acad Child Adolesc Psychiatry 35:646-653
Cardasis W, Hochman JA, Silk KR (1997), Transitional objects and borderline personality disorder. Am J Psychiatry 154:250-255
Coid J, Allolio B, Rees LH (1983), Raised plasma metenkephalin in patients who habitually mutilate themselves. Lancet 2:545-546
Crabtree CH, Grossman WK (1974), Administrative clarity and redefinition for an open adolescent unit. Psychiatry 37:350-359
Davidson L, Rosenberg M, Mercy J, Franklin J, Simmons J (1989), An epidemiologic study of risk factors in two teenage suicide clusters. JAMA 17:2689-2692
Favazza AR (1989), Why patients mutilate themselves. Hosp Community Psychiatry 40:137-145
Favazza AR, Conterio K (1989), Female habitual self-mutilators. Acta Psychiatr Scand 79:283-289
Favazza AR, Rosenthal RJ (1993), Diagnostic issues in self-mutilation. Hosp Community Psychiatry 44:134-140
Ghaziuddin M, Tsai L, Ghaziuddin N (1992), Mood disorder in a group of self-cutting adolescents. Acta Paedopsychiatr 55:103-105
Gould MS, Shaffer D (1986), The impact of suicide in television movies: evidence of imitation. N Engl J Med 315:690-694
Herpetz S (1995), Self-injurious behaviour: psychopathological and nosological characteristics in subtypes of self-injurers. Acta Psychiatr Scand 91:57-68
Hillbrand M, Krystal JH, Sharpe KS, Foster HG (1994), Clinical predictors of self-mutilation in hospitalized forensic patients. J Nerv Ment Dis 182:9-13
Jones IH, Daniels BA (1996), An ethological approach to self-injury. Br J Psychiatry 169:263-267
Kaminer Y (1986), Suicidal behavior and contagion among hospitalized adolescents (letter). N Engl J Med 315:1030
King CA, Franzese R, Gargan S, McGovern L, Ghaziuddin N, Naylor MW (1995), Suicide contagion among adolescents during acute psychiatric hospitalization. Psychiatr Serv 46:915-918
Matthews PC (1968), Epidemic self-injury in an adolescent unit. Int J Soc Psychiatry 14:125-133
Modestin J, Wurmle O (1989), Role of modelling in in-patient suicide: a lack of supporting evidence. Br J Psychiatry 155:511-514
Offer D, Barglow P (1960), Adolescent and young adult self-mutilation incidents in a general psychiatric hospital. Arch Gen Psychiatry 3:194-204
Phillips D, Carstensen L (1986), Clustering of teenage suicides after television news stories about suicide. N Engl J Med 315:685-689
Pies RW, Popli AP (1995), Self-injurious behavior: pathophysiology and implications for treatment. J Clin Psychiatry 56:580-588
Rissmiller DJ, Rissmiller F (1990), Inpatient suicide epidemics and suggestion for prevention. Hosp Community Psychiatry 41:922-924
Robbins D, Conroy R (1983), A cluster of adolescent suicide attempts: is suicide contagious? J Adolesc Health Care 3:253-255
Rosen PM, Walsh BW (1989), Patterns of contagion in self-mutilation epidemics. Am J Psychiatry 146:656-658
Ross RR, McKay HB (1979), Self-Mutilation. Toronto: Lexington Books, DC Heath
Sacks M, Eth S (1981), Pathological identification as a cause of suicide on an inpatient unit. Hosp Community Psychiatry 32:36-40
Schmidtke A, Hafner H (1988), The Werther effect after television films: new evidence for an old hypothesis. Psychol Med 18:665-676
Simeon D, Stanley B, Frances A, Mann JJ, Winchel R, Stanley M (1992), Self-mutilation in personality disorders: psychological and biological correlates. Am J Psychiatry 149:221-226
Simkin S, Hawton K, Whitehead L, Fagg J, Eagle M (1995), Media influence on parasuicide: a study of the effects of a television drama portrayal of paracetamol self-poisoning. Br J Psychiatry 167:754-759
Solomon Y, Farrand J (1996), "Why don't you do it properly?" Young women who self-injure. J Adolesc 19:111-119
Taiminen T, Salmenpera T, Lehtinen K (1992), A suicide epidemic in a psychiatric hospital. Suicide Life Threat Behar 22:350-363
Taiminen TJ (1992), Projective identification and suicide contagion. Acta Psychiatr Scand 85:449-452
Taiminen TJ, Helenius H (1994), Suicide clustering in a psychiatric hospital with a history of a suicide epidemic: a quantitative study. Am J Psychiatry 151:1087-1088
Vaisanen EJ, Hagglund T-B (1981), Destruction preceding suicide: an episode in the large series of adolescents' suicides. In: Depression et Suicide. New York: Pergamon, pp 670-675
