|Home | About | Journals | Submit | Contact Us | Français|
Traumatic physical injury can result in many disabling sequelae including physical and mental health problems and impaired social functioning.
To assess the effectiveness of psychosocial interventions in the prevention of physical, mental and social disability following traumatic physical injury.
The search was not restricted by date, language or publication status. We searched the following electronic databases; Cochrane Injuries Group Specialised Register, CENTRAL (The Cochrane Library 2009, Issue 1), MEDLINE (Ovid SP), EMBASE (Ovid SP), PsycINFO (Ovid SP), Controlled Trials metaRegister (www.controlled-trials.com), AMED (Allied & Complementary Medicine), ISI Web of Science: Social Sciences Citation Index (SSCI), PubMed. We also screened the reference lists of all selected papers and contacted authors of relevant studies. The latest search for trials was in February 2008.
Randomised controlled trials that consider one or more defined psychosocial interventions for the prevention of physical disability, mental health problems or reduced social functioning as a result of traumatic physical injury. We excluded studies that included patients with traumatic brain injury (TBI).
Two authors independently screened the titles and abstracts of search results, reviewed the full text of potentially relevant studies, independently assessed the risk of bias and extracted data.
We included five studies, involving 756 participants. Three studies assessed the effect of brief psychological therapies, one assessed the impact of a self-help booklet, and one the effect of collaborative care. The disparate nature of the trials covering different patient populations, interventions and outcomes meant that it was not possible to pool data meaningfully across studies. There was no evidence of a protective effect of brief psychological therapy or educational booklets on preventing disability. There was evidence from one trial of a reduction in both post-traumatic stress disorder (PTSD) and depressive symptoms one month after injury in those who received a collaborative care intervention combined with a brief psycho-educational intervention, however this was not retained at follow up. Overall mental health status was the only disability outcome affected by any intervention. In three trials the psychosocial intervention had a detrimental effect on the mental health status of patients.
This review provides no convincing evidence of the effectiveness of psychosocial interventions for the prevention of disability following traumatic physical injury. Taken together, our findings cannot be considered as supporting the provision of psychosocial interventions to prevent aspects of disability arising from physical injury. However, these conclusions are based on a small number of disparate trials with small to moderate sample sizes and are therefore necessarily cautious. More research, using larger sample sizes, and similar interventions and patient populations to enable pooling of results, is needed before these findings can be confirmed.
Injuries account for 9% of the world’s deaths and 12% of the world’s disease burden (WHO 2002), and may arise from road traffic crashes, poisoning, falls, fires, drowning, interpersonal violence and war, self-inflicted injuries, as well as other sources. The World Health Organization (WHO) states that “while mortality is an important indicator of the magnitude of a health problem, it is important to realise that for each injury death, there are several thousand injury survivors who are left with permanent disabling sequelae” (WHO 2002, p 3). Even in patients with physical problems that do not arise from traumatic injury, the ability to function effectively is strongly influenced by factors such as mood, coping skills and social support. Interventions that influence these factors are therefore likely to contribute to better health and more cost-effective outcomes (Sobel 1995).
Schnyder et al found that 25.5% of severely injured accident victims showed some form of psychiatric morbidity after one year and that this “can be predicted to some degree by mainly psychosocial variables” (Schnyder 2001, p 653), in particular patients’ early cognitive appraisal of their accident. However, while NICE (Gersons 2005) has endorsed the effectiveness of longer-term psychosocial interventions in the aftermath of psychological trauma, the effectiveness of short-term ‘debriefing’ has not been demonstrated (Rose 2002). While psychosocial interventions following physical injury may alleviate psychiatric co-morbidity, we wish to focus on the benefits of preventive interventions rather than therapeutic interventions per se. To date, there has been no attempt to systematically review the effects of psychosocial interventions that seek to alleviate the distress of those who have acquired physical injuries, including subsequent disability. This review will provide this evidence. For the purposes of this review, by disability we mean any diminution in an individual’s psychological, social or physical functioning that has arisen following, and as a consequence of, a traumatic physical injury.
