Journal of Traumatic Stress. VoL 11. No. 2, 1998

Posttraumatic Stress Disorder and Family Functioning in Adolescent Cancer David Pelcovitz,1,2 Barbara Goldenberg Libov,1 Francine Mandel,1 Sandra Kaplan,1 Mark Weinblatt,1 and Aliza Septimus1

Twenty three adolescents with a history of cancer, 27 physically abused adolescents, and 23 healthy, nonabused adolescents were administered structured posttraumatic stress disorder (PTSD) interviews and self-report questionnaires regarding family functioning. Thirty five percent of adolescent cancer subjects met criteria for lifetime PTSD as compared to only 7% of the abused adolescents: 17% of the cancer subjects and 11% of the abuse subjects met criteria for current PTSD. Adolescents with cancer viewed their mothers and fathers as significantly more caring and more protective than the comparison and abused adolescents. Cancer subjects who met criteria for lifetime PTSD say their families as significantly more chaotic than those who did not have PTSD. Eighty three percent of cancer subjects who had lifetime PTSD also had mothers who had PTSD. KEY WORDS: adolescence: cancer; posttraumatic stress disorder.

This study presents the results of an investigation of the prevalence of posttraumatic stress disorder (PTSD) in a sample of adolescents with a history of cancer. In addition, the association between PTSD symptoms in the adolescents and family functioning is investigated. Unlike victims of other major stressors such as violence or natural disaster, the child faced with a life-threatening illness has to deal with a trauma that is chronic in nature. One can easily view childhood cancer and its treatment as one of the most traumatic events that a child and his/her family can endure. In 1North Shore University Hospital-NYU 2To whom correspondence should be

School of Medicine New York, New York 14853. addressed at the Division of Child and Adolescent Psychiatry, North Shore University Hospital, 300 Community Drive, Manhasset, New York 11030.

205 0894-9867/%O4DO-020iSlS.O<00/1 C 1998 International Society (or Traumatic Stress Studies

206

Ptlcovitz,Libov.Mandel, Kaplan, Weinblatt, and Septimus

addition to the overwhelming emotional impact of being faced with a lifethreatening illness at an age that is not usually associated with such serious life-threats, the nature of the treatment poses an additional source of trauma. During the active treatment phase, the child with cancer generally faces frequent spinal taps, venipunctures. hone marrow aspirations, broviac insertions, and blood transfusions. In addition, chemotherapy combinations may result in a wide range of painful and uncomfortable sequelae including nausea, acne, weight gain, and hair loss. Then, after the active treatment phase is concluded, the child and his/her family are faced with the constant threat of recurrence of the illness. Childhood cancer has a number of other characteristics which are associated with risk for PTSD. Severity of illness and the extent of the threat of the illness to one's life are important predictors of PTSD. In a review of risk factors associated with PTSD in crime victims, Kilpatrick and Resnick (1993) reported significantly higher PTSD prevalence rates in crimes which posed a threat to the life of the victim. There is also evidence that when a traumatic event is particularly unexpected the victim is more likely to develop PTSD (Davidson, 1993). Adolescents arc particularly vulnerable to the psychological effects of trauma. Dealing with a life threatening illness during a developmental stage in which peers are enjoying a moratorium from adult responsibilities is particularly overwhelming. Researchers have documented the particular vulnerability of adolescents to the effects of major stressors such as combat or rape (Burgess, 1985; Kulka, et al., 1990; van der Kolk, 1985). These findings are not surprising in light of the impact which a trauma such as cancer is likely to have on the major developmental tasks of adolescenceidentity formation and separation from one's family of origin. In some ways, the adolescent who is dealing with a hie threatening illness is at an advantage in comparison to younger children because of the adolescent's greater cognitive and emotional maturity. However, dealing with a serious illness at this developmental stage may affect the adolescent's ability to integrate past, present, and future expectations into a lasting sense of identity (Pynoos, 1993). This may have particular impact on the adolescent's ability to plan constructively for the future. In addition, a number of subjective factors associated with PTSD risk are relevant to adolescents with cancer including extreme fear, and lack of control (Davidson, 1993). Previous studies of cancer survivors have investigated the prevalence of various psychiatric symptoms, especially depression and anxiety (Fritz & Williams, 1988; Teta et al., 1986). Most studies, however, rely on measures that do not specifically focus on stress related to the illness. Researchers tend to rely on self-report measures of general psychological symptomatology, such as depression or anxiety, without specifically assessing stress re-

PTSD and Family Functioning

207

actions related to coping with a serious illness. Assessment of posttraumatic symptoms in children with cancer allows for a more focused assessment of the specific impact of dealing with the stress of a life-threatening illness.

