This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dougall, A. L. Articles by Baum, A. Search for Related Content PubMed PubMed Citation Articles by Dougall, A. L. Articles by Baum, A. Related Collections PTSD

Predictors of Posttraumatic Stress Among Victims of Motor Vehicle Accidents

From the University of Pittsburgh (A.L.D., D.M.P., A.B.), Pittsburgh, Pennsylvania; and the Uniformed Services University of the Health Sciences (R.J.U., C.S.F.) Bethesda, Maryland.

Address requests for reprints to: Andrew Baum, PhD, University of Pittsburgh Cancer Institute, Department of Behavioral Medicine and Oncology, 3600 Forbes Ave., Suite 405, Pittsburgh, PA 15213. Email: baum{at}pcicirs.pci.pitt.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
OBJECTIVE: This study identified factors that predict individual vulnerability to psychological trauma by examining the relationships among situation and person variables and symptoms of posttraumatic stress disorder (PTSD) 1, 6, and 12 months after a serious motor vehicle accident (MVA).

METHODS: Background characteristics, exposure variables (ie, injury severity and accident characteristics), and psychosocial variables (ie, perceived loss of control, social support, and coping) were used to predict symptoms of PTSD and recovery in 115 injured MVA victims. All participants were injured during the MVA and provided data prospectively over the course of a year after their accidents.

RESULTS: Along with background and exposure variables, use of wishful thinking coping distinguished between victims with and without symptoms of PTSD.

CONCLUSIONS: Psychosocial variables such as wishful thinking coping can be used to identify MVA victims who are at risk of developing chronic posttraumatic stress and warrant further investigation.

Key Words: posttraumatic stress disorder • social support • coping • stress • trauma.

Abbreviations: ANOVA = analysis of variance; DFA = discriminant function analysis; DSM-III-R = Diagnostic and Statistical Manual of Mental Disorders, third edition revised; MANOVA = multivariate analysis of variance; MVA = motor vehicle accident; PTSD = posttraumatic stress disorder; SCID = Structured Clinical Interview for DSM-III-R; SES = socioeconomic status.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Posttraumatic stress disorder (PTSD) is a profound outcome of psychological trauma that can result in severe disability across several domains of functioning (1–3). Of those exposed to traumatic events, a substantial number develop PTSD, but most do not. Instead, most people exhibit relatively rapid adaptation and recovery after the stressor. Selective vulnerability to the imperiling features of traumatic events has drawn considerable attention to the search for predictors of PTSD or susceptibility to other effects of psychological trauma. The study reported in this article systematically evaluated several predictors of PTSD symptoms among men and women involved in serious MVAs. The contribution of modifiable psychosocial variables to the prediction of posttraumatic stress was examined, as was the contribution of more static risk factors such as demographic and exposure variables.

The diagnosis of PTSD requires exposure to a traumatic event and persistent symptoms of reexperiencing the traumatic event, avoidance of trauma-related stimuli, emotional withdrawal or numbing, and heightened physiological arousal (1). In general, more frequent or prolonged exposure and greater intensity of exposure to violence, injuries, or death are associated with posttraumatic stress (4–8). However, not all studies find effects for all of these exposure variables (5, 8), and their influence on symptoms of posttraumatic stress seem to weaken over time (8). Most trauma victims do not develop PTSD even when exposures are considered to be severe (2), suggesting that other vulnerability factors affect the emergence of and recovery from PTSD symptoms. However, the sheer number of factors that might affect how traumatic events are experienced limits vulnerability analyses and typically cannot simultaneously account for variance due to exposure, person variables (including ethnicity, gender, SES, and education), and psychosocial variables (2, 3). Women and ethnic minorities are at a higher risk for PTSD than men and non-Hispanic whites (3, 5, 9–11). Additionally, low SES and less education are associated with a higher risk for developing posttraumatic stress (12–14).

MVAs were used as a model for psychological trauma. MVAs are among the most common traumatic events, affecting approximately 23% of adults in their lifetime (3). Previous studies of vulnerability factors have examined the predictive value of exposure and demographic risk factors to the development of MVA-related PTSD, predicting up to 70% of PTSD diagnoses in a sample of injured MVA victims 1 to 4 months after their accidents (15). Fear of dying in the MVA, injury severity, litigation, and prior diagnosis of major depression predicted PTSD outcomes. In another report, Blanchard et al. (4) were able to classify 6-month recovery from PTSD in 84% of their sample of injured MVA victims using initial PTSD symptoms, injury severity, recovery from injuries, and the occurrence of a new trauma to a family member as predictors. Ehlers et al. (5) expanded on this work, showing that psychological maintaining factors such as rumination, negative interpretations of intrusions, thought suppression, and anger cognitions improved the prediction of PTSD 1 year after a MVA.

This last finding is particularly important because these psychological maintaining factors are potentially modifiable. Background and person variables that predict risk for posttraumatic stress are typically not as amenable to change. These psychosocial risk factors have not been adequately studied in investigations of the emergence and maintenance of posttraumatic stress after exposure to MVAs. Research with victims of other traumatic stressors suggests that less perceived control and available social support affect risk for PTSD (12, 16–19). Use of avoidance and other emotion-focused coping strategies are also associated with more posttraumatic stress (16, 20–22).

