This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Green, B. L. Articles by Stern, N. Search for Related Content PubMed PubMed Citation Articles by Green, B. L. Articles by Stern, N. Social Bookmarking                            What's this?

Trauma History as a Predictor of Psychologic Symptoms in Women With Breast Cancer

From the Department of Psychiatry, Georgetown University, Washington, DC.

Address reprint requests to Bonnie L. Green, PhD, Department of Psychiatry, Georgetown University, 310 Kober Cogan Hall, Washington, DC 20007; email greenb{at}gunet.georgetown.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To identify predictors of psychiatric problems in women with early-stage breast cancer.

PATIENTS AND METHODS: One hundred sixty women with early-stage breast cancer were recruited from three treatment centers. They filled out self-report questionnaires, including a medical history and demographic survey, the Trauma History Questionnaire, Life Event Questionnaire, Brief Symptom Inventory, Beck Depression Inventory, and Duke-UNC Functional Social Support Questionnaire, and were evaluated using the Structured Clinical Interview for DSM-III-R.

RESULTS: Hierarchical regression analyses indicated that four of five variable sets made a significant incremental contribution to outcome prediction, with 35% to 37% of the variance explained. Outcomes were predicted by demographic variables, trauma history variables, precancer psychiatric diagnosis, recent life events, and perceived social support. Cancer treatment variables did not predict outcome.

CONCLUSION: The findings highlight the important roles of trauma history and recent life events in adjustment to cancer and have implications for screening and treatment.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
A NUMBER OF MODELS, based on empirical studies, have been proposed to help identify risk factors for psychologic distress after treatment for breast cancer. These models are potentially useful for identifying survivors in possible need of psychologic intervention and for understanding the experiences and backgrounds associated with distress in order to design appropriate and efficacious treatments. Factors that have been investigated include medical aspects of the disease itself; medical treatment received; demographic factors such as age, education, and marital status; psychiatric history; and environmental supports.^1–5 Although not all of these factors have been associated with psychologic outcomes in each study, the studies, taken together, are useful for understanding vulnerability and risk across cancer populations. To date, studies of psychologic responses to cancer diagnosis and treatment have tended to focus on general anxiety and depression symptoms and have not examined either the impact of the patient’s trauma history on her adjustment to cancer or recent life events that may have taxed the patient’s capacity to cope with her illness.

Posttraumatic stress disorder⁶ (PTSD) has been well studied as a possible consequence of exposure to traditional stressors such as rape, disasters, or combat, and many of these studies have examined factors that predict the disorder and its associated symptoms.^7–10 The required conditions and symptoms of this disorder according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV)⁶ are listed in Table 1. Briefly, for a diagnosis of PTSD to be made, the patient must have experienced a traumatic event and, after the event, must exhibit symptoms of persistent re-experiencing of the trauma, persistent avoidance and numbing, and persistent arousal. Life-threatening illness was not designated a criterion A stressor event for PTSD (ie, was not mentioned as a "traumatic event") until the most recent edition of DSM (DSM-IV).⁶ However, cancer is often perceived as life-threatening, and many patients react to this diagnosis with fear or helplessness. In addition, many of the symptoms reported by cancer survivors are similar to the PTSD criteria. These include intrusive thoughts about cancer illness and treatment, including distress about the effects of treatment (eg, hair loss, premature menopause, disfigurement, or changes in interpersonal relationships); re-experiencing of aspects of the illness (eg, being told the diagnosis or receiving chemotherapy); deliberate avoidance of reminders of the treatment or disease, such as trying not to think or talk about the illness; having a sense of foreshortened future; and experiencing symptoms typical of hyperarousal, including difficulty sleeping and concentrating and feelings of nervousness, anxiety, and irritability. The kinds of conditioned responses often seen in PTSD related to other traumatic events (eg, rape, war, and disaster) are also common in cancer (eg, conditioned nausea and vomiting after chemotherapy and anticipatory anxiety around treatment, testing, and follow-up visits).

In a recent study assessing the usefulness in predicting PTSD of the Andersen model,¹ originally developed to predict general responses to cancer, Andrykowski and Cordova² evaluated 82 women using the PCL-C. They used variables summarized by Andersen as predicting distress (history of depression, education, social support, disease staging, physical comorbidity, and type of cancer treatment), as well as several variables that they proposed might explain additional variance (time since treatment, number of precancer stressors [or traumas], age at diagnosis, and taking of tamoxifen). Precancer stressors or traumas were assessed by asking an open-ended question about stressful or disturbing events that may cause a person to feel intense fear, helplessness, or horror, the question leading the PTSD section of the Structured Clinical Interview for DSM-III-R (SCID). With multiple regression analysis, their expanded model accounted for 52% of the variance in PTSD symptoms. In the best-fit model (evaluated with a stepwise procedure using all of the variables found significant in the earlier equations), significant predictors of PTSD symptoms were low social support, later disease stage at diagnosis, recency of treatment, and higher number of precancer traumas. Of note, disease stage was significant only in the shorter, best-fit model.

Examining a number of related predictor variables, Tjemsland et al¹⁴ studied 106 breast cancer patients before and 6 weeks after surgery. In their study, the Intrusion subscale of the Impact of Event Scale was used as a measure of PTSD symptomatology. In a multiple regression analysis, they found that a serious accident or illness or the death of a close relative in the past year, and elevated scores on the General Health Questionnaire presurgery, predicted high Impact of Event Scale intrusion scores. Age, marital status, breast cancer in a first-degree relative, social support before surgery, stage of cancer, and adjuvant treatment were not correlated with response. Using several instruments, they determined that approximately 8% of their sample met diagnostic criteria for PTSD. In these recent studies, little attention has been given to non-PTSD psychologic outcomes.