PTSD and Cancer In recent years, a number of investigators have systematically investigated PTSD in adults suffering from a variety of serious illnesses such as coronary disease (Doerfler, Pbert, & DeCosimo, 1994), breast cancer (Alter et al., 1996), asthma (Yellowless, Hanes, Potts, & Ruffin, 1988), and burns (Powers, Cruse, Daniels, & Stevens, 1994). Empirical study of PTSD in children with serious illnesses is not as extensive. However, an increasing number of studies have been recently published investigating the role of PTSD in childhood cancer. In a clinical description of the emotional correlates of treatment for childhood cancer, Nir (1985) reported that PTSD is the most common psychiatric diagnosis that he found in childhood cancer patients. These symptoms included intrusive recollections, nightmares, and feelings of alienation from peers. Nir described how, in some families, any sign of somatic distress in the cancer survivor evokes such high levels of fear in the patient and family members that they frequently engage in unnecessary, "emergency" medical consultations. Although Nir was the first to conceptualize the stress associated with childhood cancer from the theoretical framework of PTSD, his clinical descriptions were not based on the use of standardized interviews or questionnaires. Stuber and her colleagues have conducted a series of studies systematically investigating the role of PTSD in survivors of childhood cancer. In a pilot study of six children who were given bone marrow transplants, Stuber and her colleagues (Stuber, Nader, Yasuda, Pynoos, & Cohen, 1991) reported that the number of children who met criteria for PTSD increased over time with three of the six children meeting criteria for PTSD one year following transplantation. The most prominent symptoms included reexperiencing and avoidance. In contrast to children traumatized by violence, arousal symptoms were relatively rare. In another study of 30 childhood cancer survivors, ages 8-19, appraisal of treatment intensity was significantly correlated with the severity of PTSD symptoms (Stuber, Meeske, Gonzalez, Houskamp, & Pynoos, 1994). Parental reactions are crucial predictors of how well a child adjusts to trauma. A child is more likely to develop PTSD when the parents show high levels of anxiety, distress, or PTSD symptoms (Bromet & Connell, 1984; Dawes, Tredoux, & Feinstein, 1989; Green et al., 1991). A family

208

Pelcovitz, Libov Mandel, Kaplan, Weinblatt, and Septimus

psychiatric history of anxiety and depression also places children at risk for PTSD (Davidson, Swartz, & Storck, 1985). In light of these findings, it is of particular note that mothers of pediatric cancer survivors have been reported to develop PTSD in response to the trauma associated with having a child with a life threatening illness (Pelcovitz et al., 1996). In their study of parents of 30 childhood cancer survivors, Stuber et al. (1994) hypothesized that mothers of childhood cancer survivors play a crucial role in mediating their child's PTSD symptomatology.

PTSD and Physical Abuse Studies of survivors of sexual abuse find PTSD rates ranging from 20.7% (Deblinger, McLeer, Atkins, Ralphe, & Foa, 1989) to 90% (Kiser, Heston, Millsap , & Pruitt, 1991). Although PTSD incidence of 35.8% has been reported in victims of severe childhood physical abuse (Famularo, Fenton, Kinschaff, Ayoub, & Barnum, 1994), much lower rates have been reported in victims of less severe physical abuse (Pelcovitz, Kaplan, Goldenberg, Lehane, & Guarrera, 1994). Adolescents faced with cancer differ from physically abused adolescents in a number of ways. The element of surprise and unpredictability have been associated with risk for PTSD symptoms in cancer patients (Cella, Mahon, & Donavan, 1990). Cancer is often viewed by the adolescent as a random act which has no clear etiology. In contrast, physical abuse is, by definition, caused by a deliberate act on the part of one's parents. Furthermore, the literature on adolescent physical abuse has noted that the adolescents are often perceived by themselves and others as more actively involved in the conflict (Pelcovitz, Kaplan, Samit, Krieger, & Ganeles, 1984). For example, abuse on the part of parents of adolescents is typically preceded by arguments over issues such as appropriate dress, friends, or curfews. In that these fights often have a more or less predictable pattern, the adolescent is likely to experience the abuse very differently than the more passive experience of being treated for adolescent cancer. Present Study The specific hypotheses of this study were that adolescents with a history of cancer would be at particular risk for PTSD as compared to a sample of physically abused adolescents, and healthy adolescents. In addition, we investigated the relationship between family functioning and the development of PTSD symptoms by looking at whether high levels of family

PTSD and Family Functioning

209

cohesiveness, flexibility, and support act as buffers against developing PTSD symptoms in adolescents with a history of cancer. Finally, we investigated the association between general psychiatric symptoms in the mothers of the adolescents with a history of cancer, the mothers who themselves had PTSD, and the development of PTSD in their ill child.

Method Twenty three adolescents with a history of cancer, 27 physically abused adolescents, and 23 comparison subjects with no history of chronic illness, life-threatening accidents, or physical abuse were administered structured PTSD interviews. In addition, subjects were administered self-report questionnaires regarding family functioning, as well as interviews regarding history of exposure to traumatic events. Mothers of the pediatric cancer survivors were administered structured PTSD interviews and a psychiatric symptoms checklist. AH interviews were conducted at the subjects' homes at their convenience. Written informed consent from all subjects was obtained from the custodial parent as well as the adolescent. The consent forms were reviewed and approved by the hospital's research review board. Subjects Adolescents with a history of cancer were recruited from a list of consecutively referred patients with childhood cancer at a teaching hospital's Division of Pediatric Hematology/Oncology. Subjects who met the following criteria were asked to participate: (1) the child's diagnosis was made before age 18; (2) the child was off treatment, or about to enter the maintenance phase of treatment; (3) the subject was at least age 13 at the time of the study. We refer to our sample as "adolescents" even though three of the subjects were over 20 at the time of the interview. We justify this because the older subjects were still living at home, attending school, and financially dependent on their parents. All of these factors suggest that they were developmentally at a similar stage as the younger subjects. The list of patients eligible for the study was generated by a database report. This database was a registry of all patients ever treated by the hospital's division of pediatric hematology/oncology. Fifty-one names of patients meeting the criteria were generated. Of those 51, eleven had either moved out of the area, or no addresses were available. Of the remaining 40 patients, 16 did not participate. The following are the reasons given by the subjects for nonparticipation: topic too painful to discuss (31%); moved