This review will focus on primary preventive rather than therapeutic interventions, as intervening early after the traumatic event to a general population of trauma patients presents a novel approach to the reduction of disability following traumatic injury. Psychosocial interventions are interventions that have their primary mode of action through psychological or social processes. Such interventions include, for instance, direct therapeutic work, health education and social support.
Physical injury may result in impairment of physical functioning. The way in which people respond to such impairment, along with the social and environmental context they live in, determines the degree of disability associated with the injury. A physical injury may disable people in terms of their physical, mental or social functioning. An individual’s own ability to cope with physical impairment, as well as their broader social situation, offers opportunities to reduce the extent to which physical injury results in disability. By providing people with psychological and social resources that assist their coping responses to physical impairment, psychosocial interventions may be able to prevent physical impairment resulting in physical, mental and social disability.
This is the first systematic review to consider the effectiveness of psychosocial interventions designed to prevent disability following an injury that produces physical impairment.
Randomised controlled trials that considered one or more defined psychosocial interventions for the prevention of physical disability, mental health problems or reduced social functioning as a result of traumatic physical injury. We included cross-over trials, cluster-randomised trials and factorial trials.
We excluded non-randomised intervention studies.
Patients who have suffered a traumatic physical injury.
We excluded trials which included people with traumatic brain injury (TBI) unless they could be disaggregated from other physically injured people who received the intervention, or comprised less than 20% of trial participants. We excluded trials which included people without traumatic physical injury (for example, psychological de-briefing following a traumatic incident not resulting in physical injury) if the results could not be disaggregated to include only those who had suffered a traumatic physical injury or those without physical injury comprise less than 20% of trial participants. We retrieved the full text of trials including patients with TBI or without traumatic physical injury and only excluded trials if the above conditions were not met. We excluded musculoskeletal conditions incurred other than through a physical injury. We only included sexual assault if it resulted in a physical injury.
We define ‘psychosocial interventions’ as being any intervention that focuses on psychological and/or social factors rather than biological factors (definition taken from Ruddy 2005). This may include, but is not limited to:
We included psychosocial interventions as long as they were sufficiently described by trialists to facilitate replication. Interventions may be administered by any health professional (for example, psychologist, medical practitioner, nurse or occupational therapist) or lay person and in any form, for example, individual or group therapy, over the telephone, or in the form of written material. Psychosocial interventions may be offered to enhance a person’s coping resources without any suggestion that they are currently suffering through any type of psychopathology or psychological disorder. Such interventions may prevent circumstances arising where their ability to cope is exceeded, resulting in a state of psychological disorder or variously defined psychopathologies. It may well be the case that people who are currently experiencing psychological disorder benefit from interventions designed to be preventive, and in this sense they may be understood as being therapeutic. However, our focus here is on prevention and not treatment.
Psychosocial interventions were compared with:
Trials were excluded if:
The search was not restricted by date, language or publication status.
We searched the following electronic databases;
Full details of the search strategies can be found in Appendix 1.
We scanned the reference lists of all selected papers and contacted authors of relevant studies to seek out additional studies.
The Cochrane Injuries Group’s Trials Search Co-ordinator ran the relevant search strategies across the appropriate databases. Two authors (MDS and MM) separately screened the titles and abstracts of the citations identified by the search to determine which papers met the pre-determined criteria. In case of doubt or disagreement, we obtained the full article for inspection. Full text copies of all potentially relevant studies were obtained and independently assessed by MDS and MM to determine whether they met the inclusion criteria. In the event of a disagreement, the third author (DD) was asked to give his opinion to resolve the issue. We stored all identified study records using electronic bibliographic software (Endnote XI).
Two authors (MDS and MM) extracted data from the trial reports using a purposefully designed data extraction form. In the event of a disagreement, the third author (DD) was asked to give his opinion. We contacted trial authors for missing data where appropriate.
Two authors (MDS and MM) independently assessed the methodological quality of selected trials. In the event of a disagreement, the third author (DD) was consulted.