Walsh BW, Rosen P (1985), Self-mutilation and contagion: an empirical list. Am J Psychiatry 142:119-120
Page Number: 211
Authors: Tero J. Taiminen, MD, PhD; Kristiina Kallio-Soukainen, MD; Hannele Nokso-Koivisto, MD; Anne Kaljonen, MSc; Hans Helenius, MSc
Self-mutilation is usually defined as the deliberate destruction of body tissue without a conscious intent to die (Herpetz, 1995). However, it is often difficult to determine whether some forms of deliberate self-harm (DSH), such as wrist cutting, are better understood as self-mutilation or parasuicidal behavior. Therefore, in the present article, we use the term DSH to refer all self-injurious behavior, both with and without suicidal tendencies.
Major self-mutilation includes eye enucleation and amputation of limbs or genitals, and it is usually associated with psychotic states or severe gender identity disturbances (Favazza, 1989; Pies and Popli, 1995). Minor self-mutilation includes skin cutting and burning, venipuncture and blood letting, self-hitting and scratching, hair pulling, and bone breaking (Favazza, 1989; Pies and Popli, 1995). Minor self-mutilation is relatively common, with a prevalence ranging from 400 to 1,400 per 100,000 (Favazza and Rosenthal, 1993). It is encountered predominantly in four groups of patients: in those with borderline personality disorder, in those with mental retardation and other organic conditions, in those with eating disorders, and in incarcerated men with antisocial personality disorder (Favazza, 1989; Herpetz, 1995; Hillbrand et al., 1994; Pies and Popli, 1995; Simeon et al., 1992).
In patients with borderline personality disorder, self-mutilation typically begins in adolescence and may persist for decades (Favazza and Conterio, 1989; Simeon et al., 1992). Although the psychological motivations for the first acts of self-mutilation in adolescence are heterogeneous, self-mutilation and the act of letting of blood may quickly become a regular method for relieving anxiety or anger (Favazza, 1989; Solomon and Farrand, 1996). Favazza and Conterio (1989) even found that the majority of female habitual self-mutilators considered their self-injurious behavior to be an addiction. It has been suggested that self-injurious behavior stimulates the production and release of endorphins, which may eventually lead to addiction (Coid et al., 1983; Pies and Popli, 1995).
Davidson et al. (1989) discussed two types of exposure to suicide contagion among teenagers: direct and indirect. They called the exposure direct if the object of imitation was actually known, and indirect if the suicide was known to the person only through news accounts or word of mouth. We suggest that this division into direct and indirect exposure is applicable to contagion of all types of DSH.
Contagion of both completed suicide and parasuicidal behavior has been documented among adolescents both in macro-level surveys concerning indirect exposure (Davidson et al., 1989; Gould and Shaffer, 1986; Phillips and Carstensen, 1986; Schmidtke and Hafner, 1988) and in micro-level studies, which have considered both direct and indirect exposure (Brent et al., 1989; Kaminer, 1986; Robbins and Conroy, 1983; Vaisanen and Hagglund, 1981). However, negative results have also been reported (Brent et al., 1996; Simkin et al., 1995). Psychiatric inpatients with personality disorders and schizophrenia seem to be particularly susceptible to suicide contagion (Rissmiller and Rissmiller, 1990; Taiminen et al., 1992). It is probably weak ego and diffuse identity that make both adolescents and psychiatric inpatients susceptible to various forms of identification (Sacks and Eth, 1981; Taiminen et al., 1992). Taiminen (1992) suggested that as a part of the empathic process, an adolescent may project his or her best qualities onto the internal representation of the suicide victim. Thereafter the adolescent may identify himself or herself with this representation, and with suicidality as well. Against this background it is quite surprising that suicide contagion seems not to be a frequent problem in psychiatric units for adolescents (King et al., 1995).