210

Pelcovitz. Libov, Mandel. Kaplan, Weinblatt, and Septimus

out of state (25%); did not speak English (25%); and too time consuming (19%). The adolescents with a history of cancer were interviewed as part of the DSM-IV PTSD field trials, which included structured interviews and several self-report questionnaires, some of which were not germane to this study. Adolescent cancer survivors were paid S25 for their participation in the interview which lasted approximately 2 hr. Twenty seven adolescents who were indicated as physically abused following a child protective services investigation were recruited from a list of consecutively indicated cases of physical abuse reported to the New York Stale register for child abuse. Subjects who had a history of intrafamilial sexual abuse either by child protective services report, or by self-report. were excluded. A more detailed presentation of psychiatric diagnoses, behavioral disturbance, and PTSD symptoms in the abuse sample are reported elsewhere (Pelcovitz et al., 1994). A nonabused community comparison sample, living in the same neighborhood as the ill subjects, was recruited from a larger pool of 99 adolescents who were interviewed as the comparison group in the study of physically abused adolescents. The subsample consisted of subjects who matched the cancer survivors with respect to age, race, gender, and socioeconomic status (Hollingshead, 1975). Both the physically abused and comparison subjects were interviewed for approximately an additional 3 hr to gather data which were not relevant to this study. The adolescent's share of the money paid to these subjects averaged $75. Chronic illness and life threatening accidents were ruled out in the comparison group using a structured interview, the K-SADS (Orvaschel, Puig-Antich, Chambers, & Abrizi, 1982), which was administered as part of the adolescent physical abuse study. This interview contains a comprehensive section asking parents about their child's medical history and history of life threatening accidents. Demographic characteristics of the subjects in the three groups are presented in Table 1. The gender distribution as well as the sociodemographic status in each of the three groups was approximately equal (chi square test). There was a significant difference in the proportion of White subjects in the three groups (Fisher exact test, p < .01). There were also significant differences between the three groups in age. The adolescent cancer group was significantly older than the subjects in the abuse group, F(2, 23) = 7.5, p < .001. The abuse group mean age was 15.1 (+/-1.7), comparison group mean age, 16.1 (+/-1.9); and the cancer group mean age was 17.6 (+/-3.1). The implications of these differences will be addressed in the discussion section.

PTSD and Family Functioning

211

Table 1. Demographics for the Adolescent Cancer (n = 23). the Adolescent Abuse (n = 27). and the Healthy Control (n = 23) Groups Variable Age* Median Inter-quartile range Gender (Frequency & Percentage) Girls Boys Race* Caucasian African-American Hispanic Asian Socioeconomic level (Hollingshead Index) I II III IV V

Adolescent Cancer

Adolescent Abuse

Healthy Control

17 15-19

15

14-16

16 15-17

12 (52%) 11 (48% )

16(59%)

11 ( 4 1 % )

12(52%) 11 ( 4 8 % )

17 (74%)

27 (100%)

23 (100%)

4(15%) 13(48%) 9 (33%) I (40) 0(0%)

6(26%) 7(30%) 6 (26O) 4 (18%) 0 (0%)

4 (18%) 1

(40)

1 (4%)

8 (35%) 8 (35% ) 2 (90) 4 (18%) 1(4%)

•Differences significant at p < .01.

Severity of abuse was calculated from Child Protective Services records recording their investigations conclusion regarding a number of physical abuse variables. These records were used in a rating scale scored by unanimous agreement of three child abuse experts who had an average of 15 years experience working with victims of abuse. Abuse victims received severity scores based on points received for the following variables: (1) Type of abuse was given a score ranging from 1 to 5 (i.e., 1 point for pushing or hitting to 5 points for "torture"); (2) Physical injuries resulting from the abuse were assigned 1 point for fractures, internal injuries and burns; (3) Treatment of abuse was given 1 point for visits to a physician or emergency room and 3 points if hospitalized. The total number of points was multiplied by 3 if CPS investigators determined that the abuse was "frequent." Severity scores for the 99 physically abused subjects, from which the 27 adolescents in the current study were recruited, ranged from 1 to 36, with an interquartile range of 2-6. A score of 3 or less was received by 50% of the sample and 86% of the sample fell at or below a severity score of 6. Although no norms are available, this suggests that the level of severity of abuse was relatively mild.

