| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Departments of Psychology (M.R.W., P.B.J.) and Internal Medicine (K.K.F.), University of South Florida; and the Psychosocial Oncology Program (M.R.W., P.B.J.) and the Blood and Marrow Transplant Program (K.K.F.), H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL.
Address reprint requests to: Paul B. Jacobsen, PhD, Psychosocial Oncology Program, H. Lee Moffitt Cancer Center and Research Institute, MOD3-PSY, 12902 Magnolia Dr., Tampa, FL 33612. Email: jacobsen{at}moffitt.usf.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Participants were 23 males and 79 females treated with BMT an average of 20 months previously (range = 3–62 months). Past and current psychiatric diagnoses were assessed through a structured clinical interview. PTSD symptomatology and other psychological variables were assessed using standardized self-report measures.
RESULTS: Results indicated that 5% of participants met diagnostic criteria for current PTSD. Participants reported an average of three to four symptoms of PTSD (range = 0–16). Univariate analyses confirmed predictions that increased PTSD symptomatology would be associated with more negative appraisals of the BMT experience, greater use of avoidance-based coping strategies, lower levels of social support, and greater social constraint (p < .05). Regression analyses indicated that each of these variables accounted for significant (p < .05) variability in PTSD symptomatology above and beyond relevant demographic and medical variables.
CONCLUSIONS: Results of the present study confirm and extend prior research regarding the prevalence of PTSD and PTSD symptoms among patients treated for cancer. In addition, the study identified a set of theoretically derived psychological characteristics that seem to place patients at risk for greater PTSD symptomatology after BMT.
Key Words: posttraumatic stress disorder • cancer • bone marrow transplantation.
Abbreviations: BMT = bone marrow transplantation; DSM-III-R = Diagnostic and Statistical Manual of Mental Disorders, third edition, revised; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, fourth edition; PCL-C = PTSD Checklist–Civilian Version; PTSD = posttraumatic stress disorder; SCID = Structured Clinical Interview for DSM-III-R; SCID-I/NP = Structured Clinical Interview for DSM-IV Disorders, Non-Patient Version.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Alter et al. (2) used a modified version of the SCID (3) to assess PTSD symptomatology in 27 female cancer patients an average of 4.6 years after treatment. Four percent of patients met DSM-IV criteria for a current diagnosis of PTSD. In another study that used the SCID (3), Green et al. (4) interviewed 160 breast cancer patients treated with surgery and radiation or chemotherapy an average of 6.5 months previously. Five percent of patients met diagnostic criteria for PTSD at some point after their cancer diagnosis, with 2.5% of patients meeting diagnostic criteria for current PTSD. Using the SCID-I/NP (5), Andrykowski et al. (6) interviewed 82 breast cancer patients treated with surgery with or without chemotherapy and/or radiotherapy an average of 37 months previously. Five percent of patients met diagnostic criteria for current cancer-related PTSD. An additional 4% met diagnostic criteria for lifetime cancer-related PTSD. In another report based on the same sample, Andrykowski and Cordova (7) used the self-report PCL-C (8) to examine the prevalence and correlates of PTSD symptomatology. Depending on the PCL-C scoring system used, 5 to 6% of patients were likely to meet DSM-IV criteria for current PTSD. Greater PTSD symptomatology was associated with more advanced disease, less time since treatment completion, greater current physical comorbidity, less current social support, and a greater history of trauma before cancer diagnosis.
Two additional studies can be identified that used the PCL-C to assess PTSD symptomatology in adult cancer patients. Cordova et al. (9) examined the prevalence and correlates of PTSD symptomatology in 55 breast cancer patients treated with surgery with or without chemotherapy and/or radiotherapy an average of 30.5 months previously. They found that 5.5 to 10.9% of patients were likely to meet DSM-IV diagnostic criteria for current PTSD depending on the PCL-C scoring method used. Greater PTSD symptomatology was associated with younger age, less education, and lower income. In the other study, Jacobsen et al. (10) examined the prevalence and correlates of PTSD symptomatology in 43 breast cancer patients treated with autologous BMT an average of 19.4 months previously. We found that 12 to 19% of patients were likely to meet DSM-IV criteria for current PTSD depending on the PCL-C scoring method used. Consistent with prior research (7, 9), more advanced disease and less education were associated with greater PTSD symptomatology. A longer hospital stay was also found to be related to increased PTSD symptoms.
