Anxiety, Acute- and Post-Traumatic Stress Symptoms Following Involvement in Traffic Crashes

Abstract

Anxiety and traumatic stress symptoms are common post-crash. This study documents generalised anxiety responses post-crash, and examines the association between Acute Stress Disorder and Post-Traumatic Stress Disorder (PTSD) with personality and coping styles. Sixty-two patients aged 18–60 admitted to hospital were interviewed prior to discharge, at 2-months and at 6–8 months post-crash. Anxiety symptoms were common, with 55% of participants experiencing moderate-severe levels prior to discharge, with this decreasing to 11% and 6.5% at 2-months and 6–8 months post-discharge. Females reported significantly higher levels of anxiety and acute distress. Neuroticism and generalised coping styles were associated with acute stress responses but not PTSD. These results have important theoretical and practical implications, and indicate that females are at risk of poorer acute anxiety outcomes following injury.

Acute distress, Acute Stress Disorder (ASD), and Post-Traumatic Stress Disorder (PTSD) following road crashes are reported to be frequent and disabling [for reviews see Bryant & Harvey, 2000; Mayou, 2002]. The clinical features of ASD and PTSD are presented in Table 1. The incidence of ASD has been estimated to be 13–14%, while PTSD has been found to range from 8–40% in survivors of traffic crashes 1–6 months post-crash in those requiring some form of medical treatment [for a review see Blanchard & Hickling, 1999; Bryant & Harvey, 2000]. Persistent PTSD was reported at 1-year post-crash in 17% of 1148 consecutive presentations to hospital, with no difference in PTSD rates between roaduser groups being found [Mayou & Bryant, 2003].

Table 1.

Clinical features of ASD and PTSD

Acute Stress Disorder (ASD)
▪ Exposure to an extreme traumatic stressor involving actual or threatened death or serious injury, where intense fear, helplessness or horror is perceived ▪ Either during or after the event, the individual experiences: a sense of numbing; detachment; reduced awareness of surrounds (‘daze’); de-realization; depersonalization; dissociative amnesia (dissociative criterion) ▪ Intrusive thoughts, recollections and dreams of the event; flashbacks (re-experiencing criterion) ▪ Avoidance of thoughts, activities, people, associated with the stressor (avoidance criterion) ▪ Increased arousal: insomnia, nightmares; exaggerated startle, irritable (hyper-arousal criterion) ▪ Time criteria: 2 days – 4 weeks
Post-Traumatic Stress Disorder (PTSD)
▪ As per above, with exception of dissociative symptoms ▪ Require a minimum of 1 re-experiencing symptom, 3 avoidant symptoms, and 2 arousal symptoms ▪ Time criteria: >4 weeks

Source: APA, 1994

The question of which factors are associated with ASD and PTSD is important for theoretical and practical treatment reasons. It has been suggested that acute responses to trauma reflect psychological factors associated with vulnerability, while PTSD is moderated by factors relating to a diverse range of adjustment issues [Bryant & Harvey, 2000]. Predictors of ASD include pre-existing depression severity, history of psychiatric treatment, history of PTSD, perceived severity of the trauma, and being female. Factors such as on-going pain, being female, perceived threat-to-life, pre-existing psychological health, length of stay, and the presence of ASD have been shown to be associated with the onset and persistence of PTSD [Harvey & Bryant, 1999; Jeavons, 2000]. The relationship between objective injury severity, such as Injury Severity Score (ISS) [Baker, O’Neill, Haddon et al., 1974] and Abbreviated Injury Scale (AIS) severity [AAAM, 1998], and PTSD is mixed, however self-rated injury severity has been reported to be an important predictor of PTSD [Blanchard & Hickling, 1999].

THE ROLE OF COPING STYLES AND PERSONALITY IN DETERMINING ADAPTATION POST-CRASH

Recently, attention has turned to understanding vulnerability or susceptibility factors for the emergence, persistence, and even the resolution of stress responses post-crash. The logic of this shift is that these factors may be modifiable. While coping responses and personality have been studied extensively across a range of medical conditions, little research has been undertaken examining those involved in traffic crashes.

There are two dominant schools of thought regarding factors that influence the coping response, and hence adaptation, to stressful events. The situational model of coping states that cognitive and behavioural responses depend on the nature of the stressor itself, and that the coping response is a dynamic process that changes over time as the scenario unfolds [Lazarus & Folkman, 1984]. In contrast, the trait-based view of coping views ‘coping styles as personality dispositions, such as neuroticism and extraversion, that transcend the influence of the situation or time, thus emphasising stability in coping rather than change’ [Porter & Stone, 1996: 133].