212

Pelcovitz, Libov, Mandel. Kaplan, Weinblatt, and Septimus

For the purpose of further describing the cancer sample, type of cancer diagnosis, as well as characteristics of the illness and treatment are presented in Table 2. Illness severity was determined by a pediatric oncologist, who divided the patients into three risk categories (mild, moderate, and severe), according to prognosis for survival at the time of diagnosis. These levels of severity of illness were based on guidelines established in the pediatric oncology literature for each of the pediatric malignancies. Our assumption in designing the study was that, even though the types and age of onset of cancer varied, the stress on the survivors and their families was similar in that all subjects were exposed to the trauma of coping with a life threatening illness which required painful treatments with significant side effects. Measures Structured Clinical Interview- for DSM (SCID). The posttraumatic stress disorder module of the SCID was administered to all subjects. The SCID Table 2. Medical Information for the Adolescents with Cancer (N = 23) Variable

Frequency

Percentage

Type of cancer Leukemias Lymphomas Carcinomas Wilm's tumor

12 7 3 1

52%

Treatment type Chemotherapy only Radiation only Chemotherapy + radiation

i 10 12

30% 13% 4% •! "r

43%

52%

Degreec of risk (Oncologist's subjective rating)

Low Intermediate High

3 12 8

13%

52% 35%

Occurrence of relapse

Yes No Ages (mean and range) Current age Age of onset Age off treatment Years since active treatment

3

20 16.0 10.5 12.5

3.3

13% 87%

14-23 2-18 5-19 0-11

PTSD and Family Functioning

213

was also administered to the mothers of adolescents with a history of cancer regarding their own reactions to the stress of coping with their child's illness (Spitzer, Williams, & Gibbon, 1987). The interview assessed the presence of current and lifetime PTSD based on DSMIII-R criteria. Current PTSD was defined as presence of symptoms within the last 6 months, and lifetime refers to symptoms present at some point during the persons life. Interviewers were graduate students in doctoral programs in clinical psychology. Training of interviewers consisted of reading a SCID-PTSD training manual which clearly delineated the standards necessary to meet the criteria for a DSMIII-R diagnosis of PTSD. In addition, training sessions were held for the interviewers to listen jointly to audiotapes of previous SCID-PTSD interviews. As noted earlier, the chronic illness sample was interviewed as part of the DSM-IV PTSD field trials which recruited subjects from five sites. Inter-rater kappa coefficients measuring the reliability of interviewers from the authors' site was .72 for current and lifetime PTSD (Kilpatrick, Resnick, & Freedy, 1992). The SCID has been used on adolescent subjects in a number of studies, and according to the senior author of the instrument has been found to be suitable for this age group (Spitzer, personal communication, January 1991). Modified-Diagnostic Interview Schedule, PTSD Module. The design of the DSM-IV PTSD field trials included a mechanism for assessing the connection between positive PTSD symptoms and to which stressors the symptoms were related. (Kilpatrick et al., in press). This was done by administering a modified version of the Diagnostic Interview Schedule PTSD module, which contained follow-up probes to elicit event-specific information. In the present study, this measure was used to determine whether positive PTSD symptoms were linked to the patient's exposure to having a life threatening illness, and in the mother's interview, their exposure to having a child with cancer. The SCID rather than the Modified DIS was used for the diagnosis of PTSD, since in the field trials the SCIDPTSD module was viewed as the "gold standard" for PTSD diagnosis. Measures Administered to Adolescents Only Parental Bonding Instrument (PBI). The PBI (Parker, 1983) was developed as a self-report scale measuring subjects' perceptions of two dimensions underlying parental attitudes and behaviors—care and protection. Parker (1983) places PBI responses into different categories. The categories which are relevant for this study are "optimal parenting," "affectionate constraint," and "affectionless control." What these categories measure will be described later.

214

Pelcovitz, Libov, Manddel, Kaplan, Weinblatt, and Septimus

The author reports that the PBI has good stability, with 3-week testretest correlations of .76 for care and .63 for protection. Split-half reliability coefficients were .88 and .67 for the care and protection scales, respectively (Parker, 1983). Twenty items, rated from 1 to 4, describe the adolescents' attitudes toward their mothers in one version, and their fathers in another. Care scores range from 0 to 36 and protection scores from 0 to 39. Family Adaptability and Cohesion Evaluation Scale (FACES III). This scale is a 20 item self-report measure which has been designed to determine an individual's sense of his or her family's cohesiveness and flexibility (Olson, Porter, & Lavee, 1985). Alpha reliability for FACES III, as reported by the measures authors, ranges from .75 to .90. Test-retest reliability is reported as .84. Measures administered to mothers of adolescents with a history of cancer only Symptom Checklist-90-R (SCL-90-R). This measure (Derogatis, 1977), which was administered to the mothers of the cancer survivors, is a 90-item self-report instrument which assesses the presence of psychological distress. For this study, we examined the Global Severity Index (GSI) which is a global measure of intensity of emotional distress. Internal consistency of the nine symptom dimensions measured by this instrument ranges from .77 to .90. Test-retest reliability ranges from .8 to .9. Derogatis cites numerous studies which support the concurrent and construct validity of the instrument. Statistical Analyses Chi-square analyses were conducted to determine diagnostic differences between groups. Kruskal-Wallis tests were used to compare the three groups on all of the following variables: mother and father care, mother and father protection, family adaptability and family cohesion. Significance levels were set at .05 for all tests (two-tailed). Logistic regressions models were computed to estimate the odds ratios for risk of developing PTSD.