We used the ‘Risk of bias’ tool to assess the risk that a study over- or under-estimates the true intervention effect. This tool involves a description and a judgement for the following criteria: sequence generation; allocation sequence concealment; blinding of outcome assessment; incomplete outcome data; selective outcome reporting; intention-to-treat analyses and other potential sources of bias. Each criterion was judged ‘Yes’ indicating low risk of bias, ‘No’ indicating high risk of bias, or ‘Unclear’ indicating either lack of information or uncertainty over the potential for bias. Plots of ‘Risk of bias’ assessments were created in Review Manager. We assessed missing data and attrition rates for each of the included studies, and reported the number of participants who were included in the final analysis as a proportion of all participants in the study. Reasons given for missing data are provided in the narrative summary. We ascertained the extent to which the results were altered by missing data. This qualitative quality assessment was not used as a threshold for inclusion of studies, but as a possible explanation for differences between studies when interpreting the results of the review (Schulz 1995).
We contacted authors of all studies included in the review in order to obtain information absent from the published reports.
As the studies were too disparate to allow pooling of results in a meta-analysis, we described the results of the trials using a qualitative summary. We performed no subgroup or sensitivity analyses.
The search strategy generated 1420 citations; 1417 from database searching, two from searching reference lists and one from contacting authors. MDS and MM independently checked the titles and abstracts of these citations and excluded 1350 as clearly irrelevant. We identified 70 citations as potentially relevant and located these for full text screening. We were unable to locate the full text of two studies (see ‘Characteristics of studies awaiting classification’), and one study only had unpublished data. MDS and MM independently screened the full text of 67 studies. Sixty-two were excluded as they did not meet the inclusion criteria (see’Characteristics of excluded studies’). Seventeen were not RCTs, 17 evaluated interventions for the treatment rather than prevention of disability, 11 evaluated a complex intervention for which the psychosocial component could not be disaggregated, seven did not measure an aspect of disability as the outcome of the trial, eight included more than 20% of patients who had not suffered a traumatic physical injury and two did not evaluate a psychosocial intervention. In total five studies met the inclusion criteria and were included in the review.
There were no disagreements which could not be resolved between MM and MDS when screening abstracts and titles, and two disagreements when screening the full text of reports. These were referred to DD for a third opinion and in both cases the studies were excluded. The kappa score for inter-rater reliability on a sample of 413 citations was 0.94.
See: ‘Characteristics of included studies’; ‘Characteristics of excluded studies’.
Five trials, involving 756 participants, are included in this review. The included trials were very different in terms of population studied, intervention assessed and outcomes. As such, pooling the results in a meta-analysis was not possible. Instead, studies are discussed according to intervention type.
Four trials included patients admitted to a trauma centre following a traumatic physical injury (Holmes 2007; Pirente 2007; Turpin 2005; Zatzick 2001) and one included patients who had undergone surgery for hip fracture (Burns 2007).
The eight included studies examined the following group comparisons:
All but one study measured physical disability outcomes (Turpin 2005), and all included at least one mental health status outcome. Only one study assessed social functioning or health care utilisation (Burns 2007).
In total the five trials assessed 12 different outcomes. Four measured physical disability (mobility, pain and changes in pain, physical illness and physical functioning); six measured mental health status (depression and change in depressive symptoms, anxiety, PTSD and change in PTSD symptoms, alcohol abuse, substance abuse and any psychological disorder); one assessed global assessment of functioning (health related quality of life); and one assessed social functioning and health care use (length of stay in hospital). With the exception of depression, where two studies used the same tool, each study used a different tool to measure the outcomes.
Where it was possible to assess this, included trials had generally poor ratings of trial quality and many suffered from biases in addition to those assessed by the Cochrane criteria (see Figure 2 and Figure 3).
All trials had adequate sequence generation, using computer generated randomisation or random number tables. All trials also had adequate allocation concealment, using either an independent telephone randomisation service or sealed opaque envelopes.
Blinding of outcome assessment was only adequately addressed in two trials, with four trial reports containing insufficient detail and one using patient self-assessment (Menzel 2006).
Only three trials performed intention-to-treat analyses. Significant other potential sources of bias included the small sample size (especially after high losses to follow up) and therefore lack of power for some studies.
We did not pool data due to heterogeneity between trials in terms of the intervention, outcomes, time point at which the outcome was assessed and the population studied. Instead, trials’ results are discussed by intervention type (brief psychological therapies, self-help information and collaborative care). In addition, we performed no subgroup or sensitivity analyses due to the heterogeneity of the included studies. See Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8 for full results of the included studies.