There are three broad types of coping responses: Problem (task)-focussed coping includes purposeful cognitive and behavioural efforts directed at managing the problem causing distress; Emotion-focussed coping where the focus is on controlling the emotional reaction to the stressor, and includes self-blame and wishful thinking; and Avoidance coping such as activities and cognitive efforts designed to avoid confronting the stressor either through distraction or social activities. The question of which particular type of coping leads to the ‘best’ adaptation is seen to depend on the stressor itself, and whether the situation is amenable to change. Indeed, evidence exists that in situations of traumatic injury, task-focussed coping may be of little value, or worse, counterproductive, and the other types of coping might be more suited in the short-term [Jeavons, 2000].

With respect to anxiety outcomes, the few available studies of traffic crash samples that have examined coping responses have demonstrated that emotion-focussed coping and avoidant coping are associated with later PTSD and general symptoms of anxiety [Bryant & Harvey, 1995; Jeavons, 2000; Nightingale & Williams, 2000]. In a sample with mild traumatic brain injury resulting from traffic crashes, the personality trait of neuroticism and avoidant coping style was associated with ASD [Harvey & Bryant, 1998]. Neuroticism is characterised by fear and anxiousness, and so such a relationship is not unexpected.

To date, few studies have examined the relationships between post-crash anxiety outcomes such as ASD and PTSD, coping and personality in a single study. This study therefore aims to document these outcomes and relationships in a sample of otherwise healthy people of working age involved in traffic crashes.

METHOD

PARTICIPANTS

Sixty-two patients admitted to a Major Trauma Centre (88.7%) and two metropolitan teaching hospitals (11.3%) following involvement in a road traffic crash were interviewed on three occasions: prior to discharge (T1), 6–8 weeks (T2) and at 6–8 months post-crash (T3).

RECRUITMENT AND RETENTION

During the 13-month recruitment period, 734 consecutive admissions were screened for eligibility to the study, of which 67% (n=492) were excluded (e.g., GCS<13, 25%, ‘serious head injury’, 24%; over 65, 18%). A total of 242 (33%) patients were eligible for the study. Potential participants were approached as their medical health permitted. Contact was made with 110 of the 242 eligible patients (45.5%). Of these 110 patients, 31 (28%) declined participation and 74 consented to the study (67.3%). An additional five (4.5%) eligible patients were approached with a full explanation of the study given verbally, supplemented by a written ‘plain language statement’; these patients were however discharged prior to consent. In total, 132 (54.5%) patients were not approached due to being discharged, co-operative arrangements with competing research studies, and project staffing constraints.

Of the 74 patients consenting to the study, 68 (91.8%) interviews were conducted prior to discharge (T1); the attrition of 6 participants was due to withdrawal of consent prior to the commencement of the interview (n=2), discharge to rehabilitation faster than anticipated (n=3), while one interview was terminated due to patient discomfort. At T2 and T3 attempts were made to contact participants by telephone. Of the 68 participants that completed the T1 interview, 64 completed Interview 2 (T2; 94% retention) and 62 completed Interview 3 (T3; 91% retention). Those lost-to-follow-up were due to a failure to return the questionnaire booklet (n=3, all males); moved residence (n=2, both males), while one female died in a subsequent motor vehicle crash. This analysis focuses on the 62 participants that completed all three interviews. Participants were not paid.

SAMPLE REPRESENTATIVENESS

The proportion of eligible cases consenting to the study was 30.6%, with 28% (n=68) completing the in-patient (T1) interview, 26% (n=64) completing the T2 interview, and 25.6% (n=62) of all eligible cases completing all three interviews. The age, gender, roaduser status, ISS, length of stay and discharge destination (home or rehabiiltation) of the 62 participants completing all three interviews did not differ from the eligible population (p≥0.05). Compared to those completing all three interviews, those refusing to participate were more likely to be male, (83.6%; p ≤ 0.01), however they were otherwise similar on roaduser type, discharge destination, length of stay, ISS and age (p ≥ 0.05).

INCLUSION CRITERIA

Eligible patients were road-users admitted for a period greater than 24 hours, being aged 18–65 years, with a Glasgow Coma Scale (GCS) ≥13 [Teasdale & Jennett, 1974]. Participants provided informed consent, and Ethics Committee approval for the study was obtained.

EXCLUSION CRITERIA

Patients were excluded if any of the following criteria were met: presence of an AIS 3+ head, spinal, or vertebral column injury; crashes involving a fatality; crashes involving vehicle fires or where individuals sustained burn injuries resulting from a vehicle fire; post-traumatic amnesia (PTA) ≥24 hours; pre-existing cognitive impairment; deliberate self-harm; history of psychosis; illicit drug dependence; occupants of a stolen vehicle; non-English speaker; residing outside the state of Victoria, and those considered by medical staff to be medically unfit to provide informed consent.

PARTICIPANT INTERVIEWS

The T1 interview focussed on pre-crash health, details of the crash and a range of instruments to assess the emotional impact of being involved in the crash. T2 and T3 interviews focussed on similar health outcomes as the initial interview with temporally specific open-ended items and survey instruments replacing that focussed on ‘acute’ responses to trauma.