The study reported here investigated whether psychosocial variables (such as perceived control, social support, and coping) can identify MVA victims who will or will not develop acute and chronic symptoms of PTSD. All participants were injured during a serious MVA and provided data prospectively over the course of 1 year after their accidents. Previously we reported that 34% of the MVA victims at 1 month, 18% at 6 months, and 14% at 12 months met threshold criteria for diagnoses of PTSD (23). Additionally, women, ethnic minorities, and victims reporting less education had more threshold diagnoses than did men, non-Hispanic whites, and individuals with more education (23).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Overview
A total of 115 injured MVA victims participated in four assessments of predisposing factors and three assessments of PTSD. The PTSD criterion variable was redefined in the present study to capture the expression of PTSD symptoms rather than clinical diagnoses. This was accomplished by combining those victims with threshold and subthreshold PTSD diagnoses and comparing them with victims with absent diagnoses, representing a cohort with no or few PTSD symptoms. The aim of the present study was to assess the contribution of modifiable psychosocial variables (ie, perceptions of control, availability of social support, and use of coping strategies) to the prediction of posttraumatic stress above and beyond that of background and exposure variables like those discussed above. These psychosocial variables were expected to improve the accuracy of the prediction of PTSD symptoms and recovery from background and exposure variables alone.

Participants
Victims of MVAs were recruited from either a regional trauma center in a large middle Atlantic metropolitan area or local police reports of MVAs. Individuals were considered for study participation if they were taken and admitted to the trauma center or another area emergency room or trauma unit. Most were sampled from a regional trauma center, but some were identified from public accident reports. Two physicians systematically reviewed MVA admissions to the trauma center and excluded patients if there was evidence of organic brain syndrome. A complex sampling scheme was designed to rotate periods at which all consecutive admissions who met eligibility criteria were asked to participate. Blocks of time were rotated each day so that admissions at 11 AM on a weekday were as likely to be approached as admissions at 11 PM on a Saturday. Experimenters contacted all potential participants after their discharge from the hospital, briefly described the study, and set up an appointment. A total of 122 MVA victims, approximately 44% of all eligible candidates, agreed to participate. Participation rates were 50% for the hospital sample and 25% for a smaller group recruited from police reports. There were no differences in demographic or accident variables between those who did and did not agree to participate or between participants drawn from the two recruitment sources. The final sample consisted of 115 MVA victims (53 women and 62 men). The sample consisted of participants who identified themselves as white (72%), African American (16%), Latin American (8%), Asian American (2%), or from another ethnic group (2%), reflecting the ethnic group composition of the surrounding communities. Participants ranged in age from 18 to 64 years with an average age of 35 years (SD = 13 years).

Participants completed four sessions over the course of the year following their MVAs, first at 2 to 3 weeks and then at 3, 6, and 12 months after the accident. The SCID (24) was administered at a separate session by one of two trained clinical social workers. The interview was administered approximately 1 week after the first, third, and fourth assessments (at 1, 6, and 12 months, respectively) and allowed examination of the onset and progression of PTSD. The first interview occurred an average of 34 days after the MVA (SD = 8, range = 26–71 days). About one-third of the sample failed to complete all three interviews. Noncompleters were comparable to completers in most ways, except they tended to be less educated and younger than those who completed all assessments. Completers and noncompleters did not differ in symptoms reported or percentage who met DSM criteria for PTSD with and without controlling for education.

Procedure
Participants were seen in their homes for each of the four assessments. The protocol followed for each session was the same, except at the first session the study was described in detail and informed consent was obtained. Instructions for the questionnaires were given, and arrangements were made for the experimenter to retrieve the completed questionnaires the next day. Participants were compensated $20 for their time and cooperation at each assessment.

Measures
Subjects completed several inventories that provided information about their background as well as their perceptions of the accident. Intervening factors were also measured, and the presence of PTSD symptomatology was established through the use of a clinical interview.

Symptoms of PTSD.
Victims with threshold and subthreshold diagnoses of PTSD were combined into one symptom group and were compared with victims in the "absent" group, who had no diagnoses of PTSD. The presence of PTSD was assessed through the administration of the PTSD supplement to the SCID (24, 25). The SCID is a semistructured clinical interview designed to accurately diagnose adult patients or research participants; the interview is based on DSM-III-R (26). Interviewers were trained to 90% interrater reliability. All interview sessions were audiotaped, and the supervising psychiatrist (R.J.U.) reviewed half of these tapes on an ongoing basis to ensure agreement. Diagnoses were determined by consensus of the interviewers and the psychiatrist (R.J.U.).