This study examined 170 patients over 60 years old who had undergone surgery for hip fracture and who had been classified as not at risk of suffering from depression (score of six or less on the Geriatric Depression Scale). The intervention consisted of up to seven sessions of CBT delivered by an assistant psychologist. The control group received treatment as usual. Outcomes were assessed at six weeks (79.4% follow up), three months (66.5% follow up) and six months (64.7% follow up). There were no significant differences in any of the outcomes (depression, mobility, pain, physical illness, functioning, or length of hospital stay) at any of the follow-up points. It is possible that the selection of participants who screened negative for depression may have reduced the effectiveness of the intervention on the subsequent development of mental health problems. Indeed, 18% of the intervention group (and 11% of the control group) were already taking antidepressants at baseline, suggesting that the screening tool used was not sensitive, creating commonality between groups and possibly further reducing the effect of the intervention.
In a similarly sized trial, the authors randomised 171 severely injured trauma patients to receive up to eight sessions of individual CBT or treatment as usual. Only the 92 patients with complete outcome data at both six and 12 months were analysed (53.8%). There were significant between group differences at baseline with a higher proportion of anxiety (57.8% versus 40.4%, P = 0.006) and depression (73.3% versus 44.7%, P = 0.014) in the intervention compared to the control group. There were no significant differences at any follow-up time between the two groups in any of the outcomes (health-related quality of life (HRQoL), depression or anxiety), possibly because of the initial differences between groups. There was significant within group reduction in depression among the intervention group from surgical ward to discharge (P = < 0.001), from surgical ward to six months follow up (P = 0.004) and from six to 12 months follow up (P = 0.013). The same reduction in anxiety was apparent in the intervention group from surgical ward to discharge (P = 0.001) and from surgical ward to six months follow up (P = 0.002).
One study examined the effect of interpersonal therapy (Holmes 2007). No significant differences between groups were found for any outcome.
Ninety major physical trauma patients admitted to two trauma centres were randomised to an average of 5.9 Interpersonal Counselling (IPC) sessions delivered by a clinical psychologist. Follow up at six months was 63.3%. There was a high rate of drop-out from the intervention group (24/51, 47.1%), though they did not differ in characteristics from completers. No significant between group differences were found for any of the outcomes (depression, anxiety, PTSD, alcohol and substance abuse, any psychiatric disorder and physical functioning - effect estimates and P values not reported). The lack of differences between groups may be partly explained by the high degree of ‘non-specific psychological support’ (mean 22.6 hours) and psychiatric/psychological treatment (mean 0.8 hours) the control group received as well as the small sample size for analysis. Participants with a past history of major depression who received IPC had significantly higher levels of depression at six months (P = 0.018), indicating that the intervention may be harmful to a vulnerable group of individuals.
One study examined the effect of self-help information delivered as a booklet (Turpin 2005). The intervention had no protective effect on any disability outcome.
Two hundred and ninety-one Accident and Emergency patients who had sustained a physical injury were randomised to receive either a self-help information booklet six to eight weeks after hospital attendance, or a letter without the booklet. The booklet described and normalised common physiological, psychological and behavioural reactions to traumatic injury and provided advice on non-avoidance, emotional help and seeking further help. Only 10% (291/2818) of those eligible agreed to participate. There were significant differences between consenters and non-consenters in terms of age, gender and trauma type. Follow up was poor with only 34% of those randomised followed up at six months. This was partly due to an administrative error which resulted in 66 participants who had completed baseline measures being removed from the analysis. There were no significant differences between groups in anxiety or PTSD symptoms. However, there was evidence that the booklet may have a detrimental impact on mental health status. At three months follow up there was a greater reduction in PTSD cases in the control than the intervention group (P = 0.06), and in an intention-to-treat analysis at six months there was a higher proportion of depressed patients in the intervention rather than the control group (18% versus 7%, P = 0.054).
One study assessed the effect of collaborative care. The intervention had no protective effect on any disability outcome.