ASSESSMENT INSTRUMENTS

The assessment instruments reported represent a sub-set of measures used in the interview process. They are as follows:

Road-user interview

The interview covered a range of demographic questions, factors related to the crash and pain scales. Pain was reported on a 100-point Visual Analogue Scale (VAS) from the Health Assessment Questionnaire (HAQ) [Fries, Spitz, & Young, 1982; McDowell & Newell, 1996], with respondents indicating their pain prior to the crash (T1), and over the last few hours at T2 and T3. On a 100-point Likert scale, participants rated their perception of control of during and after the crash, feelings of helplessness, responsibility for the crash and that of another party, injury severity, and threat-to-life.

SF-36 [Ware & Kosinski, 2001]

The SF-36 is a 36-item measure of general health status and quality of life. In-patient administration focused on health in the month prior to the crash, and was used to index pre-existing (pre-crash) health. The measure is designed to make reference to the previous four weeks, and has excellent test-retest reliability [Brazier, Harper & Jones, 1992].

Beck Anxiety Inventory – (BAI) [Beck & Steer, 1993]

The BAI is a 21-item self-report survey for measuring the severity of anxiety by assessing the extent to which symptoms ‘bother’ the respondent. Items are rating using a 4-point scale (0–3), ranging from 0 (Not at all bothered) - 3 (Severely, I could barely stand it). The time reference was ‘over the past week, including today’. Anxiety severity categories are: minimal (0–7); mild (8–15); moderate (16–25); severe (26–63). The BAI was administered at each interview. T1 scores were used to describe anxiety responses post-crash and leading up to the interview.

Acute Stress Disorder Scale (ASDS) [Bryant, Moulds, & Guthrie, 2000]

The ASDS is a 19-item self-report scale that measures the presence and severity of Acute Stress Disorder. Respondents rated items using a 5-point Likert scale ranging from 1 (not at all) - 5 (very much) as to how they have felt since the crash. Following Bryant et al. (2000), a score ≥9 for dissociative items plus a combined score of ≥28 on the re-experiencing, avoidance and arousal dimension was used to classify those a experiencing ASD. A total severity score (ASDS) was also calculated. The ASDS was administered at T1.

The PTSD Checklist-Civilian (PCL-C) [Weathers, Litz, Herman et al., 1993]

The PCL-C is a 17-item self-report survey designed to assess PTSD symptoms in a civilian population, and is valid for use as a screening device for the presence of PTSD. The items of the PCL correspond to the Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV) [APA, 1994] symptoms for PTSD. Respondents rated ‘how much you have been bothered by that problem in the past month’ on a 5-point Likert scale ranging from 1 (not at all) - 5 (extremely). Respondents used the crash as the frame of reference for answering the items. The recommendations of Blanchard et al. [1996] were adopted for PTSD classification (i.e., score of ≥44 and meeting Criterion B, C, D). Study participants were classified as ‘sub-syndromal PTSD’ if they met Criterion B (re-experiencing) and C (re-experiencing) or D (arousal) but not both, and not-PTSD otherwise [Hickling & Blanchard, 1992]. The PCL was administered at T2 and T3.

Coping Inventory for Stressful Situations (CISS) [Endler & Parker, 1990]

The CISS is a 48-item self-report measure of three coping styles: task-oriented, emotion-coping, and avoidance-oriented coping, which can be further split into social diversion and distraction coping. Respondents answered items using a 5-point scale, ranging from 1 (not at all) – 5 (very much) to indicate their use of each behaviour in situations in the past they have perceived as ‘stressful, difficult or upsetting’. Standardised scores were calculated for each coping dimension. The CISS was administered at T1.

NEO-FFI [Costa & McCrae, 1992]

The NEO-FFI is a 60-item self-report questionnaire that measures normal personality traits, providing a measure of the five major dimensions, or domains, of personality. The NEO returns domain scores for neuroticism (N), extraversion (E), openness (O), agreeableness (A), and conscientiousness (C), with standardised scores being derived.

The NEO-FFI was administered at T2 as past research has indicated that high levels of acute anxiety temporarily influence personality assessment domain scores, including those indexed by the NEO [Costa, Herbst, McCrae et al., 2000; Costa & McCrae, 1992; Reich, Noyes Jr, Coryell et al., 1986]. By delaying the administration of the NEO-FFI, it was anticipated that the likely confounding influence of anxiety would be minimised, as in-patient anxiety levels were expected to be high. This decision was also premised upon two aspects related to the reliability of the NEO itself. First, the 3-month test-retest reliability as reported by Costa and McCrae (1992b) has been reported to be high, with these correlations being 0.79, 0.79, 0.80, 0.75, and 0.83 for N, E, O, A and C respectively. Second, personality is largely stable from 20–30 years of age, whereupon the personality maturation process reaches a plateau; beyond 30 years of age there is little change in personality traits associated with normal ageing [Costa et al., 2000; McCrae & Costa, 1994].

