Abstract
The aim of the study was to estimate the incidence of trauma-related distress and mood disorders in the early stages after acute traumatic hand injuries and identify characteristics associated with these reactions. Data were obtained from 112 patients by means of mailed questionnaires and medical records. Nearly half of the patients had increased levels of intrusive and avoidance symptoms, indicating trauma-related distress. One-third showed signs of a mood disorder. Mood disorders were associated with the need for help with activities of daily living, pain and avoidance symptoms. The study showed that emotional problems in the early stages after injury are related to the consequences of both the injury and the traumatic experience. Negative reactions to the sight of the hand were associated with both trauma-related distress and mood disorders, suggesting that observation of the reactions to the sight of the hand could help to identify patients in need of psychological support.
INTRODUCTION
Patients with complex acute traumatic hand injuries have to face several negative consequences of the injury and the accident, particularly during the first weeks following the accident. A major problem is impairment of hand function, but there are also problems with pain, cosmesis and trauma-related distress (Gilbert, 1996; Grunert and Maksud-Sagrillo, 1998; Grunert et al., 1988, 1992a, b; Smith et al., 1985).
Intrusive symptoms, the re-experience of the trauma in distressing dreams or flashbacks, and avoidance symptoms, the avoidance of thoughts, feelings or places associated with the trauma, are typical of trauma-related distress (Horowitz, 1997). Previous studies have found intrusive symptoms in 70–90% and avoidance symptoms in 20–50% of the hand-injured patients during the weeks after the accident (Grunert et al., 1988, 1992a, b). One of the studies also reported symptoms of anxiety in 48% and symptoms of depression in 62% of the patients (Grunert et al., 1992a). However, most studies only recruited severely hand-injured patients, and it is thus difficult to draw inferences about the magnitude of psychological problems in patients with less severe injuries. Unfortunately, the short period of inpatient treatment after most hand injuries makes it difficult to identify patients in need of emotional support. The aim of the present study was to estimate the incidence of trauma-related distress and mood disorders in the early stages after acute traumatic hand injuries and to identify characteristics associated with these psychological response-reactions.
PATIENTS AND METHOD
Sample selection and procedure
The study included patients with acute traumatic hand injuries requiring inpatient surgical treatment, who were aged more than 18 years and capable of reading and understanding Swedish. Drug abusers and patients with injuries resulting from a suicide attempt were excluded because of psychosocial problems prior to the trauma. Patients were recruited at the Department of Hand Surgery in a Swedish university hospital between August 1997 and May 1999 (22 months). Out of 173 consecutive inpatients with acute traumatic hand injuries, 142 met the inclusion criteria. However, 30 dropped out for various reasons (Table 1), and the final study group consisted of 112 patients (79%).
All patients received written information about the study before discharge from the hospital. Informed consent was then obtained by means of a telephone call a few days after discharge. Patients who consented to participate answered questionnaires by mail on average 12 days (SD, 5) after the day of the accident. The study was approved by the Research Ethics Committee at the university hospital.
Patients
The 112 participating patients were between 20 and 77 years old with a mean age of 45 (SD, 14) years. Ninety-nine patients (88%) were men, and 81 (72%) were married or living with a partner. The right hand was dominant in 108 patients (96%). At the time of the accident, 91 patients (81%) were employed, 11 (10%) were retired and ten (9%) were temporarily unemployed, long-term sick-listed, studying or on military service. Ratings of health and life-situation prior to the accident on verbal ordinal scales showed that 92 patients (82%) had considered their health as “very good” or “excellent” and 79 patients (71%) were “very satisfied” or “extremely satisfied” with their previous life-situation.
Medical records were examined to identify differences between the participants (n = 112) and the patients who dropped out (n = 30). No statistical differences were found between them with regard to age, sex, marital status, length of stay at hospital, type of accident or injury. However, there was a significant difference in their work situation (chi-square (χ2) test; P<001). Most participating patients (91/112), but less than half of the dropouts (13/30), were employed when the accident happened.