Studies of PTSD in response to other traumatic events have indicated that pre-event factors (such as demographics and prior trauma), aspects of exposure to the event, and postevent variables (such as social support and recent life event stressors) all contributed to current psychologic status, including PTSD.^15–17 Prior trauma, more severe exposure, and recent stressors are usually shown to be associated with more distress, whereas social support is associated with less disturbance. Green et al¹⁸ proposed that factors in the individual (eg, pretrauma personality and coping), along with aspects of the recovery environment (eg, social support, demographics, and additional stressors), affect the level and timing of adaptation to a traumatic event. These models and studies bear a strong similarity to those in the cancer literature predicting psychologic risk, mentioned earlier, and suggest a fair amount of overlap in models and risk factor studies in the two fields.

In the present study, we assessed PTSD symptoms by diagnostic interview, as well as general distress and depression symptoms, in a large sample of women with early-stage breast cancer and evaluated the role of a comprehensive set of risk factors in statistically predicting these outcomes. A particular focus was the extent to which trauma history variables and recent life events contribute to the prediction of current psychiatric status, the hypothesis being that more exposure to these adverse or stressful events would be associated with higher postcancer psychologic distress.

In an earlier report on the present sample,¹⁹ we described findings that were similar to those from other studies with regard to the PTSD diagnosis. Five percent of women met postcancer and 3% met current criteria for the cancer-related PTSD diagnosis by interview. However, PTSD symptoms were more common, especially when assessed for any time postcancer. Younger age predicted cancer-related PTSD, but type of treatment (mastectomy v lumpectomy, adjuvant chemotherapy) and time since diagnosis or end of treatment were not correlated with PTSD symptoms.

In the current study, we examined additional predictors of PTSD as well as other indicators of current distress. The potential predictors were tested in a hierarchic conceptual framework that would suggest possible ordering of the influences on early responses to breast cancer and its treatment. Results were seen as potentially useful both in recommending clinical strategies for evaluating and treating distressed breast cancer survivors and in addressing the nature of the potential link between PTSD, other mental health outcomes, and life-threatening illness.

	PATIENTS AND METHODS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Participants
A total of 160 women with invasive stage I or stage II (T1 or T2, N0, M0) node-negative breast cancer were recruited from three cancer treatment centers in the Washington, DC, area. Patients were identified from a variety of sources, including tumor registries, surgical schedules, pathology rosters, and medical oncology rosters. To be eligible for the study, participants had to be between 25 and 75 years of age, have no earlier history of cancer, and have finished cancer treatment in the previous 4 to 12 months (which ensured that subjects were no longer experiencing the acute physical effects of treatment but were close enough to diagnosis to recall reactions to illness and therapy). Women receiving ongoing oral hormonal therapy (tamoxifen) were eligible for the study, and psychiatric history was not an exclusion criterion.

We attempted, and believe we were able, to obtain the names of most of the eligible patients at participating centers during the study period. Of 264 women invited to participate, 160 (61%) were eventually interviewed. These 160 represented 65% of those who were reached and were eligible for the study. Among nonparticipants, 5% had a recurrence of breast cancer or had not finished cancer treatment in the time specified, making them ineligible, and another 13% could not be reached. The majority of nonparticipants (82%) declined either by returning our reply form (see Procedure) or by stating, when contacted by telephone, that they were too busy, would have difficulty getting to the hospital for an interview, were not interested, or were too distressed. Participants were younger (mean age [± SD], 53.40 ± 9.66 years) than nonparticipants (mean age, 60.41 ± 11.55 years), and African-American women were less likely than white women to participate (29% v 76%).

Instruments
A self-report questionnaire completed by each woman contained specific questions about the woman’s age, marital status, ethnicity, and education. Medical history questions developed for the study were asked about the type of breast surgery (mastectomy with or without reconstruction or lumpectomy), whether chemotherapy was part of the treatment, whether the subject was taking tamoxifen (Nolvadex; Zeneca Pharmaceuticals, Wilmington, DE), family history of cancer, the degree of involvement in the care of a relative with breast cancer, and the presence of any other serious medical condition. A number of standardized instruments were also administered. Each was selected because it had been used previously in other oncology or trauma studies.

The Trauma History Questionnaire (THQ)²⁰ is a self-report instrument that was developed at our center to gather data about a subject’s lifetime history of exposure to traumatic events, including crime, sexual assault or abuse, physical assault or abuse, war, tragic death, and disaster. Based on the high-magnitude–stressor interview from the DSM-IV PTSD field trials,²¹ the THQ has been used with several clinical and nonclinical samples and is currently undergoing revision. It has good test-retest reliability over a several-week period (r = .60 to 1.00 for individual, personally experienced events) and produces significant correlations with self-reported distress in college students and with measures of distress and functional status in clinic patients (for total events, as well as for physical assault or trauma and for sexual assault or trauma). Total and subscale scores also predicted PTSD in cocaine-dependent outpatients.²²

In the present study, responses to the THQ were used to rate extent of sexual and physical assault and abuse that the subjects reported in their backgrounds, as well as whether participants had had a life-threatening illness before developing cancer. Self-report responses were followed up during the face-to-face interview to gather additional details. Sexual trauma was assessed by three questions (eg, "Has anyone ever made you have intercourse or oral or anal sex against your will?"). These items were summed, so scores could range from 0 to 3. Physical trauma was assessed by four questions (eg, "Has anyone, including family members or friends, ever attacked you without a weapon and seriously injured you?"). These items were also summed, yielding scores of 0 to 4. Answers to the question about life-threatening illness were scored yes or no, and less serious illnesses were screened out.

In the Life Event Questionnaire,²³ subjects are asked to indicate which of 38 life events in seven categories (death of a loved one, change in a relationship, illnesses, work changes, and so on) occurred within the past year, excluding those related to the diagnosis and treatment of their breast cancer.

In the Brief Symptom Inventory (BSI),²⁴ a 52-item short form of the Symptom Checklist–90, the subject is asked to report symptom-related distress on a scale from 0 to 4. The BSI has nine symptom-specific scales and a global severity index (GSI). This measure has been used in numerous studies of PTSD and of cancer.