Results Eight out of the 23 (34.8%) adolescents with a history of cancer met criteria for lifetime PTSD as compared to only 2 of the 27 abused adolescents (7%), and 1 of the 23 comparisons (4%) x2(2, N = 23) = 10.3, p < 0.01. There were no significant differences among the three groups on the proportion of adolescents who met criteria for current PTSD (17.4% of the adolescents with a history of cancer, 11.1% of the abused adolescents,

PTSD and Family Functioning

215

and none of the adolescents from the comparison group). All of the adolescents with a history of cancer described their symptoms as stemming from the diagnosis or treatment of cancer. In contrast, as noted elsewhere (Pelcovitz et al., 1994), the three adolescent abuse victims who met criteria for PTSD described their symptoms as related to extrafamilial sexual abuse incidents, and not the physical abuse. On the Parental Bonding Instrument, significant differences were found in all four categories. For mother's care (cancer median 30, Interquartile Range (IQR) = 28-30, abuse median = 25, IQR = 20-30, comparison median = 32.5, IQR = 28.5-35.5: p> < .01), pairwise comparisons determined that the abuse group reported significantly less maternal care than the cancer or comparison groups. Similarly, for father's care (cancer median = 30.5, IQR = 29-32, abuse median = 19, IQR = 14-25, comparison median = 29.5, IQR = 25-34.5, p < .001), pairwise comparisons also determined that the abuse group reported significantly less paternal care. The adolescent's perception of mother's level of protection (cancer median = 30, IQR = 27-33, abuse median = 10, IQR == 6-16, comparison median = 11.5, IQR = 6-16.5, p < .001) revealed on pairwise comparisons that the mothers of the adolescents with cancer were viewed as significantly more protective than the abuse or comparison group. The cancer adolescents also perceived their fathers as more protective (cancer median = 32, IQR = 30-36, abuse median = 16, IQR = 10-19, comparison median = 9, IQR = 7-14.5, p < .001). Parker (1983) places PBI responses into different categories. The comparison group adolescents describe their parents in a manner which falls into what Parker terms "optimal parenting;" this reflects a perception on the part of the healthy adolescents that their parents arc appropriately caring and protective. Compared to Parker's norms, the adolescents with a history of cancer viewed their mothers and fathers as overly protective and highly caring. Their response pattern fit into what Parker terms the "affectionate constraint" category, which reflects a level of discomfort on the part of the adolescents regarding what they perceive as overly high levels of parental concern. The abuse group fit Parker's category of "affectionless control" reflecting a view of their parents as overprotective in an uncaring way. Although abused adolescents perceive both parents in this manner, the scores describing paternal levels of protection and caring were more deviant from Parker's optimal parenting category than were the scores reflecting their view of their mother's level of protection and care. PBI scores in cancer survivors with PTSD were not significantly different than cancer survivors without PTSD. On FACESIII, there were no significant differences in perceived familial adaptability and cohesion among the three groups. However, when

216

Pelcovitz, Libov, Mandel, Kaplan, Weinblatt, and Septimus

comparing the adolescents with a history of cancer who met criteria for lifetime PTSD to adolescents who did not, adolescents with PTSD saw their families as significantly more chaotic than the ones who did not have PTSD. All of the cancer survivors who had current PTSD also had mothers who had current PTSD (p < .05). Five of the six (83%) cancer survivors who had lifetime PTSD had mothers with lifetime PTSD (p < .05). If a mother had current PTSD, her child was seven times more likely to develop PTSD than if the mother did not have PTSD (odds ratio = 7, 95% confidence interval = 1.94 to 25.0). Maternal scores on the SCL-90-R Global Severity Index were not significantly related to the adolescent's PTSD status. A complete discussion of the results of the investigation of PTSD in the mothers of adolescents with a history of cancer is presented elsewhere (Pelcovitz et al., 1996)

Discussion The finding that there were significantly higher prevalence rates of lifetime PTSD in adolescents with a history of cancer than in physically abused adolescents and healthy non-abused adolescents is not surprising in light of what is known about risk factors for PTSD. Adolescence is a developmental stage which is particularly vulnerable to the stress of confronting a major trauma (Burgess, 1985; Kulka et al., 1990; van der Kolk, 1985). At a time when their peers are dealing with issues related to autonomy and identity, the adolescent cancer patient is confronted with a serious illness which forces them to be dependent on parents and medical personnel. At the same time, peer relationships are frequently disrupted and the adolescent often reports feeling isolated from his healthy peer group (Pelcovitz, 1994). The process of dealing with a chronic illness during adolescence involves a number of variables which have been associated with high risk for PTSD including life-threat, physical injury, hospitalization, and surprise (DSM-IV, 1994). Recent studies have supported the notion that as a general rule, younger survivors of a variety of traumas are at greater risk for PTSD than are older individuals (Norris, 1992). Younger adults with breast cancer have been found to be at greater risk for PTSD symptoms than older patients (Cordova et al., 1995). Medical treatment has also been associated with PTSD risk (Davidson & Smith, 1990). It should also be pointed out that PTSD prevalence in the cancer patients in this study may, in fact, be an underestimate, because the relatively high refusal rate in the cancer sample excluded potential subjects whose discomfort talking about their experience with cancer may have made it more likely that they suffered from cancer related PTSD. On the other hand, it is also possible