Data collection and instruments
Information about the injuries and the duration of hospital stay was obtained from the patients’ medical records.
Study-specific questions
Four questions in a questionnaire collected information about the place of accident, responsibility for the accident, experience at the time of the accident that the injury was severe or crippling, and stress-induced analgesia. Another four questions about experienced problems were constructed on the basis of empirical experiences of hand-injured patients and earlier research (Gilbert, 1996; Grunert and Maksud-Sagrillo, 1998; Grunert et al., 1988, 1992a, b; Smith et al., 1985). These questions inquired about limited ability to perform physical activities (5-point ordinal scale from “not at all” to “very much”), need for help with activities of daily living (eating, washing and dressing, etc.), need for help with other less vital activities (fastening shoelaces, taking money out of a wallet, opening bottles and tins) and reactions to the sight of the hand. The patients were also asked to rate their pain for the last 24 h on a numerical rating scale (NRS) from zero to 10. Zero indicated “no pain at all” and 10 “unendurable pain”. A score of 4 or above was considered unacceptable pain in the hand clinic setting, and was used as a cut-off for the identification of patients with troublesome pain. NRS has proved reliable for measuring pain in trauma patients (Berthier et al., 1998), and good correlations have been shown between NRS and other pain rating scales (Berthier et al., 1998; Downie et al., 1978).
Impact of event scale
Trauma-related distress was measured with a Swedish translation of the Impact of Event Scale (IES) (Horowitz et al., 1979). IES consists of 15 items in two separate subscales. The subscale for intrusive symptoms has seven items (summed scores range from zero to 35) and measures the degree of persistent re-experience of the trauma. The subscale for avoidance symptoms has eight items (summed scores range from zero to 40) and measures the degree of psychological defence against re-experiences. Summed scores below 9 on each subscale indicate low levels, scores of 9–19 medium levels and scores above 19 high levels of traumatic stress symptoms (Horowitz, 1982). The IES is a reliable and valid instrument for identifying patients with trauma-related distress but does not diagnose a post-traumatic stress disorder (Briere and Elliott, 1998; Joseph, 2000). In the present study, Chronbach’s alpha was 0.88 for both the intrusion subscale and the avoidance subscale. This is comparable to alpha values reported in previous studies (Briere and Elliott, 1998; Horowitz et al., 1979; Joseph, 2000) and showed that the IES had satisfactory internal consistency (Streiner and Norman, 1995).
Hospital anxiety and depression scale
Mood disorders were assessed with a previously tested Swedish version (Sullivan et al., 1993) of the Hospital Anxiety and Depression (HAD) Scale (Zigmond and Snaith, 1983). The HAD scale was specifically developed for the detection of emotional problems in patients with somatic illness and does not diagnose a psychiatric disorder. It comprises two subscales, one for anxiety and one for depression. Each subscale comprises seven items. Summed subscale scores range from zero to 21, with summed scores below 8 on each subscale classifying the patients as “non-cases”, scores of 8–10 as “possible cases” and scores of over 10 as “definite cases” (Zigmond and Snaith, 1983). In this study, Chronbach’s alpha was 0.83 for the anxiety subscale and 0.72 for the depression subscale. According to a review study, several other studies have reported higher alpha (0.81) values for the depression subscale (Herrmann, 1997). However, both the anxiety and the depression subscales had satisfactory internal consistency in the present study (Streiner and Norman, 1995).
Statistics
The χ2 test was used to examine if one variable had the same distribution in two independent groups. Spearman’s analysis of rank correlation (rho) was used to measure linear associations between ordered variables, and also to identify significant relationships (P<005) between injured anatomical tissues. Injury groups were formed on the basis of the results of these analyses. An experienced hand surgeon (A.A.) verified the grouping of the patients. Differences between the injury groups were examined with the Kruskal–Wallis test. The smallest injury group (only three patients) was excluded from analyses.