To receive a threshold diagnosis of PTSD, all five diagnostic criteria had to be met. Subthreshold diagnoses were coded on the basis of the following definition: criteria A (a traumatic event) and E (duration of 1 month or longer) and at least two of criteria B, C, and D (reexperiencing, avoidance, and hyperarousal symptoms, respectively) had to be met (cf, Refs. 15, 27, and 28). In this sample, 27 of 29 participants who met subthreshold criteria at 1 month did not meet criterion C (emotional numbing and avoidance). The other two participants did not meet criterion D (hyperarousal). At 6 months all 25 participants who met subthreshold criteria met criteria B and D but not criterion C. Likewise, at 12 months 15 of 16 participants with subthreshold diagnoses did not meet criterion C. The one other participant did not meet criterion B (reexperiencing). Any diagnosis not meeting criteria for threshold or subthreshold PTSD was coded as absent.

Exposure.
Exposure was assessed by examining information regarding specific characteristics of the MVA that was obtained during the first session as well as the extent of injuries incurred and victims’ perceptions of threat during their MVAs. Information on injuries was obtained from medical and police reports and was confirmed during a brief interview at the first session during which the MVA victims described their accidents. Injury severity was assessed by two independent raters (interrater reliability = 0.99) using the Injury Severity Score from the Abbreviated Injury Scale developed by the Association for the Advancement of Automotive Medicine (29). Scores can range from 1 to 75, although scores in the upper range of the scale reflect severe injuries that are often untreatable.

Perceived threat was measured at 3 months by asking participants to indicate on a five-point scale how strongly they agreed with the statement that the accident was very threatening (30). Higher scores on this item reflected more perceived threat.

Psychosocial variables.
Potential variables that could affect the relationship between trauma exposure and PTSD symptomatology included availability of social support, perceptions of personal control, and the types of coping strategies used to deal with the MVA. Social support was assessed by measuring perceptions of emotional support using a six-item questionnaire (Cronbach’s = 0.82, test-retest reliability = 0.70; see Ref. 31). Each item is scored on a seven-point Likert scale, and total scores range from 6 to 42 with higher scores indicating more emotional support. Perceived control was assessed by a single item in the same questionnaire as the social support items (31). The item asked the participant to indicate on a seven-point scale to what extent they thought "one can control what happens to him/her."

Coping with the accident was assessed using the Ways of Coping Inventory (30). This scale measured the frequency of 69 types of coping behaviors. Five scales derived from previous factor analyses were used in the current analyses: problem-focused coping, wishful thinking, seeking social support, self-blame, and avoidance coping (32). Scores on each coping scale were adjusted for the total amount of coping used by calculating relative scores (33). Although the Ways of Coping Inventory was administered at every time point, participants were instructed to answer with regard to their MVA only at the second and fourth assessments. At the first and third assessments, participants were instructed to complete the inventories while considering any stressful event of their own choosing (this allowed evaluation of more general styles of coping). Because the focus of the present report is determinants of stress after an MVA, only responses from the second and fourth assessments were considered.

Data Analysis
Distributions of the predictor variables were examined to determine normality. Reports of perceived emotional support were negatively skewed, and raw data for this scale were squared. Repeated-measures ANOVAs and MANOVAs were conducted to evaluate changes in the psychosocial predictor variables over time. The relationships between individual predictor variables and the PTSD symptom groups (0 = absent group, 1 = symptom group) at each time point were determined by computing correlation coefficients. The relationships between individual predictors and the recovery groups (0 = maintained, 1 = recovered, 2 = absent) were determined using ² analyses for categorical predictors and ANOVAs orMANOVAs for continuous predictors. These analyses were conducted to determine which variables should be eliminated from the final logistic regression model and the final DFA model to predict the symptom and recovery groupings, respectively. The logistic regression and DFA equations were sequential with background variables entered on the first step, followed by exposure variables, and then psychosocial variables. The models were evaluated at each step by examining the overall number of cases correctly classified and the correct number of victims predicted to belong to each group. The predictive value of each variable was determined in the logistic regressions by examining its regression coefficient and odds ratio and determined in the DFA by examining its pooled within-groups correlation with the standardized canonical discriminant functions.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Background, exposure, and psychosocial variables were first evaluated as predictors of PTSD symptoms at 1, 6, and 12 months. This same set of predictors was then used to predict recovery of PTSD symptoms over the entire year.

Symptoms and Symptom Groups
Diagnostic interviews were completed for 108 victims at 1 month, 86 victims at 6 months, and 75 victims at 12 months. Victims with subthreshold or threshold diagnoses of PTSD were combined into one symptom group (coded as 1) and compared with an absent group (coded as 0) consisting of victims without diagnoses of PTSD. Three sets of symptom and absent groups were created, one for each time point. More than 61% of the MVA victims at 1 month (N = 66), 45% at 6 months (N = 39), and 37% at 12 months (N = 28) were classified into the symptom group. The majority of victims in the symptom group remained in this group at 6 and 12 months (63% and 53%, respectively). An additional five victims exhibited delayed onset of symptoms at 6 months, as did three more at 12 months.