As indicated by this review, research with adult cancer patients has focused primarily on documenting the prevalence of PTSD and identifying its demographic, medical, and treatment correlates. With the notable exception of the study by Andrykowski and Cordova (7), the relation of psychological factors to PTSD symptomatology in this patient population has not been examined. Lazarus and Folkman (11) have proposed a psychological model that is of potential relevance to the study of PTSD symptoms in cancer patients. This model posits that an individual’s appraisal of a stressful event influences the reaction to that event and the development of a stress response. An appraisal that an event is harmful, threatening, or challenging is seen as increasing the likelihood of a stress response occurring. This model is consistent with DSM-IV, which stipulates that an event must evoke a sense of intense fear, helplessness, or horror and must involve actual or threatened death or injury for PTSD to develop. The Lazarus and Folkman model (11) also posits that coping processes play an important role in determining an individual’s reaction to an event. Along these lines, it can be reasoned that greater use of avoidance-based coping processes (eg, distraction and distancing), rather than approach-based coping processes (eg, planning and problem solving), will be associated with greater PTSD symptomatology because avoidance is a hallmark feature of PTSD.
Later elaborations of the model have conceptualized social support as a form of coping (12). In that social support provides an individual with aid (eg, instrumental and informational), it may be similar to other approach-based coping processes in which aid is sought directly. Experience of social constraint may have the opposite effect (13). To the extent that individuals perceive their social environments to be unsupportive, they may be more like to engage in avoidance-based forms of coping that interfere with recovery after a traumatic event.
A limited number of studies provide empirical support for the role of subjective appraisals of the trauma experience in the development and maintenance of PTSD. Along these lines, fears of dying among accident victims (14) and perceptions of life threat among survivors of a shooting spree (15) have been shown to be associated with increased PTSD symptomatology. Research on the role of coping processes in the development of PTSD symptoms is also consistent with the proposed model. Increased use of avoidance-based coping strategies (eg, escape-avoidance, distancing, and denial) has repeatedly been shown to be associated with increased PTSD symptomatology in combat veterans and accident victims (16–22). Findings regarding the role of approach-based coping processes (eg, planful problem solving and confrontive coping) are less clearcut. Although some studies of combat veterans and accident victims have found that use of approach based or problem-focused coping processes is related to decreased PTSD symptomatology (19, 21), other studies have found no relationship (16–18, 20, 22). With regard to social support, there is evidence that higher levels of perceived support are associated with decreased PTSD symptomatology in veterans, burn victims, and community samples (23–25). Higher levels of social constraint have been shown to be related to greater intrusive thinking (a common PTSD symptom) in mothers who experienced the death of a child due to sudden infant death syndrome (13).
Taken together, these studies suggest that Lazarus and Folkman’s model is useful in explaining individual differences in PTSD symptomatology in a variety of populations. The aim of the present study was to apply this model to the study of PTSD among cancer patients who previously received allogeneic or autologous BMT. Autologous BMT, in which bone marrow is harvested from the individual’s own body and then retransplanted after chemotherapy and/or radiation, is associated with a 1 to 10% mortality rate and typically produces a number of aversive side effects, such as nausea, vomiting, pain, bleeding, and mucositis (26, 27). Allogeneic BMT, in which bone marrow is harvested from a donor and then retransplanted into the patient, is associated with a 20 to 30% mortality rate and additional potentially life-threatening complications, such as graft vs. host disease (26).
Based on Lazarus and Folkman’s model and prior research (14–25), it was hypothesized that increased PTSD symptomatology would be associated with more traumatic appraisals of the BMT experience as well as greater reliance on avoidance-based coping strategies, decreased perceptions of social support, and increased perceptions of social constraint. Based on prior research in other PTSD populations (14, 24, 28–32), it was also hypothesized that a history of mood or anxiety disorder would be associated with increased PTSD symptomatology among BMT recipients.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Of the 148 BMT recipients meeting eligibility criteria, 34 (23%) could not be reached, one (1%) declined to participate, and 11 (7%) completed the clinical interview but failed to return the questionnaire packet (see below). Thus, complete data were obtained from a total of 102 BMT recipients (69%). The 102 patients included in the final analysis were more likely to be female (
2(1) = 4.60, p < .05), white (2(1) = 4.95, p < .05), and older (t = 2.05, p < .05) than the 46 patients not included in the final analysis.