[pagebreak]
Compared with contagion of completed suicides, contagion of self-mutilation among adolescents has been much less studied. To our knowledge, there are no macro-level studies on contagion of self-mutilation. At the micro-level, there are some descriptions of contagious parasuicidal behavior and self-mutilation due to direct exposure within adolescent treatment programs (Crabtree and Grossman, 1974; Ghaziuddin et al., 1992; Matthews, 1968; Offer and Barglow, 1960; Rosen and Walsh, 1989; Walsh and Rosen, 1985). Offer and Barglow (1960) noted that self-mutilators were often group leaders whose behavior was widely imitated, and self-mutilation became a learned form of acting-out behavior with a ritualistic exhibitionistic character. Ross and McKay (1979) reported that some self-mutilating acts in a training school for delinquent girls occurred in a context of an initiation rite. Rosen and Walsh (1989) came to the conclusion that contagious self-mutilation may be viewed as a concrete display of affinity between two people. Adolescents appeared to use self-mutilation to communicate feelings and to ensure a tight bond within a relationship (Rosen and Walsh, 1989).
Studies on the quantitative importance of contagious self-mutilation among adolescent inpatients have yielded contradictory results. King et al. (1995) did not find any clustering of self-mutilation among 57 adolescent inpatients. However, two studies have found that self-mutilation occurred in statistically significant clusters in a long-term treatment facility for seriously disturbed adolescents (Rosen and Walsh, 1989; Walsh and Rosen, 1985). This is an interesting finding, because though some individual contagion incidents of completed adult inpatient suicides have been described, systematic and statistically significant temporal clustering of either completed inpatient suicides or inpatient parasuicides has not been demonstrated (King et al., 1995; Modestin and Wurmle, 1989; Taiminen and Helenius, 1994). This indicates that self-mutilation may be even more contagious than suicidal behavior among adolescent inpatients. This hypothesis is directly supported by the finding of Walsh and Rosen (1985) that, during their 1-year observation period, suicidal talks were not clustered in time across subjects though self-mutilative acts were.
The aims of the present study were (1) to determine the quantitative importance of contagion of DSH among adolescent psychiatric inpatients with statistical methods, and (2) to clarify the psychological mechanisms behind contagion by interviewing those subjects who had been involved in DSH contagion incidents.
METHOD
Setting
The setting for this study was an adolescent psychiatric ward of Turku University Central Hospital in southwestern Finland. This acute closed ward with 10 beds takes care of both voluntarily and involuntarily admitted adolescents between 12 and 18 years of age.
Subjects
For 12 months we collected data on all incidents of DSH. During this period 51 adolescents, 35 female and 16 male, were treated on the ward. During the 1-year data collection period, 64 acts of DSH were performed by 12 adolescents who are the subjects in the present study. The number of acts per subject ranged from 1 to 19 (mean 5.4, SD 5.4).
Of the 64 acts of DSH, 58 were skin cutting, 2 were drug overdoses, and 4 were self-starvation. With self-starvation episodes, only the first day of the episode was recorded as a day of DSH. All the acts of skin cutting were clearly minor self-mutilation in nature, but the drug overdoses and episodes of self-starvation can probably be better classified as parasuicides. There were no completed suicides, acts of major self-mutilation, or acts of minor self-mutilation other than skin cutting (e.g., hair pulling or self-inflicted burns) during the study period.