Correlational Analyses
Before testing the models predicting PTSD symptoms at 1, 6, and 12 months, the extent to which each background, exposure, and psychosocial variable was associated with the PTSD symptom groups was assessed. Individual predictors that were significantly associated with the symptom groups were then used in the final regression models to predict group membership. Table 1 presents these correlational analyses. Because the psychosocial variables (ie, perceived availability of social support, perceived control, and use of coping strategies) did not change over the course of the study, values for these predictors at the first time they were measured were used in these analyses. Additionally, because coping was not assessed until 3 months, relative coping scores at 3 months were used to predict group membership at 6 and 12 months but not at 1 month.

Exposure.
Contrary to expectations, injury severity was negatively associated with group membership at 1 month, indicating that victims in the symptom group actually had less severe injuries (see Table 1). No relationship between injury severity and group membership was found at either 6 or 12 months. As expected, more perceived threat was positively associated with group membership at all three assessments, indicating that victims in the symptom group reported more perceived threat (see Table 1).

Group differences in specific characteristics of the MVA, such as weather conditions, trip destination and origin, number and type of vehicles involved and damage to each, whether there were other passengers in the participant’s vehicle and whether the passengers were injured, and whether passengers were wearing seat belts, were also examined. With the exception of the presence of other passengers, these accident characteristics did not significantly distinguish between the participants in the symptom and absent groups and were not included in further analyses. Victims who reported having the presence of other passengers during the MVA were more likely to belong to the symptom group at 6 and 12 months (see Table 1). Presence of other passengers was positively associated with perceived threat (r = 0.24, p < .05), suggesting that more threat was perceived when others were involved in the accident than when the victim was alone.

Psychosocial variables.
As expected and shown in Table 1, perceived availability of social support and group membership were negatively related at both 6 and 12 months, indicating that belonging to the absent group at 6 and 12 months was associated with perceiving more available social support 2 to 3 weeks after the MVA. Unlike support, perceptions of control were not related to expression of PTSD symptoms and were not used in subsequent analyses to predict symptom vs. absent group membership. As expected, use of wishful thinking coping at 3 months and symptom group membership at 6 and 12 months were positively correlated (see Table 1). Victims in the symptom group reported more use of wishful thinking coping than did victims in the absent group. Use of seeking social support coping at 3 months and group membership at 6 months were negatively associated (see Table 1), indicating that victims in the symptom group reported less use of seeking social support coping than did victims in the absent group. The symptom vs. absent grouping was not associated with reported use of problem-focused, self-blame, or avoidance coping. Therefore, wishful thinking coping and seeking social support coping were used to predict group membership at 6 months, and wishful thinking coping was used to predict group membership at 12 months.

Logistic Regression Analyses
The results of the analyses reported above identified variables that were related to symptom and absent group membership and facilitated development of sequential logistic regression equations to predict group membership. Because there were no differences over time for social support or coping, values of these variables at the first time they were assessed were used as predictors.

One month.
At 1 month group membership was predicted by gender of the participant, followed by injury severity and perceived threat. Information regarding the gender of the MVA victim significantly differentiated between the symptom and absent groups at 1 month, (²(1,82) = 6.58, p < .01). With gender in the equation, all of the victims were predicted to belong to the symptom group, resulting in an overall success rate of 65%. Discrimination between the groups was further improved when injury severity and perceived threat were entered into the equation (²(2,82) = 8.94, p < .05). With this additional information, the overall correct classification rate rose to 77%, with 89% of the symptom group correctly classified and 55% of the absent group correctly classified. Table 2 shows regression coefficients, Wald statistics, odds ratios, and 95% confidence intervals for odds ratios for each of the predictors on the last step of the equation. Only gender and perceived threat had significant unique effects. Therefore, being a woman and having high perceived threat were associated with increased odds of belonging to the symptom group at 1 month.

Recovery
Seventy-five participants completed all three PTSD diagnostic interviews over the course of the study. Recovery was defined as a change in a threshold diagnosis or a subthreshold diagnosis to an absent diagnosis. By 6 months 30%, or 13 of the 43 victims reporting threshold or subthreshold PTSD at 1 month, showed evidence of recovery that was maintained at 12 months (recovered). By 12 months an additional eight victims (18%) had recovered. The remaining 22 victims presented with chronic posttraumatic stress (maintained), and another 22 victims received absent diagnoses at all three time points (absent). Ten victims did not fit into the three categories and were omitted from the present analyses.

As expected, male victims were the majority in the absent group (68%) and female victims were the majority in both the maintained (73%) and the recovered groups (62%) (²(2,65) = 7.98, p < .05). Additionally, the maintained group reported more perceived threat (mean = 2.82, SD = 1.37) than the absent group (mean = 1.21, SD = 0.27) (F(2,58) = 3.79, p < .05). However, the recovered group (mean = 2.05, SD = 1.31) did not differ in perceptions of threat from either the maintained or absent group. The recovered group and the absent group were also more likely to be by themselves during the accident than the maintained group (²(2,63) = 7.26, p < .05). A repeated-measures MANOVA revealed that use of coping strategies differed across the groups (F(10, 94) = 2.94, p < .01) with no time or group-by-time effects (F(5, 48) = 1.38, p < .25 and F(10,94) = 0.98, p < .47, respectively). At both 3 and 12 months the maintained group reported more relative use of wishful thinking coping (mean = 0.25, SE = 0.02) than the recovered (mean = 0.16, SE = 0.02) and the absent groups (mean = 0.12, SE = 0.02) (F(2,52) = 15.01, p < .001). However, relative use of wishful thinking coping did not differ between the recovered group and the absent group.