Procedure
Patients who underwent BMT 3 to 60 months previously were identified using a registry of BMT recipients. Patients meeting all eligibility criteria were initially contacted by means of an introductory letter from the director of the Blood and Marrow Transplantation Program. Individuals were then contacted by telephone to obtain verbal consent for study participation. BMT recipients who verbally agreed to participate were mailed an informed consent form, questionnaire packet, and stamped envelope in which to return the study materials. A research team member telephoned participants 1 week after the initial packet mailing to conduct a structured clinical interview, answer any questions, and provide a reminder to promptly return the study materials. A second reminder call was made if study materials had not been received within 2 weeks.
Measures
Demographic data.
Demographic data were obtained through use of a standard self-report questionnaire. Variables assessed included age, ethnicity, marital status, employment status, and education level.
Disease and treatment data.
A computerized patient database and patient medical charts were reviewed to obtain information regarding disease and treatment characteristics. Variables assessed included cancer diagnosis, disease status at transplant, hospitalization status at transplant (inpatient vs. outpatient), and dates of hospitalization for BMT.
Psychiatric diagnoses.
The SCID-I/NP (5) assesses selected axis I psychiatric disorders based on DSM-IV criteria. Past (lifetime) as well as current presence of individual DSM-IV disorders can be assessed. In the present study, only the modules for mood disorders, anxiety disorders, and adjustment disorders were administered because prior research suggests that these are the most common psychiatric disorders among cancer patients (33). SCID-I/NP interviews were conducted by telephone. Prior research supports the reliability and validity of telephone administration of the SCID (34). All interviews and ratings were conducted by a study investigator with extensive training in SCID-I/NP administration and scoring. Every fifth interview was audiotaped and rated independently by a second rater. Agreement between the two raters for the presence or absence of any axis I disorder (
= 0.74) and the presence or absence of PTSD ( = 1.00) was acceptable. Participants were assessed for PTSD resulting from any and all past experiences that meet DSM-IV criteria for a traumatic event. In addition, participants were instructed to separately consider their experiences with cancer diagnosis and treatment when responding to PTSD questions to distinguish PTSD symptoms related to their cancer experience from those related to other traumatic events.
PTSD symptomatology.
The PCL-C (8) is a 17-item self-report measure consisting of items that reflect DSM-IV symptom criteria for PTSD. The PCL-C yields a total score (range = 17–85) as well as subscale scores for intrusions (range = 5–25), avoidance/numbing (range = 7–35), and arousal (range = 5–25). Participants were instructed to respond to items in terms of their experience with cancer diagnosis and treatment and to rate each symptom in terms of how much it had bothered them in the past month (1 = not at all, 5 = extremely). Two methods have been developed to identify PCL-C respondents who are likely to meet DSM-IV criteria for PTSD (8). Using the symptom method, a formal diagnosis of PTSD is considered likely if participants endorse (ie, rate as "moderately," "quite a bit," or "extremely" bothersome) one or more intrusion symptom items, three or more avoidance/numbing symptom items, and two or more arousal symptoms in correspondence with DSM-IV symptom cluster criteria for PTSD. Using the cutoff method, a formal diagnosis of PTSD is considered likely if participants obtain a total score of 50 or greater. In addition to providing a dichotomous classification of individuals likely to meet diagnostic criteria for PTSD, the PCL-C can be used as a continuous measure, with higher scores indicating increased PTSD symptomatology. The PCL-C demonstrated adequate internal consistency for both the total scale (Cronbach’s 
= 0.89) and each of the subscales (Cronbach’s = 0.73–0.86) in the current study. The validity of the PCL-C was demonstrated in the current study by comparing PCL-C scores of BMT recipients with and without a current SCID-I/NP diagnosis of cancer-related PTSD. As expected, BMT recipients with a current SCID-I/NP diagnosis of cancer-related PTSD reported significantly greater levels of PTSD symptomatology on the PCL-C than did BMT recipients without a current diagnosis of PTSD (means = 42.00 vs. 29.01, p < .01).
Coping strategies.
The Brief COPE (35) is a 28-item self-report measure derived from the 60-item COPE (36). This measure can be formatted to assess situational reports of coping as well as dispositional "coping styles". With the dispositional format used in the current study, respondents indicate on a four-point Likert scale (0 = I usually don’t do this at all, 3 = I usually do this a lot) how often they use each specific coping strategy when dealing with stressful situations. For purposes of the current study, those subscales reflecting avoidance-based coping strategies (Behavioral Disengagement, Self-Distraction, Denial, and Substance Use) were grouped into a single scale by summing the relevant subscale scores. The avoidance-based scale demonstrated moderate internal consistency in the current study (Cronbach’s = 0.62).