Clinical Interviews
Seven subjects were involved in four or more contagion incidents. We attempted to have these seven subjects interviewed by three clinicians (T.J.T., K.K.-.S, and H.N.-K.) in the peak month of DSH contagion. One subject refused to discuss any issues related to DSH, and the interviews were carried out for six subjects. In semistructured interviews we asked for information particularly on the following subjects: patient's personal history of DSH, patient's subjective affective states during the acts of DSH, patient's estimation of her own motivation for DSH, patient's impression why other adolescents on the ward had been involved in DSH, patient's relationships with other adolescents involved in DSH, and patient's judgment of the role of imitation or identification on acts of DSH on the ward.
Statistical Analyses
To analyze whether the acts of DSH occurred in a nonrandom fashion, we used a method previously described by Rosen and Walsh (1989). DSH contagion was defined as two or more acts of DSH that involved two or more adolescents and occurred on the same day or consecutive days.
The data were analyzed with 11 independent sequential [[Chi.].sup.2] analyses (Rosen and Walsh, 1989). The first [[Chi].sup.2] test examined the distribution of contagion events between the subject who was involved in the most contagion incidents and the other 11 subjects (1 x 11 contingency table). The next [[Chi].sup.2] test was conducted for the subject who was involved in the second highest number of contagion incidents, but the data for the first subject were omitted (1 x 10 contingency table). This procedure was followed for each succeeding subject, which resulted in 11 independent [[Chi].sup.2] tests. The obtained [[Chi].sup.2] values were then summed to obtain an overall value.
RESULTS
Demographic and Diagnostic Characteristics of Subjects With DSH
The 12 subjects involved in acts of DSH during the 12-month period were all female. Their mean age was 15.5 years (SD 1.7, range 12.3 to 17.9); 11 were of Caucasian and 1 of African race. Their most common DSM-IV diagnoses on Axis I were major depression (seven subjects) and impulse control disorders (three subjects). Nine subjects fulfilled the criteria for one or more personality disorders, borderline personality disorder being the most common (nine subjects). Their mean length of hospitalization during the study period was 91.9 days (SD 73.2, range 7 to 200), and their mean length of hospitalization before the first DSH act was 56.2 days (SD 56.3, range 1 to 179). Age, complete DSM-IV diagnoses, length of hospitalization during the 12-month period, length of hospitalization before the first DSH act, type and number of DSH events, and number of DSH contagion incidents of the 12 subjects are presented in Table 1. The subjects in Table 1 and Figure 1 are numbered according to the temporal order of their first act of DSH during the 12-month period, e.g., subject 1 committed the first act of DSH during the study period.
Contagion of DSH
There were 37 instances of contagion during the 12-month period. The number of contagion incidents per subject ranged from 0 to 17 (mean 6.2, SD 5.8), and 10 of 12 patients with DSH were involved in contagion episodes. Their mean length of hospitalization before the first DSH act was 55.1 days (SD 55.2, range 10 to 179). An overall value obtained from the 11 independent [[Chi].sup.2] tests was significant ([[Chi].sup.2] = 9.1, df = 55, p [less than] .05). Thus, DSH occurred in nonrandom fashion and incidents of DSH were not independent of each other during the 12-month period.
Instances of contagion are illustrated with a sociogram, presented in Figure 1, in which each line connecting two subjects represents an incident of contagion. Two subjects, 4 and 5, were not involved in episodes of contagion at all. Subjects 1, 2, and 3 were involved in the first episode of contagion. Subjects 6 and 8, subjects 7 and 8, and subjects 6 and 7 were then involved in dyadic interactions of contagion. At the time of these dyadic episodes, subjects 1, 2, and 3 were no longer on the ward. Finally, there were two larger clusters of DSH, the first one involving subjects 7, 9, and 10, and the second one subjects 7, 8, 9, 11, and 12.
In most episodes of contagion the type of DSH was self-mutilation by skin cutting of wrists or arm and blood letting. Three subjects, 7, 8, and 12, also stored up blood in bottles kept in their wardrobes. However, there were also a few episodes in which skin cutting triggered other types of DSH in other subjects such as parasuicidal behavior by drug overdose or self-starving.