The three recovery groups were comparable with regard to ethnic composition, education, income, age, marital status, injury severity, perceived control, available social support, and relative use of problem-focused coping, seeking social support coping, self-blame coping, and avoidance coping. Gender, perceived threat, and presence of other passengers predicted recovery group membership. The addition of wishful thinking coping further improved classification. These four predictors correctly classified 59% of the participants, compared with 33% correctly classified by chance alone (Wilks’ = 0.56, F(8, 106) = 4.50, p < .001). The first discriminant function accounted for 95% (eigenvalue = 0.73) of the between-group variability in discriminating among recovery groups and separated the maintained group from the recovered and absent groups (Wilks’ = 0.56, ²(8,59) = 31.85, p < .001). The primary predictors for this function were the relative use of wishful thinking coping (0.87), perceived threat (0.43), and gender (0.40). These variables provided information that led to the correct classification of more than two-thirds of the maintained (71%) and absent (68%) groups. Prediction of the recovered group (37%) was unimpressive.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
This study identified factors associated with vulnerability to and recovery from trauma by examining differences in situation and person variables among MVA victims with and without symptoms of PTSD. As previously found in other samples of MVA victims, background and exposure variables predicted manifestation and recovery of symptoms of PTSD (4, 5, 15). We extended these findings by showing that psychosocial variables, such as coping, are also related to symptoms of PTSD and can be used to identify the occurrence of posttraumatic stress as well as its continued maintenance more than a year after an MVA. Because psychosocial variables are more readily modified than background and exposure variables, it is important to identify these risk factors as potential targets for intervention.

Between 72% and 77% of the participants were correctly classified into the symptom or absent group. Of the background variables, gender uniquely predicted group membership at 1 month but did not have unique effects at 6 and 12 months. Perceived threat and presence of other passengers were the two exposure variables that were consistently related to symptom and absent group membership. Of the psychosocial variables, wishful thinking coping was the only variable to uniquely predict group membership. Gender, perceived threat, and use of wishful thinking coping also predicted recovery at 12 months.

The classification rates for the symptom and absent groups (72–77%) were comparable to the 70% reported by Blanchard et al. (15) in a similar sample of MVA victims 1 to 4 months after their accidents. However, one of the four major predictors of PTSD in their sample (ie, injury severity) was negatively correlated with initial PTSD symptomatology in the present sample and did not add significant information to the discrimination between the symptom and absent groups. Additionally, Blanchard et al. assessed the contribution of prior mental health and accident characteristics to the prediction of acute PTSD. Attention in the present study was primarily focused on the contribution of traditional psychosocial mediators to the prediction of acute and chronic PTSD symptomatology. These psychosocial variables provided additional information on modifiable risk factors for the presentation and maintenance of posttraumatic stress.

Prediction of recovery was less impressive. The background, exposure, and psychosocial variables examined in the present study were poor predictors of which victims belonged to the recovered group (correct classification rate = 37%). However, they did provide information that correctly predicted 71% of the victims in the maintained group and 68% of victims in the absent group. Blanchard et al. (4) had better success at predicting 6-month recovery in their sample of MVA victims. They were able to accurately classify 84% of the victims using information pertaining to injury severity, recovery of injuries, severity of initial posttraumatic stress, and trauma to a close family member during the interim. Injury severity was not associated with recovery in the present sample, suggesting possible differences between the two samples.

Background characteristics, especially gender, had their biggest influence on PTSD symptomatology early (at 1 and 6 months). This finding is consistent with a report from Ehlers et al. (5) in which women were more likely than men to present with PTSD at 3 months after a MVA but not at 1 year. These data suggest that the mediators of chronic symptoms of PTSD may be different from those of acute symptoms. One hypothesis is that initial responses to trauma may be influenced by background variables, such as gender and education, and may in part reflect social and cultural learned responses. Over time, these cultural influences may recede and individual differences in responding may arise from the presence or lack of psychosocial resources necessary to deal with the trauma.