INJURY DATA

Injury information was obtained using the inpatient medical history. Injuries were coded according to the AIS, 1998 revision [AAAM, 1998]. The ISS was calculated for each participant, and acts as a global index of injury severity [Baker et al., 1974]. The GCS was obtained from paramedic and hospital medical records with the lowest observed value being recorded for this study. The GCS acts as a measure of conscious state and ranges from 3 (non-responsive) – 15 (alert and oriented) [Teasdale & Jennett, 1974].

STATISTICAL ANALYSIS

Patient and injury characteristics were examined, and comparisons between males and females were made using chi-square tests and Repeated Measures ANOVAs where appropriate. Planned comparisons were used to examine gender differences, if any, in ASDS total score and the four criterion sub-scores. The Wilcoxon Signed Rank Test was used to assess changes in BAI categories. The Kruskall-Wallis test was used to assess potential differences in ASD and PTSD symptom severity outcomes between roaduser groups [Siegel & Castellan, 1988]. Z-scores were used to determine differences between Australian SF-36 norms and the sample. Pearson’s r and Spearman’s rho were used to examine relationships between parameters of interest. Power analysis was performed using NCSS-PASS 2000 [Hintze, 2000]. Analyses were conducted using SPSS V.12. Statistical significance was set at p≤0.05.

RESULTS

SAMPLE DEMOGRAPHICS

Of the 62 participants, 38.7% were drivers (37.5% male) and 9.7% were passengers (100% female). Approximately one-fifth (21%) of the sample were motorcyclists (92.3% male), 25.8% were cyclists (81.3% male) while there were 3 pedestrians (4.8%, 33.3% male). Males represented 56.5% of the sample (n=35) and females 43.5% (n=27). The mean age of males (mean: 35.3 years, SD: 12.3) and females (mean: 38.7 years, SD: 12.3) was similar (p=0.3).

Pre-crash mental health status was assessed using the Mental Component Summary (MCS) score of the SF-36. Male (mean: 51.1, SD: 9.5) and female (mean: 50.7, SD: 9.3) pre-crash MCS scores were similar (p=0.9), and indicate an ‘average’ level of mental health functioning, and were within +0.08 to +0.78 standard deviations of Australian population values (z, p≥0.05) [ABS, 1995].

INJURY CHARACTERISTICS

Fifteen (24.2%) participants sustained a compromised state of consciousness with the proportion of males (25.8%) and females (22.2%) being similar (p=0.2); the period of time at GCS 13–14 was not reported. The mean length of stay (LOS) for males (mean: 6.3 days; SD: 5.9, Range: 2–25 days) and females (mean: 5.4 days; SD: 4.3; Range: 2–12 days) did not differ (p=0.4), with a similar proportion of males (25.7%) and females (29.6%) being admitted for greater than 1 week (p=0.7).

The mean injury severity score of males (mean: 9.5, SD=5.9) and females (mean: 11.1, SD: 10.7) was similar (p=0.5), with a similar proportion being classified as major trauma (M: 14.3%; F: 18.5%), indexed by an ISS greater than 15 (p=0.6). The most common non-minor injuries were AIS2+ lower extremity injuries (46.8% of the sample), AIS2 upper extremity injuries (40.3% of the sample), AIS2+ chest injuries (22.6% of the sample), and AIS2+ head injuries (19.4% of the sample). Fishers Exact tests indicated no gender differences in these injury categories (p>0.05). There was no difference in discharge status between males and females (p=0.2), with the majority being discharged home (M: 82.9%; F: 70.4%) and the balance being admitted to rehabilitation or other hospital providers.

Planned comparisons indicated a significant increase in self-reported pain from pre-crash levels (measured at T1) to T2 (p≤0.001), with a significant reduction in pain evident from T2 to T3 (p≤0.001). T3 pain levels remained significantly higher than pre-crash levels (p≤0.001). Pain levels for males (T1: 4.5; T2: 26.4; T3: 15.5) and females (T1: 7.6; T2: 31.1; T3: 19.1) did not differ in any of these three comparisons (p≥0.05).

Analysis indicated no differences between roaduser groups in ASD symptom severity (χ²_kw(4)=4.7, p≥0.05), T2 post traumatic stress symptom severity (PTSS) (χ²_kw (4)=5.8, p≥0.05), or T3 PTSS (χ²_kw(4)=6.9, p≥0.05). Hence, all subsequent analyses were performed by collapsing the roaduser categories.