Social support.
The Interpersonal Support Evaluation List–Short Form (37) is a 15-item self-report measure derived from the 40-item Interpersonal Support Evaluation List (38). The short form retains five items from each of the three subscales from the original measure: 1) tangible support, 2) appraisal support, and 3) belonging support. Respondents indicate how true or false each statement is for them using a four-point Likert scale (1 = completely false, 4 = completely true). Internal consistency for the total score was adequate in the current study (Cronbach’s = 0.81).
Social constraint.
The Social Constraint Scale (13) is a 10-item self-report measure developed to assess the degree to which individuals perceive their social relationships as constrained and the degree to which they feel constrained about discussing their thoughts and feelings. Five items measure perceived social constraint from the most important other person in the individual’s life, and five items measure the perceived social constraint from all other people in the individual’s life. In the current study, respondents indicated how much social constraint that they perceived from others regarding their cancer experience on a five-point Likert scale (1 = almost never, 5 = almost always). Internal consistency for the total score in the present study sample was adequate (Cronbach’s = 0.80).
Subjective trauma experience.
The six-item Trauma Experience Questionnaire was developed specifically for the current study to assess patients’ subjective experience of the traumatic value of BMT. Separate items assessed patients’ retrospective perceptions of helplessness, hopelessness, loss of control, pain, distress, and concerns about death during their hospitalization for BMT. Ratings are made on a four-point Likert scale (1 = none or not at all, 4 = severe or a lot). These items are similar to those used by Neumann (39) to study posttraumatic effects in myocardial infarction patients. The total score demonstrated adequate internal consistency in the present study (Cronbach’s = 0.81).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Although a range of cancer diagnoses was sampled, 49% of study participants underwent BMT due to breast cancer. Other common cancer diagnoses include non-Hodgkin’s lymphoma (15%) and chronic myelogenous leukemia (10%). Based on their disease status at time of transplant, 37% of participants were considered to be at low risk for disease recurrence or progression after BMT, 35% were considered to be at intermediate risk, and 25% were considered to be at high risk.1
The majority of participants underwent autologous BMT (83%) and received their transplants as inpatients (90%). BMT recipients treated as inpatients were hospitalized for an average of 28.39 days (SD = 11.17, range = 17–87) for transplantation. Participants were an average of 20.42 months (SD = 14.27, range = 3–62) post-BMT at the time of questionnaire completion.
Rates of Psychiatric Diagnoses
Based on SCID-I/NP interviews, 5% of BMT recipients (N = 5) received a current diagnosis of PTSD related to cancer diagnosis and treatment. No BMT recipients received a current diagnosis of PTSD related to other traumatic events. Eleven percent of BMT recipients (N = 11) received a current diagnosis of anxiety disorder other than PTSD, 4% (N = 4) received a current diagnosis of mood disorder, 2% (N = 2) received a current diagnosis of adjustment disorder, and 85% (N = 87) received no current diagnosis. Percentages total >100% because some participants had multiple diagnoses. In terms of lifetime (ie, other than current) psychiatric disorders, 4% (N = 4) received a lifetime diagnosis of PTSD related to cancer diagnosis and treatment and 2% (N = 2) received a lifetime diagnosis of PTSD related to other traumatic events. Six percent (N = 6) received a lifetime diagnosis of anxiety disorder other than PTSD, 30% (N = 31) received a lifetime diagnosis of mood disorder, and 66% (N = 67) received no lifetime diagnosis. Lifetime psychiatric disorders were further distinguished into those disorders that were present before and after transplant. Twenty-six percent of BMT recipients (N = 27) were identified as having a mood or anxiety disorder in their lifetime before transplant, with 2% of BMT recipients (N = 2) meeting criteria for PTSD due to a non–cancer-related traumatic event. Seventy-four percent of BMT recipients (N = 75) were identified as not having a mood or anxiety disorder in their lifetime before transplant.
To examine the possibility that cancer-related PTSD may have contributed to the failure to return questionnaire packets, Fisher’s Exact tests were conducted to determine the relationship between study completion status and current as well as lifetime SCID-I/NP diagnoses of cancer-related PTSD. There were no significant (p < .05) differences in the proportion of cancer-related PTSD diagnoses between BMT recipients who completed the SCID-I/NP but failed to return questionnaire packets (N = 11) and BMT recipients who completed both aspects of the study (N = 102).