Variables concerning injuries, accidents and experienced problems were tested for relationships with trauma-related distress and mood disorders. The intrusion and the avoidance variables were also tested for relationships with mood disorders. Both the explanatory variables and the outcome variable were dichotomized. Logistic regression analyses with backward stepwise elimination tested which of the variables had the best ability to explain the outcome. Results of the logistic regression analyses are presented as odds ratio (OR) with 95% confidence intervals (CI95%) and significance values (P). Information about the percentage of correctly classified cases is also given.
RESULTS
Injuries
The frequencies of patients with similar injuries are presented in Table 2. Amputation was the most frequent injury. One patient had a complete hand amputation but in most cases one or more fingers was amputated: replantation was performed in four patients with amputated fingers and three were successful. A few patients had contaminated soft-tissue injuries which required surgical cleaning and intravenous antibiotic treatment. Half the patients had injured the index, middle, ring and little fingers (Table 2). With the thumb included, injuries to the digits comprised 65% of all the injuries. The dominant hand was injured in 47 patients (42%). About half the patients (54%) remained in hospital for only 1 day and night (median, 1; range, 1–14). The majority (90%) left hospital within 4 days.
Accidents
Most accidents had happened at home (n = 45; 40%), usually when chopping wood or repairing the house. Accidents at work (n = 43; 38%) occurred mostly in heavy industry, building, transport or farming/forestry. A smaller proportion of accidents happened during sports or hobby activities (n = 13; 12%) such as riding, soccer and shooting. Some accidents happened in unspecified places as a result of external circumstances (n = 11; 10%), for example dog bites or assaults in the street. No one was injured in a road traffic accident. Sixty-eight patients (61%) held themselves responsible for the accident.
All except three patients remembered their experiences at the time of the accident. Thirty-eight patients (35%; n = 109) thought they had suffered a severe or crippling injury. Three of them, all with an arterial injury at the wrist, had thought they were going to die. Experience of threat at the time of the accident differed between the injury groups (Kruskal-Wallis test; P = 002). Patients with amputation more frequently thought they had suffered a severe or crippling injury when the accident happened than patients with other injuries (63% and 23%, respectively).
Sixty-three patients (56%) had experienced stress-induced analgesia at the time of the accident but of various durations. Two-thirds started to feel pain during their passage to hospital and one-third when they arrived at the hospital.
Experienced problems
The patients’ subjective assessment of how limited they felt to perform physical activities revealed that 44 patients (39%) felt “much” limited and 23 patients (21%) “very much” limited. Thus, more than half of the patients experienced substantially limited physical ability. A total of 71 patients (63%) stated that they needed help with everyday activities such as eating, washing and dressing, as a result of the injury. They also needed help with less important activities such as fastening shoelaces, taking money out of a wallet and opening bottles and tins. Another 24 patients (22%) only needed help with the latter. Only 17 patients (15%) managed without help. The patients’ statements about how much help they needed had a weak, but significant, correlation with experience of limited physical ability (rho=0.37; P<001). No significant differences in experience of limited physical ability or need for help were found with regard to type of injury, location of injury or whether the dominant hand was injured.
Negative reactions to the sight of the hand when the dressing was changed were reported by 52 patients (46%). Of these, 36 patients (32%) “looked but felt uneasy” and 16 patients (14%) “avoided looking at the hand if it at all was possible”. There were no significant differences between the different injury groups with regard to negative reactions at the sight of the hand.
The patients’ ratings of pain during the last 24 h had a median value 2 (range, 0–9). Thirty-four patients (31%) had experienced troublesome pain (score≥4). There was a difference in pain ratings between the injury groups (Kruskal-Wallis test; P = 0.013). Compared with the total group, the median value for pain was above average in the group of patients with a combination of an injured major nerve, arterial blood vessel and/or injured flexor tendon (median, 5; range, 0–8; n = 10) and also in patients with amputations (median, 3; range, 0–9; n = 32).
Trauma-related distress
Intrusive symptoms were somewhat more frequent than avoidance symptoms (Table 3). However, the two were strongly correlated (rho=0.74; P<0.01; n = 111) and occurred usually at the same level in the same patient. Taking both subscales into consideration, 61 patients (55%) had low levels, 36 patients (32%) had medium levels and 14 patients (13%) had high levels of either or both symptoms.