ANALYSIS OF ANXIETY POST-CRASH

Using the BAI, acute post-crash anxiety symptoms were common, with 55% of participants experiencing moderate-severe levels of anxiety prior to discharge (T1), with this decreasing to 11.3% at T2 and 6.5% at T3. Table 2 shows the percent of males and females by anxiety severity category and the pattern of decreasing anxiety severity is evident, particularly for males. The shift in symptom severity categories from T1 to T2 was statistically significant (z=5.9, p≤0.001), while there was evidence for further reductions in anxiety severity from T2 to T3 (z=1.7, p=0.08). A small proportion of females continued to demonstrate moderate-severe anxiety at T2 and T3, while no males reported moderate-severe anxiety at T3.

Table 2.

Anxiety (BAI) symptom severity

	T1		T2		T3
	Male	Female	Male	Female	Male	Female
	%	%	%	%	%	%
Minimal	20.0	3.7	80	59.3	88.6	63.0
Mild	37.1	25.9	14.3	22.2	11.4	22.2
Moderate	31.4	40.7	5.7	11.1	Nil	14.8
Severe	11.4	29.6	Nil	7.4	Nil	Nil
Mean severity score (SD)
Male	15.7 (10.7)		4.2 (4.7)		2.6 (3.1)
Female	22.7 (11.4)		9.4 (9.6)		6.7 (7.3)

The mean BAI total score for males and females, as seen in Table 2, demonstrates that anxiety severity decreases over time (p≤0.001), with the reduction being equivalent for males and females (p=0.5). Post-hoc comparisons indicate that anxiety severity decreased from T1 to T2, and T2 to T3 (p≤0.05). Anxiety symptom severity for females was higher than that reported by males during the follow-up period (p=0.001).

The most common somatic anxiety (neuro-physiological and autonomic) symptoms at T1 causing ‘moderate-severe’ bother were feeling faint (38.8%), wobbliness in legs (35.5%), feeling dizzy or light-headed (33.8%), feeling hot (32.3%), and sweating but not due to external heat sources (30.7%). The most common subjective or panic type symptoms were being unable to relax (32.3%), being scared (30.7%), fear of the worst happening (24.2%), having difficulty breathing (22.6%), being terrified (21%), being nervous (21%) and a fear of dying (18%). By T2 and T3 the proportion of participants experiencing these feelings decreased considerably, with the most common symptoms being feeling faint (T2: 14.5%, T3: Nil), difficulty breathing (T2: 13%; T3: Nil), numbness or tingling (T2: 12.9%, T3: 4.8%), being unable to relax (T2: 12.9%, T3: 12.9%), indigestion or discomfort in abdomen (T2: 11.3%; T3: 11.3%), and feeling scared (T2: 11.3, T3: 6.5%).

ACUTE STRESS DISORDER

Using the diagnostic criteria specified by Bryant et al. (2000), 11 participants (17.7%; 4 males; 7 female) were classified as an ASD ‘case’. Of interest is the finding that 40 participants (64.5%) met the dissociative criteria (score ≥9) for ASD diagnosis, a result indicating high levels of dissociation. Table 3 presents the ASDS symptom severity for each of the four criteria and the total score. The total ASDS severity score for females (Mean: 35.4, SD: 10.2; Range: 20–59) was marginally higher than for males (Mean: 31.2, SD: 8.7; Range: 19–58) (p=0.08), with planned comparisons indicating the females reported significantly higher levels of dissociative (p=0.03) and avoidance symptoms (p=0.04), while there was no difference in re-experiencing (p=0.2) and arousal (p=0.9) symptom severity.

Table 3.

ASDS symptom and total score for males and females

	Dissoc.†	Re-exp↕	Avoid‡	Arousal	Total
	X̄ (SD)	X̄ (SD)	X̄ (SD)	X̄ (SD)	X̄ (SD)
Male	9.7 (4.1)	5.6 (2.3)	5.0 (1.7)	10.8 (3.6)	31.2 (8.7)
Female	12.3 (5.1)	6.4 (2.6)	6.1 (2.3)	10.7 (3.9)	35.4 (10.2)
All	10.9 (4.7)	5.9 (2.5)	5.5 (2.1)	10.7 (3.7)	33.0 (9.5)

^†Dissociation criterion;

^↕Re-experiencing;

^‡Avoidance

POST-TRAUMATIC STRESS DISORDER (PTSD)

Using the symptom cluster method for PTSD diagnosis described earlier, three participants (4.8%; 2 females) at T2, and 2 participants (3.2%, both female) at T3 were classified as ‘probable PTSD’ (see Table 4). In addition, 8.1% of participants at T2 (M: 2.9%, n=1; F: 14.8%, n=4) were classified as ‘sub-syndromal-PTSD’, while at T3 2 females (3.2% sample) participants (F: 7.4%) were classified as ‘sub-syndromal-PTSD’.

Table 4.