[TABULAR DATA FOR TABLE 1 OMITTED]
Findings From the Clinical Interviews
Of the six subjects interviewed (7, 8, 9, 10, 11, and 12), two subjects, 8 and 11, were involved in DSH for the first time, whereas four others had committed self-mutilation before hospitalization. It is of interest that subjects 8 and 11 were also the youngest of the 12 who had been involved in acts of DSH.
Two subjects, 11 and 12, said that self-mutilation did not relieve their anxiety or anger at all. According to their own judgment, they were involved in DSH mainly because they did not want to feel themselves outsiders. Subjects 6, 7, 8, 9, and 10 found that self-mutilation instantly relieved anxiety and/or anger. Subjects 8 and 9 did not experience pain while injuring themselves.
Only in subject 10 was self-injurious behavior probably not influenced by others, which is also supported by the fact that her two skin-cutting incidents preceded self-mutilation events by subjects 7 and 9. In five subjects DSH was clearly influenced by interaction with other inpatients. In subject 9 a dyadic idealizing interaction with subject 7 become evident. Subject 9 projected esteemed qualities onto subject 7, and thereafter she identified herself with them and with self-mutilative behavior as well.
Cardasis et al. (1997) found that patients with borderline personality disorder more often brought transitional objects into the psychiatric hospital than other patients. On the other hand, it has been suggested that the blood produced by self-cutting may serve as a comforting transitional object (Favazza, 1989). Subjects 7, 8, and 12, who had a peculiar habit of storing blood in their wardrobes, were all aware of the same custom in the other two subjects. We got the impression that only for subject 7 did blood storage serve as a transitional object, while for subjects 8 and 12, the main incentive for this habit was imitation of subject 7.
With the exception of subject 10, the contagion incidents were associated with group dynamics. Skin cutting and blood letting had become an initiation rite that strengthened group cohesion. However, it seems that a prerequisite for the feeling of togetherness was a shared emotional experience of relief after an event of DSH. Subject 11, who did not feel any relief with self-mutilation, was openly treated with contempt by subjects 6, 7, 8, and 9 and was labeled by them as a fake or a pretender.
DISCUSSION
Contagion of DSH
The present study replicates the findings from two previous studies carried out by Walsh and Rosen (Rosen and Walsh, 1989; Walsh and Rosen, 1985), that microlevel contagion of DSH occurs in psychiatric units for adolescents. Moreover, our results suggest that a majority of DSH incidents on an adolescent psychiatric ward may have been influenced by contagion. The present study also demonstrates that one type of DSH behavior, e.g., self-mutilation by skin cutting, may trigger another type of DSH behavior, e.g., parasuicide by drug overdose, in another individual.
Characteristics of Adolescents Who Are at Risk for DSH Contagion
Nine of the 12 female subjects with DSH incidents fulfilled the diagnostic criteria for borderline personality disorder. This is in line with previous studies, which have found that female patients with borderline personality disorder are predisposed both to self-mutilation and to imitative behavior (Favazza, 1989; Herpetz, 1995; Phillips and Carstensen, 1986).
Two of the six subjects interviewed started their DSH behavior on the ward. They were the two youngest patients on the ward, both with borderline personality disorder. The finding that DSH can spread on a psychiatric ward to previously DSH-naive adolescents is particularly alarming when we keep in mind the assumption that self-injurious behavior may advance to an addiction (Cold et al., 1983; Pies and Popli, 1995).
Social Factors on the Closed Psychiatric Ward
It has been proposed that insecurity in the organization and the high number of inexperienced staff prepare the ground for suicide contagion in psychiatric hospitals (Taiminen et al., 1992), and this assumption may hold true also concerning contagion of other types of DSH. However, during the two larger clusters of DSH within the study period, the number of inexperienced staff on the ward was at a minimum and during the study there were no organizational changes threatening.
During the two larger clusters of DSH, the ward was slightly overcrowded and the majority of the inpatients were depressed female adolescents with borderline personality disorder. Jones and Daniels (1996) have noted that stressful events such as changes in room population often provoke self-injury in socially isolated primates. It is possible that social isolation and social stress created by the overpopulation of the closed ward, together with a patient population exceptionally vulnerable to imitation, exposed the ward to contagion incidents.