As each patient usually had both intrusive and avoiding symptoms, analyses were performed to identify characteristics associated with medium or high levels of either or both symptoms (score ≥9 on at least one of the subscales). A logistic regression model including negative reactions to the sight of the hand, experience at the time of the accident that the injury was severe or crippling, and troublesome pain had the best ability to predict whether the patient belonged to the group with increased levels of traumatic stress symptoms (Table 4). A model including these variables correctly classified 65% of the patients with medium or high level of traumatic stress symptoms and 77% of all patients. Several patients with amputation had high levels of traumatic stress, but amputation had no significant predictive value with regard to trauma-related distress according to the logistic regression analyses.
Mood disorders
The median values for ratings of anxiety and depression were equal, even though somewhat more patients were definite cases of anxiety (Table 5). There was a fairly strong correlation between anxiety and depression (rho=0.62; P<001; n = 111). Taking both subscales into consideration, 24 patients (22%) were possible cases and 11 patients (10%) were definite cases of a mood disorder, whether anxiety or depression or both.
Because there were few patients with only anxiety or only depression, analyses were performed to identify variables with the best ability to classify patients as possible or definite “cases” of a mood disorder (a score of ≥8 on at least one of the HAD subscales) or “non-cases”. Logistic regression analyses showed that most cases with signs of a mood disorder were predicted by negative reactions to the sight of the hand, the need for help with activities of daily living, medium or high level of avoidance symptoms and troublesome pain (Table 6). Troublesome pain was not significant, but increased the sensitivity of the model. A model without troublesome pain identified 58%, while a model including troublesome pain identified 71% of the patients who had signs of a mood disorder. The final model, which included troublesome pain, correctly classified 84% of all cases, both those with and those without symptoms of a mood disorder. Both intrusion and avoidance were associated with mood disorder, but avoidance had better ability to classify patients as “cases” or “non-cases” according to the logistic regression analyses.
DISCUSSION
The study included patients with acute traumatic hand injuries, which were sufficiently severe to require inpatient surgical treatment. As the aim was to assess psychological consequences, patients with known psychosocial problems prior to the accident were excluded.
Ratings on the IES, made on average 12 days after the accident, showed that nearly half of the patients had increased levels of traumatic stress symptoms. Most frequent were intrusive symptoms such as nightmares, flashbacks and involuntary recollections of the accident. Previous studies of hand-injured patients have reported higher frequencies of intrusive symptoms but a similar frequency of avoidance symptoms, but had not excluded patients with psychosocial problems prior to the accident (Grunert et al., 1988, 1992a, b). Interestingly, a study of severely injured trauma patients treated at an intensive care unit who did not have pre-existing psychosocial problems (Schnyder et al., 2001), showed similar IES scores to those recorded in this study.
Approximately one-third of the hand-injured patients had signs of a mood disorder with one out of ten being a definite case of anxiety, depression or both. The patients in this study reported more symptoms of both anxiety and depression than a reference group consisting of healthy men (Sullivan et al., 1993), but fewer symptoms than patients with an acute traumatic hand injury in an earlier study (Grunert et al., 1992a). As mentioned earlier, the latter study concerned only a limited sample of patients.
It is important to identify patients with trauma-related distress in its early stage, because early intervention may prevent progression (Rusch, 1998). Patients with intrusive symptoms are easy to identify in clinical practice by simply asking them if they suffer from nightmares or flashbacks, but patients with avoidance symptoms are more difficult to identify as they usually avoid talking about the accident. Furthermore, the hospital stays of hand patients is short, which also makes it difficult to discern patients in need of psychological support: this is why this study attempted to find variables associated with traumatic stress.