Incidence of probable & sub-syndromal PTSD

Probable PTSD	T1	T2	T3
Male	N/A	2.9% (n=1)	Nil
Female	N/A	7.4% (n=2)	7.4% (n=2)
Sub-syndromal PTSD	T1	T2	T3
Male (%)	N/A	2.9% (n=1)	Nil
Female (%)	N/A	15% (n=4)	7.4% (n=2)

Of those classified as ‘probable-PTSD’ at T2, the male resolved completely, one female continued to be classified as ‘probable-PTSD’ at T3, while one female was classified as ‘sub-syndromal-PTSD’ at T3, demonstrating some resolution of symptoms. Of those classified as ‘sub-syndromal’ at T2, one female was classified as ‘probable-PTSD’ at T3; this participant completed the T3 interview 8.9 months post-crash, meeting the ‘delayed-onset’ DSM-IV criteria (> 6-months). One of the females remained ‘sub-syndromal PTSD’ while the other two females resolved from completely. By T3, the male previously classified as ‘sub-syndromal-PTSD’ was classified as non-PTSD.

CORRELATION ANALYSIS

Correlation analysis was used to examine the relationships between ASD symptom severity (ASDS), ASD diagnosis, and PTSD diagnostic category with a range of parameters. Table 5 presents correlation coefficients for demographic, injury and a range of perceptions of the crash measured at T1.

Table 5.

Relationships between acute anxiety, distress and later PTSD with participant characteristics and perceptions

	ASDS	ASD dx.	T2 PTSD category	T3 PTSD category
ASDS total	1.0
ASD dx.	0.64‡	1.0
T2 PTSD dx.	0.24	0.45‡	1.0
T3 PTSD dx.	0.19	0.40‡	0.69‡	1.0
Gender	0.23	0.19	0.24	0.30†
SF-36 MCS	−0.22	−0.22	−0.04	−0.06
LOS – 7 days	0.23	0.09	0.09	0.28†
LOC*	0.32‡	0.03	−0.11	−0.15
T1 Pain	0.26†	0.11	0.25†	0.23
T2 Pain	0.15	0.11	0.25	0.2
T3 Pain	0.13	0.05	0.24	0.24
Control	−0.55‡	−0.41‡	−0.25	−0.30†
Threat-to-life	0.40‡	0.20	0.24	0.41‡

^‡p≤ 0.01 level;

^†p≤0.05 level;

^*LOC: loss of consciousness

ASDS was associated with a loss of consciousness, acute pain, low perceived control, and a higher threat-to-life. ASD diagnosis was associated with T2 and T3 PTSD diagnostic category, as well as low levels of perceived control. As expected, T2 PTSD diagnostic category was associated with T3 PTSD diagnostic category, but also acute pain. Finally, female gender, prolonged length of stay in hospital, low levels of perceived control, and a high threat-to-life were associated with T3 PTSD diagnostic category. Perceived control and perceived threat-to-life were strongly associated with ASD and PTSD variables, indicating a key role for subjective judgement in initial and ongoing distress. As a point of interest, use of the ASD diagnostic classification rule as described resulted in a sensitivity value for T3 PTSD diagnosis (sub-syndromal + full) being 60%, with the specificity value being 86%.

A number of correlation coefficients approached statistical significance, however these associations suffered from a lack of statistical power. For example, pain at T2 and T3 held a marginal association with T2 (r=0.25, p=0.053, Power=51%), and T3 PTSD (r=0.20, p=0.06, Power=34%), respectively. Age, perceived injury severity, and perceived responsibility were not associated with any of the outcome parameters of interest, and are not presented for reasons of space.

Table 6 examines the association between coping styles, personality factors and anxiety outcomes. Of the personality characteristics, higher levels of neuroticism were associated with higher ASDS severity, and ASD case diagnosis. Extraversion, agreeableness, openness and conscientiousness were not associated with any of the anxiety outcomes of interest.

Table 6.

Relationships between acute anxiety, distress and later PTSD with coping styles and personality characteristics

	ASDS total	ASD dx.	T2 PTSD category	T3 PTSD category
The NEO-FFI Personality dimensions
Neuroticism	0.36‡	0.32‡	0.12	0.03
Extraversion	−0.04	−0.08	−0.06	−0.08
Openness	−0.08	0.01	0.14	0.18
Agreeableness	0.06	0.01	0.06	0.15
Conscientiousness	−0.02	−0.08	0.00	−0.14
Coping styles (generalized responses)
Task-focused	−0.18	−0.22	−0.02	−0.05
Emotion-focused	0.29†	0.23	0.03	−0.08
Avoidance	−0.12	−0.11	−0.05	−0.06
Distraction	0.02	−0.03	−0.01	0.06
Social diversion	−0.24	−0.26†	−0.03	−0.04

^‡p≤ 0.01 level;

^†p≤0.05 level

Emotion-focussed coping was associated with ASDS severity score, and a trend for an association with classification as an ASD ‘case’ was evident (p=0.07, Power = 44%). Higher social diversion coping scores were associated with lower acute distress and a lower likelihood of being classified as an ASD ‘case’. The coping parameters indicate a trend toward lower scores on task-focussed coping being associated with ASD case status, however this correlation suffers low statistical power due to the relatively small sample (p = 0.08, Power = 41%). Personality characteristics and coping styles were not associated with PTSD diagnostic category at T2 or T3.