In a pilot study, 34 injured patients admitted to a trauma centre were randomised to receive either collaborative care comprising a personally assigned trauma support specialist and a brief psychoeducational intervention targeting PTSD, or treatment as usual. The trauma support specialists spent on average 1.5 hours with each patient. Follow up at four months was 74.5%. At one-month follow up in intention-to-treat analyses the intervention group had significantly decreased PTSD (effect estimate 0.99, F[1,33] = 6.8 P < 0.05) and borderline significant decreased depression (effect estimate 0.58, F[1,33] = 3.7 P = 0.07) symptoms when compared to the control group. Due to the complex intervention it is not possible to determine which aspect of the intervention had a positive impact on mental health status. At four months the intervention groups symptoms had significantly increased relative to the control groups for both PTSD (effect estimate 1.75, F[1,27] = 6.1 P < 0.05) and depression (effect estimate 1.15, F[1,33] = 6.8 P < 0.05). There were no significant between group differences at either one or four months for the other outcomes: drinking to the point of intoxication and functional limitations.
The aim of this review was to evaluate the effectiveness of psychosocial interventions for the prevention of disability following traumatic injury. Five trials were identified. The disparate nature of the trials covering different patient populations, interventions and outcomes meant that it was not possible to pool data meaningfully across studies. Three studies assessed the effect of brief psychological therapies, one assessed the impact of a self-help booklet, and one the effect of collaborative care.
Overall there was no evidence of a protective effect of brief psychological therapies or self-help booklets on preventing disability, and evidence from one trial of a reduction in both PTSD and depressive symptoms one month after injury in those who received a collaborative care intervention combined with a brief psychoeducational intervention (Zatzick 2001). Mental health status was the only disability outcome affected by any intervention. In three trials the psychosocial intervention had a detrimental effect on the mental health status of patients.
The findings from this review must be viewed in light of the small sample size and the heterogeneous characteristics of trials published in this area.
The absence of effect in the brief psychological therapy trials is surprising given the strong evidence for the effectiveness of these interventions, in particular CBT, in treating a range of mental health problems (Bisson 2007; Hunot 2007; Soo 2007) and other conditions including sleep problems (Montgomery 2003). This may be the result of low power in all three brief psychological therapy trials (the sample size for the analysis ranged from 57 to 135) and relatively large losses to follow up (range 53.8% to 64.7%). The lack of effect may also be due to the universal application of the intervention to all traumatically injured patients including those who may not be at risk of a secondary disability. In particular, one trial specifically excluded those who were considered at risk of developing mental health problems (Burns 2007), and another trial excluded those who had previously suffered a major psychiatric illness or alcohol abuse (Pirente 2007). In addition, the lack of effect of any of the brief psychological therapies on nonmental health outcomes may be due to the targeted approach of these therapies focusing relatively more strongly on mental health status and coping skills per se rather than on physical health and social functioning.
The ‘light touch’ nature of some of the interventions may explain their lack of effect, in particular the self-help booklet intervention which reported no significant between group differences despite adequate power. It is likely that a short booklet received some weeks after injury is not sufficient to significantly modify feelings and behaviour in order to have a measurable impact on disability outcomes. The only significant effect on a primary outcome in this review was for a complex intervention comprising collaborative care with a dedicated trauma specialist combined with a brief psycho-educational intervention (Zatzick 2001). However the protective effect was not maintained over the medium term once the frequency of contact decreased, possibly suggesting that only complex, time-consuming and therefore costly interventions may have a measurable effect on disability prevention.
There was evidence from three trials that psychosocial interventions have a detrimental effect on the mental health status of some patients (Holmes 2007; Turpin 2005; Zatzick 2001). In one study, subjects with a past history of depression who received interpersonal therapy had significantly higher levels of depressive symptoms at six months (Holmes 2007), while a greater reduction in PTSD ‘caseness’ between baseline and follow up was observed in the control group compared to those who received an educational booklet (Turpin 2005). In addition, while there was a reduction in depressive and PTSD symptoms in those who had received collaborative care compared to the control group one month after injury, these effects were short-lived. Once patient contact had dropped off by four months post-injury, symptoms of depression and PTSD significantly increased in the intervention group relative to the controls (Zatzick 2001). These findings may be partly explained by high rates of drop-out from the intervention group in one of the brief psychological therapy trials (Holmes 2007), perhaps indicating high participant burden resulting in increased participant stress. These results may reflect similar processes to those operating when single session debriefing following a traumatic event leads to increased symptoms of PTSD (Rose 2002). These include ‘secondary traumatisation’ of the injury and ‘medicalising’ normal distress whereby increasing awareness of potential psychological distress may paradoxically induce distress in those who would otherwise not have developed it (Rose 2002).