Analysis of PCL-C Reports of PTSD Symptomatology
The percentage of participants who endorsed each PCL-C item (ie, rated it as "moderately," "quite a bit," or "extremely" bothersome) was calculated to identify the prevalence of specific PTSD symptoms among BMT recipients (see Table 1). The most frequently reported symptoms (ie, those endorsed by at least 25% of participants) were "feeling as if your future will be cut short" (avoidance/numbing cluster, 41%), "trouble falling or staying asleep" (arousal cluster, 30%), "having difficulty concentrating" (arousal cluster, 26%), and "trouble remembering important parts of your cancer treatment or your experience with cancer" (avoidance/numbing symptom cluster, 26%). The least frequently reported symptoms (ie, those endorsed by <10% of participants) were "having physical reactions when something reminds you of cancer treatment or your experience with cancer" (intrusion cluster, 8%) and "repeated, disturbing dreams of cancer treatment or your experience with cancer" (intrusion cluster, 8%).
|
50), six participants (6%) were identified as likely to merit a formal diagnosis of PTSD. Based on application of the symptom method (one intrusion symptom, three avoidance/numbing symptoms, and two arousal symptoms endorsed), nine participants (9%) were identified as likely to merit a formal diagnosis of PTSD. All six participants identified as being likely to receive a formal diagnosis of PTSD with the cutoff method were among those individuals also identified by the symptom method as meeting PTSD diagnostic criteria. The utility of each PCL-C scoring method was evaluated by comparing a PCL-C–based diagnosis of PTSD to the criterion of a SCID-I/NP diagnosis of PTSD. When using the recommended cutoff score of 50, one participant was correctly classified by the PCL-C as meriting a formal PTSD diagnosis (true positive), five participants were incorrectly classified as meriting a PTSD diagnosis (false positives), 92 participants were correctly classified as not meriting a diagnosis of PTSD (true negatives), and four participants were incorrectly classified as not meriting a diagnosis of PTSD (false negatives). Compared with the SCID-I/NP, the PCL-C cutoff of 50 yielded sensitivity of 0.20, specificity of 0.95, positive predictive power of 0.17, negative predictive power of 0.96, and overall diagnostic efficiency of 0.91.
Using the symptom method of scoring the PCL-C, two participants were correctly classified as meriting a PTSD diagnosis, seven participants were incorrectly classified as meriting a PTSD diagnosis, 90 participants were correctly classified as not meriting a diagnosis of PTSD, and three participants were incorrectly classified as not meriting a diagnosis of PTSD. Compared with the SCID-I/NP, the symptom method of scoring yielded a sensitivity of 0.40, specificity of 0.93, positive predictive power of 0.22, negative predictive power of 0.97, and diagnostic efficiency of 0.90.
Analysis of PTSD Symptom Severity
Appraisal of the trauma experience.
As predicted, more negative appraisals of the trauma experience were all associated with increased PCL-C scores (p < .01; see Table 2). The total subjective trauma appraisal score was also significantly associated with greater PTSD symptomatology (p < .001).
|
Social support and social constraint.
As shown in Table 2, results supported predictions regarding the relation of social support and social constraint to PTSD symptomatology. Decreased perceptions of support were all associated with increased PTSD symptomatology (p < .05). In contrast, greater perceptions of social constraint from a significant other (p < .01) and other people (p < .001) were both associated with greater PTSD symptomatology. Total scores for social support and social constraint were also associated with PTSD symptomatology in the predicted directions (p < .001).
Past psychiatric history.
As predicted (see Table 3), participants with a history of anxiety or mood disorder before BMT reported greater PTSD symptomatology than participants without a history of anxiety or mood disorder before BMT (p < .05).
|
Medical and treatment variables.
For those participants who underwent BMT as inpatients (N = 92), length of hospitalization was not related to PCL-C scores (r = 0.05, p >.05). Similarly, there were no relationships between PTSD symptomatology and type of transplant or disease risk status (p > .05; see Table 3). Likewise, there were no differences in PTSD symptomatology between breast cancer patients and all other patients (p > .05; see Table 3). In addition, PCL-C scores were not associated with length of time since BMT (r = 0.01, p > .05). In contrast, patient hospitalization status was marginally (p = .08) related to PTSD symptomatology, with inpatient transplant recipients reporting greater PTSD symptomatology than outpatient transplant recipients (see Table 3).