DISCUSSION

The findings of this study demonstrate that acute post-crash anxiety symptoms were common post-crash, with 55% of participants experiencing moderate-severe levels of anxiety prior to discharge, whilst at 6–8 months post-crash this percentage was considerably lower. Females reported significantly higher anxiety symptom severity than males at all time-points.

Common anxiety symptoms rated as ‘very unpleasant but I could stand it’ (moderately) or ‘I could barely stand it’ (severely) were somatic symptoms, such as feeling faint, dizzy and hot. Common subjective-panic symptoms were being scared, being unable to relax, feeling terrified and fear-of-dying. The somatic symptoms, in particular, may be partially explained by anaesthetic administration and post-operative analgesia administered in the week prior to the T1 interview, precisely the frame of reference for the BAI. Pre-existing anxiety on these items could not be documented. Nonetheless, awareness of these subjective feelings are important for appropriate in-patient care to be provided.

ASD was included in the Diagnostic and Statistical Manual of Mental Disorders-IV in 1994 to capture those at risk of experiencing significant distress post-trauma, and to permit identification of those at risk of PTSD [Bryant & Harvey, 2000]. Using the ASDS, the proportion of participants classified as an ASD ‘case’ was 17% (64% female), a rate consistent with previous research. Analysis of ASDS sub-scores indicated that females experienced higher levels of dissociative (i.e., feeling numb; dazed; dreamlike during and after) and avoidant symptoms (i.e., things, people, thoughts) compared to males. A large proportion of the sample (64.5%; n=40) reached the diagnostic threshold for the dissociative symptoms, however 29 (72%) of these did not reach threshold on the avoidance, re-experiencing and arousal criterion, therefore did not meet the necessary criteria for ASD diagnosis. ASD diagnosis was however correlated with T2 and T3 PTSD diagnostic group, indicating robust predictive utility of the ASDS; the sensitivity of 60% was however moderate although specificity was high (87%). As the ASDS is relatively new, the diagnostic rules may require fine-tuning across different populations to improve test sensitivity. Alternatively, other acute responses may be used in conjunction with the ASDS to predict PTSD. Indeed, recent research has demonstrated the predictive potential of autonomic responses for later PTSD, suggesting these responses reflect the process of ‘fear conditioning’ to the trauma [Bryant & Harvey, 2000; Bryant, Harvey, Guthrie et al., 2003].

Of note was the finding that loss of consciousness (LOC) was correlated with the ASDS total score. This result possibly indicates a potential crossover in symptoms those associated with concussion [Emanuelson, Andersson Holmkvist, Bjorklund et al., 2003]. This symptom crossover is particularly relevant for dissociative symptoms, such as ‘feeling in a daze’. Indeed, of the ASD criteria sub-scores, only the correlation between LOC and dissociation was statistically significant (r=0.45,p<0.05). Examination of the data indicate that only three of the eleven participants classified as an ASD ‘case’ sustained a loss of consciousness. These findings suggest that ASDS diagnostic rules are sufficiently robust so as not to be biased by the crossover of symptoms associated with both dissociation and LOC.

The proportion of participants classified with probable-PTSD was low (<5%) compared to previous studies, however ‘sub-syndromal PTSD’ was somewhat more common. The PCL was used in place of a full diagnostic interview due to a considerable number of other measures administered in the study. Whether the proportion of participants diagnosed with PTSD by using a full diagnostic interview would increase is unknown. The small proportion of those with PTSD may also be a product of the sample being otherwise health people of working age, with all but one being employed and/or studying at the time of the crash.

Correlation analysis was undertaken to assess relationships between a wide range of parameters with ASD. The results indicate that ASDS was associated with a loss of consciousness as noted above, acute pain, low perceived control, and a higher threat-to-life. In addition, the personality disposition of neuroticism, characterised by anxious worry, was associated with acute stress symptoms. A marginal negative association between task-focussed coping and ASD diagnosis was evident. This result is suggestive of benefits associated with proactive behaviour in confronting the challenge of injury, however the statistical power of this association was low. Theoretically, task-focussed behaviours would be manifested in greater compliance with rehabilitation, and hence improved physical outcomes in the longer term. More short-term task-oriented coping responses might involve dealing with insurance matters, arranging hospital visits and the like.