Experience at the time of the accident that the injury was severe or crippling was associated with trauma-related distress, which is to be expected as experience of a stressful event is the main cause of traumatic stress (Horowitz, 1997). Stress-induced analgesia at the time of the accident had no relationship with trauma-related distress, even though both are initiated by stressful situations. A possible explanation is that stress-induced analgesia is a physiological response-reaction which occurs automatically in stressful situations (Stout et al., 1995), while experiences of a stressful event resulting in trauma-related distress depends on a psychological appraisal of the situation (Horowitz, 1997). Another possible explanation is poor recollection of events. The patients in the present study answered questions about their experiences at the time of the accident 1–2 weeks after the event. Even though nearly all patients answered that they remembered what happened, there could be recollection bias in the assessments.
Troublesome pain was associated with increased levels of traumatic stress symptoms. A previous study of patients with burns has also reported a positive correlation between traumatic stress and pain shortly after the injury (Taal and Faber, 1997), and a review study has suggested that there is also a relationship with pain in the later stages after an injury (Sharp and Harvey, 2001). However, the causal relationship between pain and traumatic stress is ambiguous (Sharp and Harvey, 2001; Taal and Faber, 1997). Whether it is pain that reminds the patient of the traumatic event resulting in traumatic stress symptoms, or the involuntary re-experience of the trauma which causes the pain, or both these mechanisms occurring concurrently is not known.
This study verified that traumatic stress symptoms are associated with negative reactions to the sight of the hand (Grunert et al., 1992a). Patients in Grunert et al.’s study (1992a) who found it difficult to look at the hand reported that, whenever they did look at it, they experienced flashbacks and intrusive images of the original injury. In the present study, patients who stated that they felt uneasy at or avoided the sight of the hand, not only had increased levels of traumatic stress symptoms but also signs of a mood disorder. These findings suggest that observation of patients’ reactions to the sight of the hand when the dressing is changed could help to identify patients in need of psychological support.
Symptoms of a mood disorder were associated with both intrusion and avoidance, but avoidance had better predictive value in the logistic regression model. This finding, and perhaps also the relationship between symptoms of a mood disorder and negative reactions to the sight of the hand, could be explained by Horowitz’ model describing stress response syndromes (Horowitz, 1997). According to his model, the memory of the traumatic experience has a tendency to be repeated in intrusive thoughts. If this process of recurrent repetitions creates uncontrollable or overwhelming emotions, the person tries to interrupt it by avoiding everything that recalls the traumatic event (Horowitz, 1997). However, further research is needed to explain the causal relationship between mood disorders, traumatic stress and negative reactions to the sight of the hand.
Symptoms of a mood disorder were related not only to traumatic stress but also to the consequences of the hand impairment, such as the need for help with activities of daily living. There were no differences between the injury groups with regard to need for help or symptoms of mood disorders: thus, no particular injury seemed to identify patients with emotional response-reactions. Mood disorders also were associated with troublesome pain which occurred in a third of the patients, most of whom had undergone an amputation or a major nerve injury. Both groups are known to have long-lasting troublesome pain, which may be difficult to relieve (Rosén and Lundborg, 2001; Wahren, 1991). However, our results emphasize that the acute care of the patients with hand injuries should include great efforts to relieve pain.
The small number of patients in the injury groups in this study made conclusions about different injuries somewhat uncertain. Further research with larger samples is needed to verify our results. Also, it could be questioned if the sex distribution in this study is representative of the actual patient group. However, the low frequency of women with hand injuries corresponds to what is normally found in the clinic where this study took place. Also other studies of hand-injured patients have mainly included men (Angermann and Lohmann, 1993; Grunert et al., 1992a; Rosén and Lundborg, 2001; Smith et al., 1985). Analyses of dropouts found a difference in work situation between non-participating and participating patients, though work situation had no influence on the results in this study.
Footnotes
Acknowledgments
This research was supported by grants from the Committee for Research and Development in Caring Sciences, University of Gothenburg and the Örebro County Council Research Committee. The authors would like to thank the medical staff of the orthopaedic and hand surgery ward for their invaluable help with sample selection and information to the patients, and Professor Lennart Bodin and his colleague Anders Magnusson for statistical guidance and discussions.