Two of the exposure variables, perceived threat and presence of other passengers, were both important predictors of posttraumatic stress. Presence of other passengers was correlated with perceptions of threat such that more perceived threat was reported if others were present than if the victim was alone. However, the two variables shared only 6% of their variances, suggesting that perceived threat was not the reason why presence of other passengers was associated with the presentation and recovery of PTSD symptoms. In the present study, perceived threat was measured with one item asking the participants to indicate on a five-point scale how threatening their accidents were. Additional questions regarding perceived threat to the self, other passengers, and victims in other vehicles may provide more information that would help to explain these relationships. Additionally, perceived threat was measured at 3 months after the MVA and could be affected by recall biases and intervening experiences. As a result, it may be a measure of lingering threat rather than a good marker of perceived threat at the time of the MVA. Ideally measures of perceived threat should be collected as soon as possible after the trauma. Presence of other passengers may have also reflected feelings of guilt or responsibility, which were not examined in the present study. Future research should further examine the interplay among perceived threat, presence of others, and posttraumatic stress to aid in the identification and treatment of posttraumatic stress.

Wishful thinking coping was the only psychosocial variable that was a significant and consistent predictor of chronic PTSD symptoms. It was not surprising that a passive, emotion-focused attempt at dealing with a trauma was related to the presentation and maintenance of PTSD symptomatology (34, 35). People who use wishful thinking report wishing that a miracle would happen, that they were a stronger person, that they could change what had happened, that they could change the way they felt, and that the situation would go away. They also report having fantasies about how things might have turned out. Wunschel et al. (36) have suggested that people can become so focused on this type of coping that they fail to use more effective strategies that may eliminate or resolve the situation and instead will engage in more maladaptive behaviors. It is difficult to determine why wishful thinking coping was so strongly related to PTSD symptomatology in this study. One explanation may be that individuals who used wishful thinking may have also been experiencing more frequent intrusive recollections of the event, and dealing with a trauma in this manner may have fostered the avoidant symptoms seen in PTSD. However, this does not explain why wishful thinking coping was associated with PTSD symptoms and recovery but avoidance coping was not.

Perceived availability of social support was associated with the PTSD symptom groups at 6 and 12 months, and seeking social support as a coping mechanism was associated with the symptom groups at 6 months. However, neither had unique effects in the logistic regression equations. As expected, availability of social support and seeking social support coping were related but shared only 16% to 19% of their variances, ruling out concerns about construct overlap and multicollinearity. It is possible that their relationships with PTSD symptomatology may be due to their relationships with wishful thinking coping. The literature suggests that social support has beneficial effects in reducing posttraumatic stress because it acts as a coping resource, allowing coping strategies to be more targeted and effective (37, 38). Less social support may contribute to less use of support networks and greater use of wishful thinking, leading to more PTSD symptoms.

The present study focused on the prospective prediction of posttraumatic stress and recovery using psychosocial variables measured before diagnosis. An alternative hypothesis is that the symptoms of posttraumatic stress were actually predicting changes in the psychosocial variables examined in the present study. Previous research has shown deterioration in personal and social resources after trauma, which may make a victim more susceptible to chronic stress (39, 40). The direction of causality is hard to determine in trauma populations because of the lack of data before the trauma. However, none of the psychosocial variables changed over the course of the present study, suggesting that victims did not experience a decline in psychosocial resources related to the MVA.

An additional finding was that 93% to 100% of the victims who were part of the PSTD symptomatic group but who had subthreshold diagnoses met all of the PTSD criteria except criterion C (emotional numbing and avoidance). This finding is consistent with the low prevalence rates of criterion C reported for victims of other types of trauma (11, 41). The two coping strategies related to PTSD symptoms in this sample were wishful thinking coping (a form of avoidance) and seeking social support coping (the opposite of withdrawing from others), both of which should be most closely associated with criterion C. Marsella et al. (42) have argued that this criterion is less directly related to biological states than criteria B and D (reexperiencing and hyperarousal, respectively). Consequently, criterion C could be more susceptible to culture-specific influences and may serve as a starting point for further examination of susceptibility to trauma.

Recruitment and retention are common problems in studies examining the long-term stress responding of trauma victims. In the present study, only 44% of the victims approached agreed to participate. This participation rate is low enough to introduce concerns about sampling bias. However, in this sample, noncompleters and completers were comparable in demographic and accident characteristics. Unfortunately, there was no way of determining their similarity in expression of PTSD. Furthermore, the recruitment rate in the present study (44%) was comparable to the 43% initial return rate reported by Bryant and Harvey (43) in their sample of MVA victims. An additional limitation is that one-third of the present sample failed to complete all of the assessments over the entire 12 months (ie, four questionnaire assessments and three diagnostic interviews). This completion rate is similar to the 67% completion rate reported by Ehlers et al. (5) for three assessments over 1 year. Attrition introduces questions about who is dropping out of the study and whether the most or least symptomatic participants are not completing all measures. This is a potentially serious sampling problem. However, in the present study, completers and noncompleters were comparable in expression of PTSD symptoms and on most other variables. These findings suggest that our sample was comparable to other samples of MVA victims. However, trauma researchers, especially those interested in the long-term effects of MVAs, should increase efforts to improve recruitment and retention rates.