PTSD and Family Functioning

2!7

that those who volunteered to participate in a study offering "free" psychiatric interviews may have been more likely to have PTSD symptoms. Since risk for PTSD is generally associated with traumatic events which are deliberately caused by others (Kilpatrick & Resnick, 1993), our finding that PTSD was more prevalent in victims of an "act of God" versus an "act of man" is surprising. Relatively high PTSD prevalence is reported in victims of violent crime such as single event assaults (Kilpatrick & Resnick, 1992). In a systematic investigation of PTSD in adolescent physical abuse, Pelcovitz et al. (1994) found high rates of conduct disorder and depression but virtually no PTSD. They speculated that the reason PTSD was not associated with physical abuse was because the adolescents lived in an environment that was so dominated by violence that the abuse was viewed as an ongoing process rather than a discrete "event." As Terr's clinical research suggests, different symptom patterns are often seen in survivors of a single event as compared to ongoing trauma (Terr, 1991). The low rate of PTSD in this study may have also resulted from the fact that the adolescents were still living in the abusive environment and may have needed to supress or deny their symptoms. This possibility is consistent with the finding of a recent investigation of prevalence of PTSD in a sample of children who were abused so severely that they were removed from their parent's custody. In that study, Famularo et al. (1994) found that 35.8% of the abused pre-adolescents met current criteria for PTSD on a structured interview. This is not surprising in light of recent research which has found life threat to be a core component in predicting PTSD (DSM-IV, 1994; Pynoos et al., 1993). The family variables which appeared to be most associated with PTSD risk in adolescents with a history of cancer were mother's PTSD status as well as the cancer survivor's perception of the family as chaotic. These findings are consistent with those of other investigators who report that parents who respond to a trauma with high levels of posttraumatic symptoms are more likely to have children who face difficulties adjusting (Bromet & Conneil, 1984; Dawes, Tredoux, & Feinstein, 1989; Green et al., 1991). Clinicians working with families of adolescents with cancer should he alerted to the importance of assessing the presence of posttraumatic symptoms in mothers. It is of note that, in some subjects, these symptoms were still active years after it was clear that their child was not facing an immediate risk to their health. Helping parents of adolescents with a history of cancer to learn more active coping mechanisms for dealing with the impact of this trauma should ultimately prove to be an important step in helping the adolescent cope with their illness. It is important to note that although maternal PTSD status was correlated with adolescent PTSD status, maternal scores measuring overall ad-

218

Pelcovitz, Libov. Mandel, Kaplan, Weinblatt, and Septimus

justment on the SCL-90-R were not significantly related to PTSD status in their children. This finding supports the importance of using traumaspecific assessment strategies in this population. Systematic investigation of families of children with cancer rely almost exclusively on measures that do not specifically focus on stress related to the illness (Perrin, Stein, & Drotar, 1991; Spirito et al., 1990). Failure to explore the types of psychological reactions which typically result from exposure to an extremely traumatic event may result in false negative findings. The advantage of systematically assessing the prevalence of PTSD in this population is that, unlike more general approaches, this methodology allows for specific questions about the impact of the illness as well as symptoms characteristically seen in individuals affected by extreme stress. Adolescents with a history of cancer described their parents as caring, but overly protective. This finding is consistent with descriptions in the family therapy literature of the regressive pull which chronic illness frequently places on the family (Rolland, 1994). At a time when adolescents need to feel independent and in control of their bodies, cancer pulls them into a state of dependency on their parents, who, in turn, may deal with the trauma of their child's illness by becoming overly protective. Our findings suggest that clinicians working with families of adolescent cancer survivors need to be attuned to these issues, which may continue to be present long after active treatment for the illness has terminated. Caution is advised regarding the generalizabilty of the results because of the relatively small number of subjects and the predominantly White and middle class composition of our sample. The demographic differences between the abuse and cancer subjects may have also biased the results, since the older age of the cancer subjects as weli as the fact that almost one quarter of the sample was not White may have biased the results (Burgess, 1985). In addition, the criteria we used for re-experiencing symptoms may have differed from those used in studies of PTSD in cancer which found lower PTSD prevalence (Cordova et al., 1995). As in many PTSD studies, our definition of re-experiencing included any situation where the subject had intrusive thoughts about any aspect of the illness or diagnosis including fear of future recurrence. A more stringent definition of re-experiencing which would not include future oriented ruminations may have reduced the number of subjects who met criteria for PTSD. The main purpose of this study was to systematically evaluate whether childhood cancer can result in PTSD. Further research is currently being conducted by our research team to systematically assess treatment related variables such as perceived life threat, rate of relapse, number of years since active treatment and the interaction between these variables and PTSD symptoms. We are also investigating the role of coping mechanisms

PTSD and Family Functioning

219

and the development of PTSD in childhood cancer patients. In addition, further research is needed on larger and better matched samples. Future study is needed to assess the role of fathers in the development of PTSD in the adolescent cancer survivor. Results of this study highlight the importance of systematically investigating the possibility of PTSD in adolescent cancer survivors, as well as the importance of assessing and treating the interaction between these symptoms and family functioning.