Two studies showed a trend towards a reduction in mental health symptoms over time in both the control and intervention groups, indicating natural recovery from the psychological consequences of physical trauma (Burns 2007; Pirente 2007), and re-enforcing the conclusion that a universally targeted intervention may impose an unnecessary burden on those who may recover naturally from any psychological trauma.
Taken together, our findings cannot be taken as supporting the provision of psychosocial intervention to prevent aspects of disability arising from physical injury. Our findings may indicate that the monitoring of high-risk patients followed by early intervention, whereby resources are allocated at gradually increasing levels to patients whose difficulties do not abate, may be a better strategy for reducing secondary disability arising from traumatic physical injury.
The quality of trials was generally poor. In particular the quality of reporting of the trials was generally very poor with non-significant effect estimates and the results of many outcomes not reported. With the exception of two trials the sample sizes available for analysis were small, with a large proportion of participants lost to follow up.
The wide range of papers selected from our search criteria and the diverse range of studies eligible for inclusion meant that we had to refine our inclusion criteria from the protocol, for example by removing the restriction that the physical injury must have happened within 12 months of the intervention, as many trials did not report this and we sought to be inclusive in our selection of trials. Instead, the emphasis for inclusion was whether the intervention was aimed at prevention rather than treatment.
The poor quality of reporting of outcomes in the trials may have biased the review as effect estimates for many of the non-significant outcomes were not available despite repeated attempts to contact the authors.
This review contributes to the understanding of psychosocial treatments for the prevention of disability following traumatic injury. This review excluded 62 studies which did not meet our inclusion criteria, indicating a significant body of work in this area. Eleven trials were excluded because they evaluated complex interventions which included but were not restricted to psychosocial interventions. It is possible that the combination of different preventive strategies incorporating medication, psychological interventions and physiotherapy as appropriate may be more effective in preventing the disabling sequelae of traumatic injury than psychosocial interventions alone, though this remains untested in a systematic review.
Seventeen of the excluded studies were trials of the treatment rather than the prevention of disability (primarily mental health problems) following traumatic injury, reflecting the focus of this review on the prevention rather than treatment of secondary conditions. However, given the possible harmful effects that a preventive intervention may have on the mental health status of participants found in this review, in addition to the time and economic cost of such universal interventions, it is important to assess the effect that more targeted interventions may have on ameliorating the negative consequences of traumatic injury. One way of achieving this is to target interventions at groups of patients who have been identified as being at risk of developing a mental or physical disability, or using a stepped care approach to increase the level of intervention based on individual need. A recent systematic review and meta-analysis of early trauma-focused cognitive behavioural therapy to prevent chronic post-traumatic stress disorder and related symptoms concluded that there is evidence for the effectiveness of trauma-focused CBT compared to supportive counselling in preventing chronic PTSD in patients with an initial diagnosis of acute stress disorder. The overall relative risk (RR) for a PTSD diagnosis was 0.56 (95% CI 0.42 to 0.76), 1.09 (95% CI 0.46 to 2.61) and 0.73 (95% CI 0.51 to 1.04) at three to six months, nine months and three to four years post-treatment, respectively, though this evidence came from one research team and therefore needs replication (Kornør 2008). A wider review into psychosocial interventions after crises and accidents found insufficient research evidence on other types of interventions to conclude about effects (Kornør 2007).
This review does not provide convincing evidence for the effectiveness of psychosocial interventions for the prevention of disability following traumatic physical injury. No protective effect of brief psychological therapies or self-help booklets was found. There was moderate evidence from one trial that collaborative care reduced symptoms of depression and PTSD in the short but not medium term. However, there was also evidence that psychosocial interventions may have a detrimental effect on mental health. Nonetheless, these conclusions are necessarily tentative as they are based on a small number of disparate trials with small to moderate sample sizes. More research, using larger sample sizes and comparable interventions and patient populations to enable pooling of results, is needed before these findings can be confirmed. Any such research should undertake intensive monitoring of participants’ short-term response to psychosocial interventions and be vigilant to their potential negative effects.