The Beck Depression Inventory (BDI)²⁵ is a 21-item self-report questionnaire that assesses how the patient has been feeling in the past week. Scores can range from 0 to 63, with scores of 10 to 18 indicating mild to moderate depression.

The Duke-UNC Functional Social Support Questionnaire (DSSQ)²⁶ is an eight-item self-report questionnaire that assesses perceived instrumental support (eg, help when the patient is sick in bed) and emotional support (eg, love and affection). Items are rated on a Likert-type scale, indicating the extent to which a particular type of social support occurs and ranging from 1 ("As much as I like") to 5 ("Much less than I would like").

After responses to the THQ were clarified and elaborated upon by the subjects, they completed a full SCID²⁷ for past and current axis I diagnoses, including PTSD, which was assessed with regard to both cancer and noncancer traumatic events. In the cancer-related PTSD module, the designated criterion A stressor event was "your [the patient’s] cancer diagnosis and treatment" and the subject was required to focus on her most stressful cancer-related experiences as elicited by a separate questionnaire. Because the study took place before the DSM-IV version of the interview had been finalized, we used the DSM-III-R version. Changes made in DSM-IV in PTSD symptoms were minor (ie, the symptom of physiologic distress at reminders was moved from the arousal section [D symptoms] to the re-experiencing section [B symptoms]), so all of the DSM-IV symptoms were assessed, and we were able to make the DSM-IV diagnosis. However, the A2 criterion (initial responses to the stressor in the form of fear, helplessness, or horror) was not part of the criteria in DSM-III-R, so we were not able to assess that criterion.

Interviewers
Interviewers had at least a bachelor’s degree and were female research assistants or graduate students. Training for the SCID consisted of a series of activities: viewing tapes created for training purposes by SCID developers, observing interviews conducted by the investigators who had previously been trained in the SCID, conducting role-play interviews with the investigators, and conducting interviews of subjects in the presence of the investigators. At each stage, cases were discussed with regard to interview format and scoring. A sample of 30 cases assessed by two raters (one conducting the interview and one observing) was used to assess interrater reliability. Kappa statistics were 0.79 for lifetime cancer-related PTSD, 1.00 for current cancer-related PTSD, and 0.84 for lifetime noncancer PTSD. The kappa statistic for any depression diagnosis was 0.77, and for any anxiety diagnosis it was 1.00. Given the low base rates of these disorders, the kappa statistics were considered adequate.

Procedure
After a potential subject was identified, a research nurse reviewed her record to confirm eligibility. With their doctors’ permission, women meeting inclusion criteria were sent a letter inviting them to participate in the study. Included was a reply form permitting them to decline without further contact. Those who responded affirmatively or did not return the reply form were followed up by telephone. At that time, the study was described and, if the woman agreed to participate, an appointment was scheduled. Interviews were usually conducted at the institution at which the subject received her cancer treatment. At the time of the interview, women provided written informed consent, completed the self-report measures, and were then interviewed. The entire procedure took from 90 minutes to 3 hours. Subjects were paid $30 for participation.

	RESULTS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
The mean age (± SD) of the sample was 53.40 ± 9.66 years (range, 26 to 75 years). Two thirds of the sample was white (66%), 27% was African-American, and 7% was Asian American or Latina. Fifty-eight percent of the women were married or living with partners, 19% were separated or divorced, and 23% were single or widowed. Only 19% had children younger than 18 years living with them. Most (64%) were college graduates, and the majority (67%) were employed at least part-time.

With regard to breast cancer treatment, 19% had undergone mastectomy alone and 19% had undergone mastectomy followed by adjuvant chemotherapy, for a total of 38% with mastectomy. Thirty-eight percent had had breast conservation (lumpectomy with radiation), whereas 24% had had conservation with chemotherapy, for a total of 62% with lumpectomy. Thus, 43% of the sample had received chemotherapy. Fifty-two percent were taking tamoxifen at the time of the interview. Women had finished cancer treatment an average of 6.5 months before the time of assessment.

As noted, 5% of the sample met postcancer (anytime since the cancer diagnosis) criteria for cancer-related PTSD when intrusive thoughts in addition to actual recollections of the event were considered to qualify for criterion B (re-experiencing). Only 3% met criteria for current (still active) PTSD. Postcancer major depression was diagnosed in 11% of the women. Six (4%) had their first major depressive episode after cancer was diagnosed and another 12 women had a recurrence of depression at this time. Half of the women with postcancer PTSD also met criteria for lifetime major depression, but none had another anxiety disorder. Seven percent of the women had substance abuse histories. According to SCID criteria, 12% of the sample had current psychiatric disorders.¹⁹

Participants reported an average of 1.18 current cancer-related PTSD symptoms on the SCID, with a range of 0 to 9, and an average of 2.26 postcancer symptoms (range, 0 to 13). The most common current symptom was sleep disturbance (18%), followed by sense of foreshortened future (12%), trouble concentrating (11%), and distress at exposure to reminders of illness (11%). The following symptoms were reported as current less than 5% of the time: intrusive recollections, traumatic dreams, flashbacks, avoiding thoughts of the event, amnesia for aspects of the event, diminished interest, detachment, restricted affect, and startle reaction. Intrusive thoughts about cancer in general (not recollections) were reported by 8% of the sample, and avoiding thoughts about cancer in general was reported by 10%. The mean (± SD) of the GSI of the BSI was 0.47 ± 0.4, decreasing at the 57th percentile for nonpsychiatric female patients. This score is comparable to the somewhat elevated scores found in other cancer survivor samples.^28,29 The mean score on the BDI was 6.32, below the range considered clinically significant. Approximately one quarter of the sample (23%) scored above 10, and 5% scored above 18.