Acknowledgments

This research was supported in part through the N.I.M.H. funded DSM-IV PTSD Field Trials; and N.I.M.H. Grant 1RO1MH43772-04 "Psychopathology, Suicidal Behavior and Adolescent Abuse." The opinions expressed in this article are those of the authors and do not necessarily represent the position of the American Psychiatric Association or its Task Force on DSM-IV. The authors thank Dean Kilpatrick, Heidi Resnick, Patti Resick, Susan Roth, Bessel van der Kolk, and John Freedy for their contributions to the DSM-IV PTSD Field Trials.

References Alter, C. L.. Pelcovitz, D., Axelrod. A., Goldenberg. B., Harris, H., & Myers. B. (19%). The identification of PTSD in adult cancer survivors. Psychosomatics, 37. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, D.C.: Author. Bromet, E., & Connell. M. (1984). Mental health of children near the Three Mile Island reactor. Journal of Preventive Psychiatry, 2, 275-301. Burgess A. W (1985). Sexual victimization of adolescents. In A. W. Burgess (Ed.) Rape and sexual assault: A research handbook. New York: Garland. Cella, D., Mahon, S. M., & Donavan, M. (1990) Cancer recurrence as a traumatic event Behavioral Medicine, 16, 15-22. Cordova, M. )., Andrykowski, M. A., Kenedy, D. E.. McGrath, P. C., Sloan, D. A., & Redd, W. H. (199S). Frequency and correlates of posttraumatic stress disorder like symptoms after treatment for breast cancer. Journal of Consulting and Clinical Psychology, 63, 981-986. Davidson, J. (1993). Issues in the diagnosis of posttraumatic stress disorder. In J. Oldham. M. Riba, & A. Tasman. (Eds.) Review of psychiatry, Vol. 12 (pp. 141-155). Washington DC: American Psychiatric Press. Davidson, J., & Smith, R. (1990). Traumatic experiences in psychiatric outpatients. Journal of Traumatic Stress, 3. 459-476. Davidson, ]., Swartz, M., & Storck, M. (1985). A diagnostic and family study of posttraumatic stress disorder. American Journal of Psychiatry, 142, 90-93. Dawes, A., Tredoux, C., & Feinstein, A. (1989). Political violence in South Africa: Some effects on children of the violent destruction of their community. International Journal of Menial Health, 18(2), 16-43.

220

Pelcovitz, Libov; Mandel, Kaplan, Weinblatt, and Septimus

Deblinger E., M.Leer S. V, Atkins, M. S., Ralphe D, & Foa, E, (1989). Post-traumatic stress in sexually abused, physically abused, and nonabused children. Child Abuse and Neglect, 13, 403-408. Derogatis, L. R. (1977). SCL-90: Administration, scoring and procedures manual. I. Baltimore: Johns Hopkins University. Doerfler, L. A., Pbert, L., & DeCosimo, D. (1994). Symptoms of posttraumatic stress disorder following myocardial infarction and coronary bypass surgery. General Hospital Psychiatry, 16, I93-J99. Famularo, R., Fenton, T, Kinscherff, R., Ayoub, C, & Bamum, R. (1994). Maternal and child posttraumatic stress disorder in cases of child maltreatment. Child Abuse and Neglect, 18, 27-36. Fritz, G. K., & Williams, J. R. (1988). Issues of adolescent development for survivors of childhood cancer. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 712-715. Green, B., Korol, M., Grace, M., Vary, M., Leonard, A., Gleser, G., & Smitson-Cohen, S. (1991). Children and disaster: Age, gender, and parental effects on PTSD symptoms. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 945-951. Hollingshead, A. B. (1975). Four factor index of social status. New Haven: Department of Sociology, Yale University. Kilpatrick, D. G., Resnick, H. S., Freedy, J. R., Pelcovitz, D., Resick, P., Roth, S., & van der Kolk, B. (in press). The posttraumatic stress disorder field trial: Emphasis on criterion A and overall PTSD diagnosis. In T. Whitaker A. J. Francis, H. A. Pincus, M., First, & R. Ross (Eds.) DSM-IV Sourcebook. Washington D.C.: American Psychiatric Press. Kilpatrick, D. G., Resnick, H. S., & Freedy, J. R. (1992, October). Posttraumatic stress disorder field trial report: A comprehensive review of initial results. Presented at the Annual Meeting of the American Psychiatric Association. Washington, D.C. Kilpatrick, D. G. & Resnick, H. S. (1993). Posttraumatic stress disorder associated with exposure to criminal victimization in clinical and community populations. In 3. Davidson & E. Foa (Eds) Posttraumatic stress disorder: DSM-IV and beyond (pp. 113-143). Washington, D.C.: American Psychiatric Press. Kiser, L. J., Heston, J., Millsap, P., & Pruitt, D. B. (1991). Physical and sexual abuse in childhood: Relationship with posttraumatic stress disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 30(5), 776-783. Kulka, R., Schlenger, W., Fairbank, J., Hough, R., and Jordon, B., Marmar, C., & Weiss, D. (1990). Trauma and the Vietnam War generation. New York: Brunner/Mazel). Nir, Y. (1985). Posttraumatic stress disorder in children with cancer. In S. Eth & R. Pynoos (Eds). Posttraumatic stress disorder in children (pp. 123-132). Washington, D.C: American Psychiatric Press. Olson D., Porter J., & Lavee Y. (1985). Family Adaptability and Cohesion Evaluation Scale III (FACES III). Family Social Sciences, University of Minnesota. Orvaschel, H., Puig-Antich, P., Chambers, W., & Abrizi, M. (1982). Retrospective assessment of prepubertal major depression with the Kiddie SADS. Journal of the American Academy of Child and Adolescent Psychiatry, 21, 392-397. Parker, G. (1983). Parental overprotection: A risk factor in psychosocial development. New York: Grune Stratton. Pelcovitz, D., Kaplan, S., Samit, C., Krieger, R., & Ganeles, D. (1984). Adolescent abuse: Family structure and implications for treatment. Journal of the American Academy of Child and Adolescent Psychiatry, 23, 85-90. Pelcovitz, D. (1994, October). Post-traumatic stress disorder in adolescents. Paper presented atsymposium on Adolescent Abuse, American Academy of Child Psychiatry 41st Annual Meetings, New York. Pelcovitz, D., Goldenberg, B., Kaplan, S., Weinblatt, M., Mandel, E, Myers, B., & Vinciguerra, V. (1996). Posttraumatic stress disorder in mothers of pediatric cancer survivors. Psychosomatics, 37, 116-126.