Caution must be used in interpreting the present findings because the cluster of symptoms examined is different from that of PTSD. Although the PTSD symptom groups were very useful in examining the influence of psychosocial variables on the manifestation of PTSD symptoms, the dichotomy of a clinical diagnostic classification was lost. There may be specific mechanisms of PTSD development related to the full cluster of symptoms and their occurrence together that were not captured here. However, these data suggest that psychosocial variables such as wishful thinking coping should be investigated further as mechanisms for individual differences in the manifestation of chronic posttraumatic stress.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
This research was supported by Grant MH 40106 from the National Institute of Mental Health. The authors thank David A. F. Haaga for his helpful reading of an earlier draft of this manuscript.

Received for publication September 8, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 

1. DSM-IV. Diagnostic and statistical manual of mental disorders. 4th ed. Washington DC: American Psychiatric Association; 1994.
2. Green BL. Psychosocial research in traumatic stress: an update. J Trauma Stress 1994; 7: 341–62.[Medline]
3. Norris FH. Epidemiology of trauma: frequency and impact of different potentially traumatic events on different demographic groups. J Consult Clin Psychol 1992; 60: 409–18.[Medline]
4. Blanchard EB, Hickling EJ, Forneris CA, Taylor AE, Buckley TC, Loos WR, Jaccard J. Prediction of remission of acute posttraumatic stress disorder in motor vehicle accident victims. J Trauma Stress 1997; 10: 215–34.[Medline]
5. Ehlers A, Mayou RA, Bryant B. Psychological predictors of chronic posttraumatic stress disorder after motor vehicle accidents. J Abnorm Psychol 1998; 107: 508–19.[Medline]
6. Goenjian AK, Najarian LM, Pynoos RS, Steinberg AM, Manoukian G, Tavosian A, Fairbanks LA. Posttraumatic stress disorder in elderly and younger adults after the 1988 earthquake in Armenia. Am J Psychiatry 1994; 151: 895–901.[Abstract/Free Full Text]
7. March JS. The nosology of post-traumatic stress disorder. J Anxiety Disord 1990; 4: 61–82.
8. McFarlane AC. The aetiology of post-traumatic morbidity: predisposing, precipitating and perpetuating factors. Br J Psychiatry 1989; 154: 221–8.[Abstract/Free Full Text]
9. Breslau N, Davis GC, Andreski P, Peterson E. Traumatic events and post-traumatic stress disorder in an urban population of young adults. Arch Gen Psychiatry 1991; 48: 216–22.[Abstract/Free Full Text]
10. Breslau N, Davis GC, Andreski P, Peterson EL, Schultz LR. Sex differences in posttraumatic stress disorder. Arch Gen Psychiatry 1997; 54: 1044–8.[Abstract/Free Full Text]
11. North CS, Smith EM, Spitznagel EL. One-year follow-up of survivors of a mass shooting. Am J Psychiatry 1997; 154: 1696–702.[Abstract/Free Full Text]
12. Engdahl BE, Harkness AR, Eberly RE, Page WF, Bielinski J. Structural models of captivity trauma, resilience, and trauma response among former prisoners of war 20 to 40 years after release. Soc Psychiatry Psychiatr Epidemiol 1993; 28: 109–15.[Medline]
13. Engdahl BE, Page WF, Miller TW. Age, education, maltreatment, and social support as predictors of chronic depression in former prisoners of war. Soc Psychiatry Psychiatr Epidemiol 1991; 26: 63–7.[Medline]
14. Russo NF, Zierk KL. Abortion, childbearing, and women’s well-being. Prof Psychol Res Pract Professional Psychology – Research and Pract 1992; 23: 269–80.
15. Blanchard EB, Hickling EJ, Taylor AE, Loos WR, Forneris CA, Jaccard J. Who develops PTSD from motor vehicle accidents? Behav Res Ther 1996; 34: 1–10.[Medline]
16. Carr VJ, Lewin TJ, Kenardy JA, Webster RA, Hazell PL, Carter GL, Williamson M. Psychosocial sequelae of the 1989 Newcastle earthquake. III. Role of vulnerability factors in post-disaster morbidity. Psychol Med 1997; 27: 179–90.[Medline]
17. King DW, King LA, Foy DW, Keane TM, Fairbank JA. Posttraumatic stress disorder in a national sample of female and male Vietnam veterans: risk factors, war-zone stressors, and resilience-recovery variables. J Abnorm Psychol 1999; 108: 164–70.[Medline]
18. Regehr C, Cadell S, Jansen K. Perceptions of control and long-term recovery from rape. Am J Orthopsychiatry 1999; 69: 110–5.[Medline]
19. Solomon SD. Role of perceived control in coping with disaster. J Soc Clin Psychol 1989; 8: 376–92.
20. Amir M, Kaplan Z, Efroni R, Kotler M. Suicide risk and coping styles in posttraumatic stress disorder patients. Psychother Psychosom 1999; 68: 76–81.[Medline]