With regard to predictor variables, 21% of the sample reported a prior life-threatening illness or medical event. Illnesses reported in this category included polio, toxemia of pregnancy, spinal meningitis, stroke, and ruptured aneurysm. When physical and sexual trauma variables were summed to derive two scores, it was determined that 26% of the sample reported some sexual trauma and 25% reported being physically assaulted. With regard to the woman’s prior experience with cancer, 36% of the sample reported that a first-degree relative had had cancer, and 10% stated that they were moderately or very involved in caring for a relative with breast cancer. Physical comorbidity was relatively high; 55% reported some comorbid illness, including heart disease, hypertension, diabetes, and arthritis. The total number of life events in the past year averaged 2.39 (SD = 2.33), and the total social support score on the DSSQ averaged 33.16 (SD = 6.00).

The predictor variables were conceptualized as falling into logical sets, loosely based on the model described previously,¹⁸ that could be tested for their hierarchic contribution to the prediction of outcome. The first set were demographic variables, including age, ethnicity, marital or partner status, and level of education. The second set were trauma history variables and included reported exposure to sexual trauma, physical trauma, and life-threatening illness. Historical variables, designated as the third set, comprised two types of predictors. These included number of precancer psychiatric diagnoses (primarily depression and anxiety) and prior experience with cancer (whether a first-degree relative had had cancer and the degree of personal involvement in caring for a relative with breast cancer). Cancer treatment made up the fourth variable set, which included type of surgery, whether the woman had chemotherapy, and whether she was taking tamoxifen. Finally, ongoing or current situational variables were entered, including any current physical comorbidity. Correlations among the outcome variables were moderate: number of current cancer-related PTSD symptoms was correlated 0.43 with GSI and 0.63 with the BDI. GSI and BDI (both self-report) were correlated 0.70.

The zero-order correlations between the predictor variables and the three measures of current symptom distress are listed in Table 2. In addition to these separate correlations, we have also included the R² (proportion of variance accounted for) for each set independently, not corrected for the other sets. A few general observations can be made about these relationships. First, the cancer treatment variables did not predict any outcome. Otherwise, at least one variable (usually more) in each set was associated with the outcome measures. Whereas being nonwhite and being less educated were associated with higher GSI scores, younger age was associated with higher BDI and PTSD symptom scores. Sexual and physical trauma were associated with all three measures. Although prior experience with a life-threatening illness was not associated with current distress, prior psychiatric diagnoses were associated with all three of the outcomes. Experiencing a greater number of recent life events and having lower social support predicted higher distress, as expected. However, having a comorbid physical illness was associated only with PTSD symptoms, and in a negative direction (ie, those with other medical problems were less likely to have cancer-related PTSD symptoms than were those without such difficulties). Overall, the historical variables, especially precancer psychiatric diagnosis, and current situational variables were the best predictors of present outcome.

	DISCUSSION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

Of note in this study is the fact that a reported history of traumatic event exposure played an important and consistent role in predicting psychologic outcomes. A number of studies have shown that prior traumatic life events increase the risk of PTSD associated with a new event.^17,32 Further, a study using an empirical model of PTSD related to combat found both direct and indirect (through later variables) effects of early trauma history on PTSD.³² Andrykowski and Cordova² found that the reported higher number of precancer stressors or traumas was predictive of more symptoms of PTSD, even with controlling for other variables, in women with cancer. Although we are not aware of any studies examining the impact of trauma history on non-PTSD outcomes associated with cancer diagnosis and treatment, in our study trauma predicted non-PTSD outcomes as well.

The events studied were those in which individuals were physically or sexually victimized by another person or persons. These events might have affected a participant’s adjustment to cancer in various ways. For example, prior victimization could be associated with lowered self-esteem or self-confidence, because of the experience of betrayal and/or self-blame that these assaults entail. Poor self-esteem, in turn, may be associated with less effective coping with the stress of a serious illness. These events also involved bodily violations, and, to the extent that the cancer is experienced as something that makes the body less than whole, less than healthy, or damaged in some way, it could recall prior feelings of being physically hurt or damaged. Finally, past interpersonal trauma may result in persistent feelings of vulnerability. The occurrence of cancer readily evokes and reinforces these feelings. Past trauma is also robustly associated with a variety of psychiatric diagnoses,³³ and indeed, the women with trauma histories in our study had more prior diagnoses (r = .16, P < .04; and r = .32, P < .001 for the presence of a precancer SCID diagnosis with physical and sexual trauma, respectively). However, in this study, both variables contributed independently to the prediction of outcomes, suggesting that there may be a number of pathways from exposure to interpersonal trauma and adjustment to cancer.

In addition to historical exposure to trauma, recent life events have also been shown to contribute to the prediction of outcome in a number of studies of adjustment to traumatic events.^34,35 Norris and Uhl³⁵ demonstrated that chronic stressors provided the external mechanism by which initial exposure was associated with adverse long-term outcomes after Hurricane Hugo in the United States. These studies suggest that continued or repeated exposure to stressful events may deplete an individual’s general capacity to cope successfully or to recover from an earlier stressor. They also support the general finding that cumulative exposure to serious life stressors and trauma is associated with increasing rates of depression and other disorders.³⁶

Having a precancer diagnosis of any psychiatric disorder was a strong predictor of psychologic outcome, in line with findings of other studies that psychiatric history is a significant predictor of adjustment after breast cancer³⁷ and other medical disorders.³⁸ Psychiatric diagnosis was less powerful as a predictor when trauma history was controlled for, however, suggesting that previous trauma is a risk factor for the earlier (precancer) diagnoses, and these, in turn, increase vulnerability for more negative outcomes when later stressors occur. These results, which are compatible with the cumulative adversity findings just noted, suggest the importance of assessing trauma exposure history in studies that attempt to examine how earlier vulnerability, defined by the presence of a psychiatric diagnosis, relates to dealing with recent stressors or coping ability. Investigators focusing only on earlier pathology may miss the role of prior trauma in a range of outcomes after medical illness.