Offer and Barglow (1960) noted that most of the self-mutilation incidents of adolescent inpatients occurred in the evening and during weekends, when structured activities were at a minimum. Ross and McKay (1979) found that incidence of self-mutilation of adolescent girls increased when the frequency of acting-out behaviors decreased. In the present study, the mean length of hospitalization before the first DSH contagion incident was almost 2 months. However, the ward described in the study was developed mainly for acute treatment and psychiatric evaluation, and therefore its resources for structured group activities are limited. Temporary overpopulation of the ward probably contributed to the lack of structured activities for the patients.
Clinical Implications: What Can Be Done to Prevent DSH Contagion in Psychiatric Units for Adolescents?
First, we suggest that the concentration of many female adolescents with borderline personality disorder on the same ward should be avoided. If this is not possible, the length of hospitalization should be limited to 2 weeks. These patients are vulnerable to a self-injurious subculture, and contagious self-mutilation can later become a socially shameful addiction promoting suicidal behavior (Favazza and Conterio, 1989).
Second, it seems that the method of DSH is also prone to imitation, because the majority of the subjects with DSH used the same method, skin cutting of wrist or arm, sometimes with blood letting. Therefore, access to the method of DSH chosen by an inpatient should be made as difficult as possible for other patients.
Third, in one subject of the present study the model of projective identification, previously described in the context of suicide contagion (Taiminen, 1992), seemed helpful. Projective identification with an object with DSH can often be detected by staff members if they actively look for signs of the blurring of self and object representations. After detection, the staff can discuss the differences between the subject and the adolescent with DSH, and the process of projective identification can be smothered (Taiminen, 1992).
Fourth, our results suggest that acts of DSH functioned as rites of togetherness for small groups of female adolescents. It should be evaluated whether these harmful rites of DSH could be replaced by some other group activity with a rite-like character, e.g., bungee-jumping or mountaineering.
Fifth, we believe that the issue of contagion should be openly discussed with the adolescents involved in DSH contagion. Ross and McKay (1979) noted that when they discussed in a neutral way self-mutilative behavior of adolescent girls with them, the frequency of self-mutilation decreased. We also found that the structured interviews described in the present study functioned as a therapeutic intervention. Most adolescents avoid behavior that they themselves perceive as imitative. Therefore, direct clarification of the possible contagious aspects of DSH behavior may prevent an adolescent from being involved in DSH incidents in the future.
REFERENCES
Brent DA, Kerr MM, Goldstein C, Bozigar J, Wartella M, Allan MJ (1989), An outbreak of suicide and suicidal behavior in a high school. J Am Acad Child Adolesc Psychiatry 28:918-924
Brent DA, Moritz G, Bridge J, Perper J, Canobbio R (1996), Long-term impact of exposure to suicide: a three-year controlled follow-up. J Am Acad Child Adolesc Psychiatry 35:646-653
Cardasis W, Hochman JA, Silk KR (1997), Transitional objects and borderline personality disorder. Am J Psychiatry 154:250-255
Coid J, Allolio B, Rees LH (1983), Raised plasma metenkephalin in patients who habitually mutilate themselves. Lancet 2:545-546
Crabtree CH, Grossman WK (1974), Administrative clarity and redefinition for an open adolescent unit. Psychiatry 37:350-359
Davidson L, Rosenberg M, Mercy J, Franklin J, Simmons J (1989), An epidemiologic study of risk factors in two teenage suicide clusters. JAMA 17:2689-2692