The heterogeneity of studies included in this review precluded the pooling of data across studies. In order to combine data across trials we recommend further trials of adequate power which focus on comparable psychosocial interventions, patient populations and outcome measures of disability. Interventions which target mental health as a result of traumatic injury are needed, as this remains the most common disabling sequelae of traumatic injury (Mossey 1990; Schnyder 2001; Shalev 1998; Zatzick 2002) and the outcome which holds the most promise for modification by psychosocial interventions.
The research studies reported on in this review were primarily conducted within North America and Europe. Further research on the prevention of disability following physical injury in different cultures and contexts is needed to develop a fuller understanding of efficacious interventions, particularly in low-income countries where the majority of persons with disabilities live (MacLachlan 2009).
More research is needed into stepped care approaches involving the monitoring of patients at risk of developing a mental health problem followed by early intervention whereby resources are allocated at gradually increasing levels to patients whose symptoms do not abate.
Traumatic physical injury such as that resulting from road traffic accidents, falls and fires can cause high levels of subsequent disability in the person affected. This may include physical disability as a result of the initial injury and subsequent complications, mental health problems such as depression, anxiety and post-traumatic stress disorder (PTSD) as a result of the trauma of the event which caused the injury and the resulting physical and social problems, and social problems such as loss of social life and unemployment. It is therefore important to evaluate interventions which seek to prevent these adverse secondary outcomes. Psychosocial interventions, which include psychological therapies such as interpersonal counselling and cognitive behavioural therapy (CBT), and social interventions such as befriending, social support and self-help advice, delivered soon after the injury, may help to prevent these problems.
This review identified five randomised controlled trials, involving 756 participants, which evaluated psychosocial interventions for the prevention of disability following traumatic injury. No convincing evidence was found supporting the efficacy of these interventions. In particular, self-help booklets and interpersonal therapies had no effect on preventing disability. There was some evidence that a more complex intervention involving collaborative care reduced symptoms of depression and PTSD in the short but not the medium term. There was evidence from three trials that psychosocial interventions had a detrimental effect on mental health. Taken together, our findings cannot be taken as supporting the provision of psychosocial interventions to prevent aspects of disability arising from physical injury. These results suggest that future interventions should focus on screening patients at risk of poor outcomes and only treating those who develop subsequent problems. However, the strength of these conclusions is limited by the small size and varied nature of many of the trials, which means that their results cannot be pooled.
This review was supported by a Cochrane Fellowship grant from the Health Research Board (Ireland) to Malcolm MacLachlan.
Mary De Silva is supported by an ESRC/MRC Interdisciplinary Post Doctoral Fellowship.
Vikram Patel is supported by a Wellcome Trust Senior Clinical Research Fellowship in Tropical Medicine.
We thank the Cochrane Injuries Group for assistance throughout the review process, in particular Ian Roberts for comments on the initial review design, Karen Blackhall for conducting the search and Emma Sydenham and Pablo Perel for support throughout the process.
CONTRIBUTIONS OF AUTHORS MM and MDS were jointly responsible for reviewing papers and writing the review.
MDS was responsible for co-ordinating the overall review process, the collation of review papers and for communication with co-authors regarding their feedback.
DD was responsible for offering a third opinion on any disputes between MM and MDS regarding their reviewing of papers.
DD, US, MB, VP and PG were responsible for commenting on the protocol and the final review, and confirming the inclusion of key papers in the area.
DECLARATIONS OF INTEREST None known.
SOURCES OF SUPPORT
Internal sources • No sources of support supplied
Wellcome Trust Senior Clinical Research Fellowship in Tropical Medicine.
DIFFERENCES BETWEEN PROTOCOL AND REVIEW The following text was added:
Secondary outcomes: health care utilisation.
Searching other resources: the reference sections of all selected papers were screened for additional studies.
Because of the types of trials identified, we conducted no quantitative analysis and completed the review using a qualitative summary. Should there be trials in the future that enable a quantitative analysis, these will be performed as per the original protocol.