Many studies have documented the salubrious effects of perceived social support. In the present study, the DSSQ was used to determine whether subjects considered the amount of support available to be as much as they wanted or needed. Perceived support was significantly related to all four outcome measures. We and others documented the association of social support with lower levels of symptoms after trauma^16,31,39 as well as cancer.^2,40,41

One finding that was not consistent with earlier findings was that physical comorbidity predicted fewer PTSD symptoms. Andrykowski and Cordova² found that physical comorbidity predicted higher PCL-C total scores in their initial model (but not in subsequent models). Possible explanations for the negative association found in our study are that subjects who had other physical problems may already have learned, at least to some extent, how to cope with medical illness and that the new illness (ie, cancer) may not have been as much of a shock to them as it might have been to a previously healthy person. This finding is unlikely to be an artifact due to the positive association between age and comorbid medical illness, because the final weights in the regression equation show that this relationship holds even with age partialled out. This particular association is probably complex and should be carefully evaluated in future studies.

Demographic findings were mixed with regard to support from previous studies. Education has been generally found to be negatively associated with distress,^1,42 although in this study it was negatively associated only with global severity on the BSI (general distress). For both PTSD and depression symptoms, younger age was the most potent demographic predictor of heightened distress. Whereas the diagnosis of depression is more common among younger women than older women,⁴³ PTSD is not associated with particular age groups among women in the general population.⁴⁴ However, a number of trauma studies have shown an association between younger age and PTSD (eg, combat⁹ and crime victimization⁴⁵). Several of the cancer and PTSD studies have also found that younger women are at higher risk for PTSD symptoms^12,42 and for more general distress as well.¹ As discussed,¹⁹ the present context may be important. Breast cancer is generally a disease of older women, so developing this disease at a younger age is unexpected. Further, cancer may be more aggressive and unpredictable in younger women, and treatment may lead to sterility. Younger women are also more likely to have children living at home and to be in the middle of careers. Hence, the disruption of and potential implications of developing the disease may be more far-reaching in this group.

The similarity of findings across different psychologic outcome measures, when considered in light of other study findings, raises the question of the specificity of the cancer-PTSD link. For example, none of the cancer-related variables—such as history of prior life-threatening illness, having a close relative with cancer, or being involved in that person’s care—was associated, either positively or negatively, with the outcome measures. Further, no aspect of treatment—surgery, prior chemotherapy, or current hormonal therapy status—predicted outcome. These latter findings echo those of other investigators,^2,11,14,42 who also did not find that treatment variables predicted PTSD-related outcomes. The exception to this pattern is the finding of Andrykowski and Cordova² that recency of treatment predicted higher current symptoms of PTSD. We did not find that length of time since the end of treatment predicted outcome,¹⁹ but our eligibility range of 4 to 12 months posttreatment may have been too restrictive to detect an impact. As previously noted, more than twice as many women in our sample developed major depression as developed PTSD after their cancer treatment (11% v 3% to 5%), with 4% experiencing their first major depressive episode postcancer, although current major depression was low as well (4%). Taken together, these findings suggest that PTSD symptoms after cancer are important but are perhaps only one aspect of distress. Some newer work^12,46 indicates that PTSD rates may be higher in bone marrow transplant recipients (13% to 19%). This increased risk associated with BMT suggests that some of the medical treatment procedures associated with cancer may serve as traumatic stressors in their own right or that more aggressive cancer in younger women⁴⁷ may be more likely to produce PTSD. It is likely that nonspecific distress would be higher in this group as well.

The present study raises a number of questions that should be addressed in future research. First, this model needs to be tested in a more diverse population. Although our study had a higher proportion of minorities than did earlier breast cancer–PTSD studies, the sample was relatively affluent. Further, a higher proportion of minorities declined to be enrolled, suggesting that those who did participate may have been less representative than the white cohort. Psychiatric disorders are more common in individuals of lower socioeconomic status.⁴⁸ Thus, it is important to evaluate these more vulnerable populations for risk factors and their relative importance in predicting psychologic responses to cancer. Further, the stage of cancer at which a woman is diagnosed may be correlated with socioeconomic status, placing less affluent women at particularly high risk for both medical and psychologic problems.

With regard to participation in the population studied, although our response rate (65%) was in the middle of those reported in other studies (42% to 81%),^{2,11,42,49–51} there was still a substantial minority who declined (35%). Some of the refusals were due to practical reasons, but others may have been made by more symptomatic women, especially avoiders, who may have had PTSD or other symptomatology. Further, we found that the participation rate was even lower in a follow-up study we attempted with women who had more advanced disease. Interview studies in general tend to have similar or lower (than our) participation rates (eg, 42% to 52%).^11,50,51 Andrykowski et al⁵² had higher rates (76%); however, they targeted women who had previously participated in research studies, a sample already selected for their interest in research participation. Given these sample constraints, the likelihood is that our rates of PTSD are conservative, because they are based on a more affluent population with earlier-stage illness, which probably reduced the range of outcome characteristics and attenuated, to some extent, the relationships found.

This study has a number of important clinical implications, both for assessment and for treatment of cancer survivors. First, it suggests the importance of screening for a variety of symptoms, including PTSD, when individuals are being treated for cancer.^52–55 Most women in our study were doing well, but a subset were distressed and could easily have been identified as such with a simple screening instrument. Sustained psychologic distress should not be viewed as a "normal" response to cancer diagnosis and treatment. A variety of professionals involved with the care of a patient, including oncologists, nurses, social workers, and chaplains, could help to assess the extent to which he or she might need psychologic help and could make a psychiatric referral. Women with histories of trauma and prior psychiatric diagnoses, as well as those who have recently experienced stressful events, may be at particular risk for problems adapting to illness, and assessment of these domains would be helpful in predicting who might need additional help. For these women, some attention to history might contribute positively to current adjustment to cancer. Even acknowledging and discussing the fact that the cancer might trigger emotions associated with prior unresolved events could help them understand their reactions. Younger women as well seem to be at particular risk and are likely to have a number of issues that are unique to their developmental stage, such as children still living at home, which would need special attention. Fortunately, useful screening tools for trauma history, PTSD, and general distress that can be readily incorporated into clinical practice, as well as guidelines for their use, are beginning to appear.^{13,19,52,53,55–58}