Multiple Regression Analysis of Factors Related to PTSD Symptomatology
A multiple regression analysis was performed to identify the variance in PTSD symptomatology accounted for by demographic, medical, treatment, and psychological factors. The analytic strategy was to first account for the contribution of demographic, medical, and treatment variables to provide a conservative test of the contribution of psychological factors to PTSD symptomatology. Accordingly, the demographic, medical, and treatment variables that were at least marginally (p < .10) associated with PCL-C total scores (age, pre-BMT psychiatric history, and hospitalization status) were forced into the regression analysis on the first step. Summary psychological measures found to be significantly (p < .05) related to increased PTSD symptomatology in univariate analyses (ie, total scores for trauma appraisal, social constraint, avoidance-based coping, and social support) were then allowed to enter the model in a forward selection manner.
Results of these analyses are presented in Table 4. Age, pre-BMT psychiatric history, and hospitalization status were found to account for 8% of the variance in PTSD symptomatology (p < .05). After controlling for these variables, subjective trauma appraisal entered the equation on the second step and accounted for 18% of the remaining variance in PCL-C scores (p < .01). Social constraint entered the equation on the third step and accounted for 9% of the remaining variance in PCL-C total scores (p < .01). Avoidance-based coping entered the equation on the fourth step and accounted for 3% of the remaining variance in PCL-C total scores (p < .05). In the fifth and final step of the analysis, social support entered the equation, accounting for 4% of the remaining variance in PCL-C total scores (p < .05). Thus, after controlling for age, pre-BMT psychiatric history, and hospitalization status, 34% of the variance in PCL-C total scores was accounted for by subjective trauma appraisal, avoidance-based coping, social support, and social constraint. Additional regression analyses were conducted to identify any interactions among trauma appraisal, coping, social support, and social constraint that were significantly related to PCL-C scores. No significant (p < .05) interaction effects were found among these psychological variables.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Results of the current study suggest that clinicians and researchers need to exercise caution in their use of the PCL-C as a method of screening for PTSD in cancer patients. The PCL-C demonstrated low sensitivity (0.20–0.40) and positive predictive power (0.17–0.22) when each scoring method was compared with a structured clinical interview in diagnosing current PTSD. These low estimates reflect a relatively high ratio of both false negatives and false positives to true positives for each scoring method. Although disappointing, these low estimates are not surprising given the low prevalence of PTSD in the current study (42). It should be noted that a previous study of PTSD in breast cancer patients (6) obtained somewhat higher estimates of sensitivity (0.60) and positive predictive power (0.60–0.75) for each PCL-C scoring method compared with a structured clinical interview.
Overall, results of the current study provide considerable evidence of the utility of the Lazarus and Folkman model (11). There was clear support for the view that subjective appraisals of the BMT experience would be related to differences in PTSD symptomatology. In addition, results supported predictions that increased use of avoidance-based coping strategies, decreased perceptions of social support, and increased perceptions of social constraint would be associated with greater PTSD symptomatology. Findings regarding the role of social support are consistent with those previously reported by Andrykowski and Cordova (7) in their study of PTSD symptoms after breast cancer treatment. Findings regarding the role of social constraint are consistent with research by Lepore and Helgeson (43), which found that greater social constraint was associated with higher levels of intrusive and avoidant thinking in prostate cancer patients.
Findings regarding the role of avoidance-based coping, social support, and social constraint can also be viewed as consistent with a social-cognitive model of trauma recovery (13). According to this model, the cognitive processing of a traumatic event through social interactions is an important component of psychological recovery. If an individual’s opportunity to cognitively process a traumatic event is restricted, as would be expected in the presence of avoidance-based coping processes, inadequate social support, or social constraint, the event may remain in active memory and result in intrusive thoughts as well as other PTSD symptoms.
Results concerning the role of demographic and medical variables in the development of PTSD were mixed in regard to previous research. With the exception of history of mood, anxiety, or adjustment disorder, no demographic or medical treatment variables assessed in the current study were significantly (p < .05) associated with PTSD symptomatology. In contrast, previous research has found that increased PTSD symptomatology in adult cancer patients is related to younger age (4, 9), lower income (9), less education (5, 10), more advanced disease (7, 10), and less time since treatment completion (7). Results are also inconsistent with our previous research, which found that greater PTSD symptomatology in BMT recipients was associated with a longer length of hospitalization (10).