PTSD and Family Functioning

221

Pelcovitz, D., Kaplan, S., Goldenberg, B., Mandel, E, Lehane, J., & Guarrera, J. (1994). Posttraumatic stress disorder in physically abused adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 305-312. Perrin, E. C., Stein, R. E., & Drotar, D. (1991). Cautions in using the Child Behavior Checklist: Observations based on research about children with a chronic illness. Journal of Pediatric Psychology; 16, 411-421. Powers, P. S., Cruse, C. W, Daniels, S., & Stevens, B. (1994). Posttraumatic stress disorder in patients with burns. Journal of Bum Care and Rehabilitation, 15, 147-153. Pynoos R. S. (1993). Traumatic stress and developmental psycbopathology in children and adolescents. In J. Oldham, M. Riba, & A. Tasman (Eds.) Review of psychiatry. Vol. 12 (pp. 205-238). Washington D.C.: American Psychiatric Press. Pynoos, R. S., Goenjian, A., Tashjian, M., Karakashian, M., Manijikian, R., & Manoukian, G. (1993). Posttraumaiic stress reactions in children after the 1988 Armenian earthquake. British Journal of Psychiatry, 163, 239-247. Rolland, J. S. (1994). Families, illness, and disability: An integrative treatment model. New York: Basic Books. Spirito, A., Stark, L. J., Cobiella, C., Drigan, R., Androkites, A., & Hewett, K. (1990). Social adjustment of children successfully treated for cancer. Journal of Pediatric Psychology, 15, 359-371. Spitzer, R. L., Williams, J. B., & Gibbon, M. (1987). Structured Clinical Interview for DSM-Ill-R. New York State Psychiatric Institute. Stuber, M., Gonzales, S., Meeske, K., Houskamp, B., Pynoos, R. S., & Kazak, A. (1994). Post-traumatic stress in childhood cancer survivors II: Family interaction. PsychoOncology, 3, 313-317. Stuber, M., Meeske, K., Gonzales, S., Houskamp, B., & Pynoos, R. S. (1994). Posttraumatic stress in childhood cancer survivors I: The role of appraisal. PsychoOncology, 3, 305-312. Stuber, M., Nader, K., Yasuda, P., Pynoos, R., & Cohen S. (1991). Stress responses after pediatric bone marrow transplantation: Preliminary results of a prospective longitudinal study. Journal of the American Academy of Child and Adolescent Psychiatry, 30, 952-957. Terr, L. (1991). Childhood traumas: An outline and overview. American Journal of Psychiatry, 148, 10-20. Teta, M. J., Del Po, M. C, Kasel, S. V, Meigs, J. W, Myers, M. H., & Mulvihill. J. J. (1986). Psychosocial consequences of childhood and adolescent cancer survival. Journal of Chronic Disabilities, 39, 751-759. van der Kolk, B. A. (1985). Adolescent vulnerability to posttraumatic stress disorder. Psychiatry, 48, 365-370. Yellowlees, P. M., Haynes, S., Potts, N., & Ruffin, R. E. (1988). Psychiatric morbidity in patients with life-threatening asthma: Initial report of a controlled study. Medical Journal of Australia, 149, 246-249.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.