There are many potential interventions that could be brought to bear to help patients cope with an ongoing and potentially life-threatening illness. Studies have indicated that psychosocial interventions of a variety of types are helpful in reducing distress after cancer diagnosis and treatment.^59,60 The robustness of the finding that naturally occurring social support reduces distress suggests that patients should routinely be assessed to determine the quality and quantity of their social networks and that those who do not have good support may be particularly appropriate candidates for support groups. Educational interventions are also important to consider, given that a recent study indicated that they are effective and may be even more helpful than patient discussion groups in reducing distress.⁶¹ Although a recent meta-analysis indicated that screening for distress (ie, starting with more distressed patients) did not change effect sizes for the impact of interventions on outcomes,⁶⁰ cost considerations might dictate that patients with the risk factors discussed, and those with signs of psychologic distress, should be targeted first. Because psychologic distress can affect ability to comply with treatment, and, potentially, rate of recovery and survival,⁶² addressing psychologic concerns may not only improve quality of life for the patient, but ultimately be more efficient for the providers as well. Finally, although most comprehensive cancer centers are required to offer support services to patients and their families, the range and content of these vary.^63,64 Further, it is recognized that such services may not be routinely available in the larger community, where in the majority of cases of breast cancer the disease is diagnosed and treated. For physicians treating breast cancer patients in the latter setting, identifying appropriate cancer-related as well as ancillary mental health resources within their local areas, and mechanisms to take advantage of these resources, can be a critical first step in ensuring appropriate care for their cancer patients who are at risk for or are experiencing distress.

	ACKNOWLEDGMENTS

 
Supported by grant no. RO1 MH49742 from the National Institute of Mental Health, Bethesda, MD.

We acknowledge the contribution of Caryn Steakley, who reviewed all records to determine study eligibility. We also acknowledge the help and cooperation of physicians at Georgetown University Medical Center and the Lombardi Cancer Center; John Lynch, Lawrence S. Lessin, and other physicians at the Washington Hospital Center; and Karen Weihs and other physicians at George Washington University Medical Center.

	NOTES

 
Parts of this work were presented at the Second Annual Conference on Trauma, Loss, and Dissociation, Alexandria, VA, February 29-March 4, 1996, and the Lake George Research Conference on Posttraumatic Stress Disorder, Bolton Landing, NY, March 5-8, 1998.

	REFERENCES

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION REFERENCES

 
1. Andersen BL: Predicting sexual and psychologic morbidity and improving the quality of life for women with gynecologic cancer. Cancer 71:1678-1690, 1993 (suppl)[Medline]

2. Andrykowski MA, Cordova MJ: Factors associated with PTSD symptoms following treatment for breast cancer: Test of the Andersen model. J Trauma Stress 11:189-203, 1998[Medline]

3. Holland JC: Clinical course of cancer, in Holland JC, Rowland JH (eds): Handbook of Psychooncology. New York, NY,Oxford University Press, 1989, pp 75-100

4. Rowland JH, Massie J: Patient rehabilitation and support: Psychologic reactions to breast cancer diagnosis, treatment, and survival, in Harris JR, Lippman ME, Morrow M, et al (eds): Diseases of the Breast. Philadelphia, PA,Lippincott-Raven, 1996, pp 919-938

5. Weisman D: Early diagnosis of vulnerability in cancer patients. Am J Med Sci 271:187-196, 1976[Medline]

6. American Psychiatric Association:Diagnostic and Statistical Manual of Mental Disorders (ed 4). Washington, DC,American Psychiatric Association, 1994

7. Green BL: Psychosocial research in traumatic stress: An update. J Trauma Stress 7:341-362, 1994[Medline]

8. Kessler RC, Sonnega A, Bromet E, et al: Posttraumatic stress disorder in the National Comorbidity Survey. Psychiatry 52:1048-1060, 1995

9. Kulka RA, Schlenger WE, Fairbank JA, et al: Trauma and the Vietnam War Generation. New York, NY,Brunner/Mazel, 1990

10. Resnick HS, Kilpatrick DG, Dansky BS, et al: Prevalence of civilian trauma and posttraumatic stress disorder in a representative sample of women. J Consult Clin Psychol 61:984-991, 1993[Medline]

11. Alter CL, Pelcovitz D, Axelrod A, et al: The identification of PTSD in cancer survivors. Psychosomatics 37:137-143, 1996[Abstract/Free Full Text]

12. Jacobsen PB, Widows MR, Hann DM, et al: Posttraumatic stress disorder symptoms after bone marrow transplantation for breast cancer. Psychosom Med 60:366-371, 1998[Abstract/Free Full Text]

13. Weathers FW, Keane TM: Psychologic assessment of traumatized adults, in Saigh PA, Bremner JD (eds): Posttraumatic Stress Disorder: A Comprehensive Text. Boston, MA,Allyn & Bacon, 1999, pp 219-247

14. Tjemsland L, Soreide JA, Malt UF: Traumatic distress symptoms in early breast cancer: II. Outcome six weeks post surgery. Psychooncology 5:295-303, 1996

15. Fontana A, Schwartz LS, Rosenheck R: Posttraumatic stress disorder among female Vietnam veterans: A causal model of etiology. Am J Public Health 87:169-175, 1997[Abstract/Free Full Text]

16. Green BL, Grace MC, Lindy JD, et al: Risk factors for PTSD and other diagnoses in a general sample of Vietnam veterans. Psychiatry 147:729-733, 1990

17. Kemp A, Green BL, Hovanitz C, et al: Incidence and correlates of posttraumatic stress disorder in battered women: Shelter and community samples. J Interpersonal Violence 10:43-55, 1995