21. Beaton R, Murphy S, Johnson C, Pike K, Corneil W. Coping responses and posttraumatic stress symptomatology in urban fire service personnel. J Trauma Stress 1999; 12: 293–308.[Medline]
22. Solomon Z, Mikulincer M, Flum H. Negative life events, coping responses, and combat-related psychopathology: a prospective study. J Abnorm Psychol 1988; 97: 302–7.[Medline]
23. Ursano RJ, Fullerton CS, Epstein RS, Crowley B, Kao T, Vance K, Craig KJ, Dougall AL, Baum A. Acute and chronic posttraumatic stress disorder following motor vehicle accidents. Am J Psychiatry 1999; 156: 589–95.[Abstract/Free Full Text]
24. Spitzer RL, Williams JBW, Gibbon M, First MB. User’s guide for the Structured Clinical Interview for DSM-III-R. Washington DC: American Psychiatric Association; 1990.
25. Kulka A, Schlenger WE, Fairbank JA, Jordan BK, Hough RL, Marmar CR, Weiss DS. Assessment of posttraumatic stress disorder in the community: prospects and pitfalls from recent studies of Vietnam veterans. Psychological assessment. J Consult Clin Psychol 1991; 3: 547–60.
26. DSM-III-R. Diagnostic and statistical manual of mental disorders. 3rd ed revised. Washington DC: American Psychiatric Association; 1987.
27. Green MM, McFarlane AC, Hunter CE, Griggs WM. Undiagnosed post-traumatic stress disorder following motor vehicle accidents. Med J Aust 1993; 159: 529–34.[Medline]
28. Weiss DS, Marmar CR, Schlenger WE, Fairbank JA, Jordan BK, Hough RL, Kulka RA. Prevalence of lifetime and partial post-traumatic stress disorder in Vietnam theater veterans. J Trauma Stress 1992; 5: 365–76.
29. Association for the Advancement of Automotive Medicine. The abbreviated injury scale (1990 revision). Des Plaines (IL): Association for the Advancement of Automotive Medicine; 1990.
30. Folkman S, Lazarus RS. Manual for the Ways of Coping Questionnaire. Palo Alto (CA): Consulting Psychologist Press; 1988.
31. Fleming R, Baum A, Gisriel MM, Gatchel RJ. Mediating influences of social support on stress at Three Mile Island. J Hum Stress 1982; 8: 14–22.
32. Vitaliano PP, Russo J, Carr JE, Maiuro RD, Becker J. The Ways of Coping Checklist: revision and psychometric properties. Multivariate Behav Res 1985; 20: 3–26.
33. Vitaliano PP, Maiuro RD, Russo J, Becker J. Raw versus relative scores in the assessment of coping strategies. J Behav Med 1987; 10: 1–18.[Medline]
34. Felton BJ, Revenson TA. Coping with chronic illness: a study of illness controllability and the influence of coping strategies on psychological adjustment. J Consult Clin Psychol 1984; 52: 343–53.[Medline]
35. Long BC, Sangster JI. Dispositional optimism/pessimism and coping strategies: predictors of psychosocial adjustment of rheumatoid and osteoarthritis patients. J Appl Soc Psychol 1993; 23: 1069–91.
36. Wunschel SM, Rohsenow DJ, Norcross JC, Monti PM. Coping strategies and the maintenance of change after inpatient alcoholism treatment. Soc Work Res Abstr 1993; 29: 18–22.
37. Cohen S. Psychosocial models of the role of social support in the etiology of physical disease. Health Psychol 1988; 7: 269–97.[Medline]
38. Lepore SJ, Silver RC, Wortman CB, Wayment HA. Social constraints, intrusive thoughts, and depressive symptoms among bereaved mothers. J Pers Soc Psychol 1996; 70: 271–82.[Medline]
39. Kaniasty K, Norris FH. A test of the social support deterioration model in the context of natural disaster. J Pers Soc Psychol 1993; 64: 395–408.[Medline]
40. Lane C, Hobfoll SE. How loss affects anger and alienates potential supporters. J Consult Clin Psychol 1992; 60: 935–42.[Medline]
41. Carlier IVE, Gersons BPR. Stress reactions in disaster victims following the Bijlmermeer plane crash. J Trauma Stress 1997; 10: 329–35.[Medline]
42. Marsella AJ, Friedman MJ, Gerrity ET, Scurfield RM, editors. Ethnocultural aspects of posttraumatic stress disorder: issues, research, and clinical applications. Washington DC: American Psychological Association; 1996.
43. Bryant RA, Harvey AG. Avoidant coping style and post-traumatic stress following motor vehicle accidents. Behav Res Ther 1995; 33: 631–5.[Medline]

This article has been cited by other articles:

				N Wilson and G Thomson Deaths from international terrorism compared with road crash deaths in OECD countries Inj. Prev., December 1, 2005; 11(6): 332 - 333. [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]

				N. Kawana, S.-i. Ishimatsu, Y. Matsui, S. Tamaki, and K. Kanda Chronic Posttraumatic Stress Symptoms in Victims of Tokyo Subway Sarin Gas Attack Traumatology, June 1, 2005; 11(2): 87 - 102. [Abstract] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dougall, A. L. Articles by Baum, A. Search for Related Content PubMed PubMed Citation Articles by Dougall, A. L. Articles by Baum, A. Related Collections PTSD