With respect to T3 PTSD diagnostic category, gender, a prolonged length of stay in hospital, low levels of perceived control, and a high threat-to-life was associated with T3 PTSD diagnostic category. In addition, having a high predisposition for emotion-focussed coping and low social diversion coping (not seeking out others in person or by phone) in difficult situations was associated with T3 PTSD. Perceived control and perceived threat-to-life were also strongly associated with ASD and PTSD variables, indicating a key role for subjective judgement in initial and on-going distress. Critically though, neither coping styles nor personality were related to later PTSD.

A number of correlation coefficients approached statistical significance, however these associations suffered from a lack of statistical power. It remains important that larger sample sizes are used to fully explore these relationships.

Previous authors have highlighted two key points in relation to factors associated with the onset and persistence of traumatic stress responses: 1) that short-term distress is likely associated with vulnerability type factors; in contrast longer-term distress is mediated by a diverse range of factors, including social functioning [Lazarus & Folkman, 1984; Porter & Stone, 1996]; and 2) that coping styles and personality play a role in adaptation [Dougall, Ursano, Posluszny et al., 2001], although only a limited amount of research of this type has been conducted in traffic crash samples. The results of the analysis support the notion of vulnerability factors being associated with short-term distress. It is likely that other factors, including resolution of injuries and a range of psychosocial outcomes are related to on-going distress.

The findings that emotion-focussed coping and the personality trait of neuroticism, characterised by anxious worry, were related to distress supports previous research [Harvey & Bryant, 1998; Malt, 1992]. Similar findings have been reported in the burns outcome literature with respect to neuroticism and PTSD [Fauerbach, Lawrence, Schmidt et al., 2000]. These findings indicate that these factors act as points of vulnerability for poor outcomes Other studies have shown emotion-focussed coping to be associated with PTSD [Jeavons, 2000], however this was not the case in this study. Sample differences and PTSD assessment may explain these differences.

The findings for social diversion are important to highlight. Social diversion as a coping construct is classified as an avoidance behaviour, where escape from the difficult, stressful or upsetting situation through interaction with others is the goal. The results reported here suggests that those with a lower predisposition to engage in this type of avoidance coping behaviour demonstrate a higher likelihood of being classified as an ASD ‘case’, and vice versa. Endler and Parker (1999) specifically note the importance of social support, and regard it as a resource and moderator of coping activities as distinct from a coping response or dimension per se. Social support cuts across the three coping dimensions by being a source of information (task-focussed), an outlet for emotion and for support (emotion-focussed), and by offering opportunities for social diversion (avoidance) (Ender & Parker, 1999). It has been reported that social support has a protective effect against PTSD [Dougall et al., 2001; Mayou, 2002], and dissatisfaction with social support is a risk factor for PTSD [Fuglsang, Moergeli, & Schnyder, 2004]. While this study does not report a direct link between social support and PTSD, social diversion coping behaviour, and the social support it most likely arouses, appears to moderate ASD. It might be speculated that those engaging in lower social diversion spend greater time ruminating and cognitively reworking the trauma, both seen to be risk factors for distress [Mayou, 2002].

The marginal negative association between task-focussed coping and anxiety has been reported by only one previous study of traffic crash survivors [Curran, Ponsford, & Crowe, 2000]. While this study examined the role of generalised coping styles, that is, participants were asked what they normally do in stressful, difficult or upsetting situations, coping measured directly as a consequence of the crash and associated injuries might be more useful in predicting later PTSD. As noted above, the situational view of coping would state that coping styles change in response to the demands of the stressor. It might prove that personality and coping styles interact so as to be either protective of (in the case of task-focussed coping), or represent a vulnerability to later PTSD.

This study demonstrated that acute anxiety and distress are commonly experienced post-crash and subside relatively quickly. That pain is associated with acute distress reinforces the need for improved pain management post-crash. The severity of generalized anxiety and acute stress symptoms prior to discharge and in the longer term may be alleviated by teaching patients to recognise negative thoughts and to substitute appropriate adaptive coping strategies, such as social diversion coping strategies, at least in the short-term, and task-focused coping when able. Clinicians should be aware that females were more likely to be at-risk of suffering anxiety symptoms post-crash so that appropriate steps can be taken to mitigate these outcomes.

STUDY LIMITATIONS

The sample represents a relatively small number of otherwise healthy crash-involved individuals hospitalised as a consequence of a road traffic crash where there was no fatality. Fatalities were excluded as they represent a small proportion of total road accident victims and are an extreme form of trauma. These results do not necessarily apply to those outside of the inclusion criteria. Analysis did however indicate that the sample was representative from the eligible population from which it was drawn. The focus of this paper was examining relationships between ASD and PTSD with a range of factors. Future studies would be well served to use multiple regression and structural equation modelling to examine the relationship between perceived social support and situational-specific coping responses with post-traumatic stress symptoms. Further understanding of these relationships may assist in the accurate identification of those at-risk of poor post-crash health.