18. Green B, Wilson J, Lindy J: Conceptualizing post-traumatic stress disorder: A psychosocial framework, in Figley CR (ed): Trauma and Its Wake: The Study and Treatment of Post-Traumatic Stress Disorder. New York, NY,Brunner/Mazel, 1985, pp 53-69

19. Green BL, Rowland JH, Krupnick JL, et al: Prevalence of posttraumatic stress disorder (PTSD) in women with breast cancer. Psychosomatics 39:102-111, 1998[Abstract/Free Full Text]

20. Green BL: Trauma History Questionnaire, in Stamm BH, Varra EM (eds): Measurement of Stress, Trauma and Adaptation. Lutherville, MD,Sidran, 1996, pp 366-369

21. Kilpatrick DG, Resnick HS, Freedy JR, et al: Report of Findings From the DSM-IV PTSD Field Trial: Emphasis on Criterion A and Overall PTSD Diagnosis. Washington, DC,National Institute of Mental Health, 1992

22. Najavits LM, Gastfriend DR, Barber JP, et al: Cocaine dependence with and without PTSD among subjects in the National Institute on Drug Abuse Collaborative Cocaine Treatment Study. Am J Psychiatry 155:214-219, 1998[Abstract/Free Full Text]

23. Horowitz M, Schaefer C, Hiroto D, et al: Life event questionnaires for measuring presumptive stress. Psychosom Med 39:413-431, 1977[Abstract/Free Full Text]

24. Derogatis LR, Spencer PM: The Brief Symptom Inventory (BSI). Baltimore, MD,Clinical Psychometric Research, 1982

25. Beck AT, Steer RA: Beck Depression Inventory: Manual. San Antonio, TX,Psychological Corporation, 1987

26. Broadhead WE, Gehlbach SH, de Gruy FV, et al: The Duke-UNC Functional Social Support Questionnaire: Measurement of social support in family medicine patients. Med Care 26:709-723, 1988[Medline]

27. Spitzer RL, Williams JBW, Gibbon M, et al: Structured Clinical Interview for DSM-III-R: Non-Patient Edition (SCID-NP, Version 1.0). Washington, DC,American Psychiatric Press, 1990

28. Cella DF, Tross S: Psychological adjustment to survival from Hodgkin’s disease. Consult Clin Psychol 54:616-622, 1986

29. Rowland JH, Holland JC, Chaglassian T, et al: Psychological response to breast reconstruction: Expectations for and impact on postmastectomy functioning. Psychosomatics 34:241-250, 1993[Abstract/Free Full Text]

30. Cohen J, Cohen P: Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Hillsdale, NJ,Lawrence Erlbaum Associates, 1975

31. Green BL, Grace MC, Gleser GC: Identifying survivors at risk: Long-term impairment following the Beverly Hills Supper Club fire. Psychol 53:672-678, 1985

32. King D, King L, Foy D, et al: Prewar factors in combat-related posttraumatic stress disorder: Structural equation modeling with a national sample of female and male Vietnam veterans. J Consult Clin Psychol 64:520-531, 1996[Medline]

33. Breslau N, Davis GC, Andreski P, et al: Traumatic events and posttraumatic stress disorder in an urban population of young adults. Arch Gen Psychiatry 48:216-222, 1991[Abstract/Free Full Text]

34. Gleser GC, Green BL, Winget CN: Prolonged Psychosocial Effects of Disaster: A Study of Buffalo Creek. New York, NY,Academic Press, 1981

35. Norris F, Uhl G: Chronic stress as a mediator of acute stress: The case of Hurricane Hugo. J Appl Soc Psychol 23:1263-1284, 1993

36. Turner RJ, Lloyd DA: Lifetime traumas and mental health: The significance of cumulative adversity. J Health Soc Behav 36:360-376, 1995[Medline]

37. Maunsell E, Brisson J, Deschenes L: Psychological distress after initial treatment of breast cancer. Cancer 70:120-125, 1992[Medline]

38. Epstein SA, Krahn L, Clauw DJ, et al: Psychiatric aspects of the eosinophilia-myalgia syndrome. Psychosomatics 36:22-25, 1995[Abstract/Free Full Text]

39. Kaniasty K, Norris FH: A test of the social support deterioration model in the context of natural disaster. J Pers Soc Psychol 64:395-408, 1993[Medline]

40. Bloom JR: Social support, accommodation to stress and adjustment to breast cancer. Soc Sci Med 16:1328-1338, 1982

41. Ganz PA, Karin H, Sim M, et al: Predicting psychosocial risk in patients with breast cancer. Med Care 31:419-431, 1993[Medline]

42. Cordova MJ, Andrykowski MA, Kenady DE, et al: Frequency and correlates of posttraumatic stress disorder like symptoms after treatment for breast cancer. J Consult Clin Psychol 63:981-986, 1995[Medline]

43. Kessler R, McGonagle K, Zhao S, et al: Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Arch Gen Psychiatry 51:8-19, 1994[Abstract/Free Full Text]

44. Kessler RC, Sonnega A, Bromet E, et al: Epidemiological risk factors for trauma and PTSD, in Yehuda R (ed): Risk Factors for Posttraumatic Stress Disorder. Washington, DC,American Psychiatric Press, 1999, pp 23-59

45. Kilpatrick D, Saunders B, Amick-McMullan A, et al: Victim and crime factors associated with the development of crime-related post-traumatic stress disorder. Behav Ther 20:199-214, 1989

46. DuHamel K, Redd W, Smith M, et al: PTSD in bone marrow transplant survivors. Presented at the meeting of the Society of Behavioral Medicine, Washington, DC, March 13-16, 1996

47. Spinelli C, Ricci E, Berti P, et al: Relationship between five-year survival and age at diagnosis in breast cancer: Multivariate analysis. Breast Dis 8:209-216, 1995

48. Holzer CE, Shea BM, Swanson JW, et al: The increased risk for specific psychiatric disorders among persons of low socioeconomic status: Evidence from the epidemiologic catchment area surveys. Am J Soc Psychiatry 6:259-271, 1986