The current study attempted to address some of the limitations of previous research concerning PTSD in cancer populations by using a theoretical framework, including a more heterogeneous sample of transplant recipients, and using a structured diagnostic interview to identify cancer-related PTSD. Despite these methodological improvements, several limitations remain. One major limitation is the use of a cross-sectional design in which time since BMT ranged from 3 to 62 months. This design does not permit conclusions to be drawn regarding the causal relations between the psychological factors that were measured and the development of PTSD symptomatology. The collection of retrospective data, particularly for accounts of the subjective experience of the transplant, is a second limitation. These recollections could be influenced by PTSD symptomatology rather than in the opposite manner as hypothesized. The use of a prospective, longitudinal research design could address both of these limitations. In addition, the inclusion of multiple follow-up assessments as part of a longitudinal study would allow for identification of the time course of PTSD symptom development and recovery. Sampling bias is a third limitation. Compared with nonparticipants, study participants were more likely to be female, white, and older. Consequently, it is possible that current study findings do not accurately reflect the full extent of PTSD symptoms in BMT recipients. It should be noted, however, that there were no differences in PTSD symptomatology between participants who completed all study measures and participants who completed only the SCID-I/NP.
In terms of clinical implications, the current study yielded a psychological profile of individuals likely to experience greater PTSD symptomatology. These results raise the possibility of offering preventive treatment to individuals with this profile before their development of PTSD symptoms after BMT. Preventive treatment might involve psychological preparation for the aftereffects of transplantation. Alternatively, it may be useful to provide debriefing, as done in response to other traumatic events (44), before the patient leaves the hospital environment. For patients who go on to develop symptoms of PTSD, the use of cognitive-behavioral interventions effective with other forms of PTSD (45, 46) merit investigation. Results from the present study indicating that social constraint was related to increased PTSD symptomatology suggest that support groups and supportive psychotherapy may also be useful. In light of evidence indicating that many cancer patients experience PTSD symptoms for months or even years after BMT, identifying ways to prevent or relieve these symptoms should be considered a priority for future research.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication August 16, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. Banou, S. E. Hobfoll, and R. D. Trochelman Loss of Resources as Mediators between Interpersonal Trauma and Traumatic and Depressive Symptoms among Women with Cancer J Health Psychol, March 1, 2009; 14(2): 200 - 214. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Stoppelbein and L. Greening Brief Report: The Risk of Posttraumatic Stress Disorder in Mothers of Children Diagnosed with Pediatric Cancer and Type I Diabetes J. Pediatr. Psychol., March 1, 2007; 32(2): 223 - 229. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. J. Thijssens, J. E. H. M. Hoekstra-Weebers, R. J. van Ginkel, and H. J. Hoekstra Quality of Life After Hyperthermic Isolated Limb Perfusion for Locally Advanced Extremity Soft Tissue Sarcoma Ann. Surg. Oncol., June 1, 2006; 13(6): 864 - 871. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Bruggimann, J. M. Annoni, F. Staub, N. von Steinbuchel, M. Van der Linden, and J. Bogousslavsky Chronic posttraumatic stress symptoms after nonsevere stroke Neurology, February 28, 2006; 66(4): 513 - 516. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. A. Hamann, T. J. Somers, A. W. Smith, S. S. Inslicht, and A. Baum Posttraumatic Stress Associated With Cancer History and BRCA1/2 Genetic Testing Psychosom Med, September 1, 2005; 67(5): 766 - 772. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Perkonigg, H. Pfister, M. B. Stein, M. Hofler, R. Lieb, A. Maercker, and H.-U. Wittchen Longitudinal Course of Posttraumatic Stress Disorder and Posttraumatic Stress Disorder Symptoms in a Community Sample of Adolescents and Young Adults Am J Psychiatry, July 1, 2005; 162(7): 1320 - 1327. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Lee, S. Joffe, H. T. Kim, G. Socie, A. L. Gilman, J. R. Wingard, M. M. Horowitz, D. Cella, and K. L. Syrjala Physicians' attitudes about quality-of-life issues in hematopoietic stem cell transplantation Blood, October 1, 2004; 104(7): 2194 - 2200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Gurevich, G. M. Devins, and G. M. Rodin Stress Response Syndromes and Cancer: Conceptual and Assessment Issues Psychosomatics, August 1, 2002; 43(4): 259 - 281. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