Favazza AR (1989), Why patients mutilate themselves. Hosp Community Psychiatry 40:137-145
Favazza AR, Conterio K (1989), Female habitual self-mutilators. Acta Psychiatr Scand 79:283-289
Favazza AR, Rosenthal RJ (1993), Diagnostic issues in self-mutilation. Hosp Community Psychiatry 44:134-140
Ghaziuddin M, Tsai L, Ghaziuddin N (1992), Mood disorder in a group of self-cutting adolescents. Acta Paedopsychiatr 55:103-105
Gould MS, Shaffer D (1986), The impact of suicide in television movies: evidence of imitation. N Engl J Med 315:690-694
Herpetz S (1995), Self-injurious behaviour: psychopathological and nosological characteristics in subtypes of self-injurers. Acta Psychiatr Scand 91:57-68
Hillbrand M, Krystal JH, Sharpe KS, Foster HG (1994), Clinical predictors of self-mutilation in hospitalized forensic patients. J Nerv Ment Dis 182:9-13
Jones IH, Daniels BA (1996), An ethological approach to self-injury. Br J Psychiatry 169:263-267
Kaminer Y (1986), Suicidal behavior and contagion among hospitalized adolescents (letter). N Engl J Med 315:1030
King CA, Franzese R, Gargan S, McGovern L, Ghaziuddin N, Naylor MW (1995), Suicide contagion among adolescents during acute psychiatric hospitalization. Psychiatr Serv 46:915-918
Matthews PC (1968), Epidemic self-injury in an adolescent unit. Int J Soc Psychiatry 14:125-133
Modestin J, Wurmle O (1989), Role of modelling in in-patient suicide: a lack of supporting evidence. Br J Psychiatry 155:511-514
Offer D, Barglow P (1960), Adolescent and young adult self-mutilation incidents in a general psychiatric hospital. Arch Gen Psychiatry 3:194-204
Phillips D, Carstensen L (1986), Clustering of teenage suicides after television news stories about suicide. N Engl J Med 315:685-689
Pies RW, Popli AP (1995), Self-injurious behavior: pathophysiology and implications for treatment. J Clin Psychiatry 56:580-588
Rissmiller DJ, Rissmiller F (1990), Inpatient suicide epidemics and suggestion for prevention. Hosp Community Psychiatry 41:922-924
Robbins D, Conroy R (1983), A cluster of adolescent suicide attempts: is suicide contagious? J Adolesc Health Care 3:253-255
Rosen PM, Walsh BW (1989), Patterns of contagion in self-mutilation epidemics. Am J Psychiatry 146:656-658
Ross RR, McKay HB (1979), Self-Mutilation. Toronto: Lexington Books, DC Heath
Sacks M, Eth S (1981), Pathological identification as a cause of suicide on an inpatient unit. Hosp Community Psychiatry 32:36-40
Schmidtke A, Hafner H (1988), The Werther effect after television films: new evidence for an old hypothesis. Psychol Med 18:665-676
Simeon D, Stanley B, Frances A, Mann JJ, Winchel R, Stanley M (1992), Self-mutilation in personality disorders: psychological and biological correlates. Am J Psychiatry 149:221-226
Simkin S, Hawton K, Whitehead L, Fagg J, Eagle M (1995), Media influence on parasuicide: a study of the effects of a television drama portrayal of paracetamol self-poisoning. Br J Psychiatry 167:754-759
Solomon Y, Farrand J (1996), "Why don't you do it properly?" Young women who self-injure. J Adolesc 19:111-119
Taiminen T, Salmenpera T, Lehtinen K (1992), A suicide epidemic in a psychiatric hospital. Suicide Life Threat Behar 22:350-363
Taiminen TJ (1992), Projective identification and suicide contagion. Acta Psychiatr Scand 85:449-452
Taiminen TJ, Helenius H (1994), Suicide clustering in a psychiatric hospital with a history of a suicide epidemic: a quantitative study. Am J Psychiatry 151:1087-1088
Vaisanen EJ, Hagglund T-B (1981), Destruction preceding suicide: an episode in the large series of adolescents' suicides. In: Depression et Suicide. New York: Pergamon, pp 670-675
Walsh BW, Rosen P (1985), Self-mutilation and contagion: an empirical list. Am J Psychiatry 142:119-120
Credits
©Journal of the American Academy of Child and Adolescent Psychiatry
Navigation
Back to Professional
Back to Articles
Back to Resources