49. Ganz PA, Rowland JH, Desmond K, et al: Life after breast cancer: Understanding women’s health-related quality of life and sexual functioning. Oncol 16:501-514, 1998

50. Ganz PA, Schag CAC, Lee JJ, et al: Breast conservation versus mastectomy: Is there a difference in psychological adjustment or quality of life in the year after surgery? Cancer 69:1729-1738, 1992[Medline]

51. Taylor SE, Lichtman RR, Wood JV, et al: Illness-related and treatment-related factors in psychological adjustment to breast cancer. Cancer 55:2506-2513, 1985[Medline]

52. Andrykowski M, Cordova M, Studts JL, et al: Posttraumatic stress disorder after treatment for breast cancer: Prevalence of diagnosis and use of the PTSD Checklist-Civilian Version (PCL-C) as a screening instrument. Psychol 66:586-590, 1998

53. Holland JC: Preliminary guidelines for the treatment of distress. (Huntingt) 11:109-114, 1997

54. Kelly B, Raphael B, Smithers M, et al: Psychological responses to malignant melanoma: An investigation of traumatic stress reactions to life-threatening illness. Hosp Psychiatry 17:126-134, 1995

55. Zabora JR: Screening procedures for psychosocial distress, in Holland JC (ed): Psycho-Oncology. New York, NY,Oxford University Press, 1998, pp 653-661

56. Payne DK, Hoffman RG, Theodoulou M, et al: Screening for anxiety and depression in women with breast cancer: Psychiatry and medical oncology gear up for managed care. Psychosomatics 40:64-69, 1999[Abstract/Free Full Text]

57. Roth AJ, Kornblith AB, Batel-Copel L, et al: Rapid screening for psychologic distress in men with prostate carcinoma: A pilot study. Cancer 82:1904-1908, 1998[Medline]

58. Goodman LA, Corcoran C, Turner K, et al: Assessing traumatic event exposure: General issues and preliminary findings for the Stressful Life Events Screening Questionnaire. J Trauma Stress 11:521-542, 1998[Medline]

59. Devine EC, Westlake SK: The effects of psychoeducational care provided to adults with cancer: Meta-analysis of 116 studies. Oncol Nurs Forum 22:1369-1381, 1995[Medline]

60. Meyer TJ, Mark MM: Effects of psychosocial interventions with adult cancer patients: A meta-analysis of randomized experiments. Health Psychol 14:101-108, 1995[Medline]

61. Helgeson VS, Cohen S, Schulz R, et al: Education and peer discussion group interventions and adjustment to breast cancer. Arch Gen Psychiatry 56:340-347, 1999[Abstract/Free Full Text]

62. Richardson J, Shelton D, Krailo M, et al: The effect of compliance with treatment on survival among patients with hematologic malignancies. J Clin Oncol 8:356-364, 1990[Abstract]

63. Presberg BA, Levenson JL: A survey of cancer support groups provided by National Cancer Institute (NCI) clinical and comprehensive centers. Psychooncology 2:251-217, 1993

64. Coluzzi PH, Grant M, Doroshow JH, et al: Survey of the provision of supportive care services at National Cancer Institute-designated cancer centers. J Clin Oncol 13:756-764, 1995[Abstract/Free Full Text]

Submitted July 8, 1999; accepted October 15, 1999.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

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]

				K.-A. Phillips, R. H. Osborne, G. G. Giles, G. S. Dite, C. Apicella, J. L. Hopper, and R. L. Milne Psychosocial Factors and Survival of Young Women With Breast Cancer: A Population-Based Prospective Cohort Study J. Clin. Oncol., October 1, 2008; 26(28): 4666 - 4671. [Abstract] [Full Text] [PDF]

				B. L. Green, J. Y. Chung, A. Daroowalla, S. Kaltman, and C. DeBenedictis Evaluating the Cultural Validity of the Stressful Life Events Screening Questionnaire Violence Against Women, December 1, 2006; 12(12): 1191 - 1213. [Abstract] [PDF]

				Y. Matsuoka, M. Inagaki, Y. Sugawara, S. Imoto, T. Akechi, and Y. Uchitomi Biomedical and Psychosocial Determinants of Intrusive Recollections in Breast Cancer Survivors Psychosomatics, June 1, 2005; 46(3): 203 - 211. [Abstract] [Full Text] [PDF]

				M. Okamura, S. Yamawaki, T. Akechi, K. Taniguchi, and Y. Uchitomi Psychiatric Disorders Following First Breast Cancer Recurrence: Prevalence, Associated Factors and Relationship to Quality of Life Jpn. J. Clin. Oncol., June 1, 2005; 35(6): 302 - 309. [Abstract] [Full Text] [PDF]

				G. E. Wyatt, T. B. Loeb, K. A. Desmond, and P. A. Ganz Does a History of Childhood Sexual Abuse Affect Sexual Outcomes in Breast Cancer Survivors? J. Clin. Oncol., February 20, 2005; 23(6): 1261 - 1269. [Abstract] [Full Text] [PDF]

				S. C. Palmer, A. Kagee, J. C. Coyne, and A. DeMichele Experience of Trauma, Distress, and Posttraumatic Stress Disorder Among Breast Cancer Patients Psychosom Med, March 1, 2004; 66(2): 258 - 264. [Abstract] [Full Text] [PDF]

				M. M. Clark, J. M. Bostwick, and T. A. Rummans Group and Individual Treatment Strategies for Distress in Cancer Patients Mayo Clin. Proc., December 1, 2003; 78(12): 1538 - 1543. [Abstract] [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]

				M. Stefanek, L. Hartmann, and W. Nelson Risk-Reduction Mastectomy: Clinical Issues and Research Needs J Natl Cancer Inst, September 5, 2001; 93(17): 1297 - 1297. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Green, B. L. Articles by Stern, N. Search for Related Content PubMed PubMed Citation Articles by Green, B. L. Articles by Stern, N. Social Bookmarking                            What's this?