Intervention Review

You have free access to this content

Psychosocial interventions for the prevention of disability following traumatic physical injury

  1. Mary De Silva1,
  2. Malcolm MacLachlan2,*,
  3. Declan Devane3,
  4. Deirdre Desmond4,
  5. Pamela Gallagher5,
  6. Ulrich Schnyder6,
  7. Muireann Brennan7,
  8. Vikram Patel1

Editorial Group: Cochrane Injuries Group

Published Online: 7 OCT 2009

Assessed as up-to-date: 3 FEB 2008

DOI: 10.1002/14651858.CD006422.pub3


How to Cite

De Silva M, MacLachlan M, Devane D, Desmond D, Gallagher P, Schnyder U, Brennan M, Patel V. Psychosocial interventions for the prevention of disability following traumatic physical injury. Cochrane Database of Systematic Reviews 2009, Issue 4. Art. No.: CD006422. DOI: 10.1002/14651858.CD006422.pub3.

Author Information

  1. 1

    London School of Hygiene & Tropical Medicine, Nutrition & Public Health Intervention Research Unit, London, UK

  2. 2

    Trinity College, University of Dublin, Centre for Global Health and School of Psychology, Dublin, Ireland

  3. 3

    National University of Ireland Galway, School of Nursing and Midwifery, Galway, Ireland

  4. 4

    National University of Ireland, Department of Psychology, Maynooth, Co. Kildare, Ireland

  5. 5

    Dublin City University, School of Nursing, Dublin, Ireland

  6. 6

    University Hospital, Department of Psychiatry, Zurich, Switzerland

  7. 7

    Centers for Disease Control and Prevention, International Emergency & Refugee Health Branch, Atlanta, GA, USA

*Malcolm MacLachlan, Centre for Global Health and School of Psychology, Trinity College, University of Dublin, Dublin, Ireland. malcolm.maclachlan@tcd.ie.

Publication History

  1. Publication Status: New
  2. Published Online: 7 OCT 2009

SEARCH

 
Characteristics of included studies [ordered by study ID]
Burns 2007

MethodsRandomised controlled trial


Participants170 patients over 60 years old who have undergone surgery for hip fracture in orthopaedic units in Manchester, UK, in the previous 2 weeks and who were not depressed (scored 6 or less on the Geriatric Depression Scale (GDS)). 85 in the intervention group and 85 control. Mean age 81 years and 50% were female.


InterventionsUp to 7 sessions of individual CBT delivered by an assistant psychologist, supervised by a clinical psychologist versus treatment as usual.


OutcomesPrimary outcome (assessed at 6 weeks):

  • Depression: Hospital Anxiety and Depression Scale (HADS)


Other outcomes assessed at 6 weeks, 3 and 6 months:

  • Fear of falling: Modified Falls Efficacy Scale
  • Pain: short form McGill pain questionnaire and the Wong-Baker pain rating scale
  • Mobility: Timed-Up-and-Go Test and the modified gait test


NotesThe trial had a separate arm for patients who scored over 6 on the GDS. As these patients had been screened for depression and the CBT was administered to treat rather than prevent depression, this arm of the trial was excluded from this review.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low riskComputer generated randomisation stratified by hospital, block size 4.

Allocation concealment?Low riskIndependent central telephone randomisation scheme.

Incomplete outcome data addressed?
All outcomes
Unclear risk6-week follow up available for 75% of intervention group and 84% of treatment as usual group. No significant differences in baseline characteristics of those lost to follow up and those included.

3-month follow up: 66.5% and 6-month follow up 64.7%

Free of selective reporting?High riskNo reporting of questionnaires other than HADS and new measures not previously introduced. Non-significant effect estimates and many 6-month outcomes not reported.

Free of other bias?High risk1. Selection of non-depressed via the GDS may reduce effectiveness of intervention by restricting the sample to those at low risk of depression

2. Use of anti-depressants in intervention and treatment as usual groups questions validity of psychometrics and/or clinical judgement and creates commonality between groups thereby possibly reducing intervention effect size

3. Small sample size and lack of power for secondary outcomes 

Intention to Treat analysis?Low risk

Blinding of outcome assessmentUnclear riskInsufficient information

Holmes 2007

MethodsRandomised controlled trial.


ParticipantsAll patients admitted to 2 trauma centres in Melbourne, Australia, over an 18-month period, who were 18 years or older and had suffered a major physical trauma (defined as one or more of Injury Severity Score > 15; serious injury to 2 or more body systems; urgent surgery for non-limb injuries; or injuries requiring mechanical ventilation for > 24 hours) were eligible.

90/146 eligible patients randomised (51 intervention, 39 control). Mean age 37 years, 69.3% male.


InterventionsInterpersonal Counselling (IPC), from clinical psychologists with specific training in IPC.  Mean number of sessions 5.9 (SD = 1.1). Treatment as usual comprised seeking help for psychological distress through primary care. On average the control group received an average of 22.6 hours of non-specific psychological support (physical and occupational) and saw a psychologist or psychiatrist for a mean of 0.8 hours.


OutcomesOutcomes assessed at 3 and 6 months:

  • Psychiatric diagnosis: Structured Clinical Interview for DSM IV (SCID)
  • Depression and anxiety: Beck Depression Inventory, Hospital Anxiety and Depression Scale
  • PTSD: Post-Traumatic Checklist
  • Alcohol use: Alcohol Use Disorders Identification Test
  • Pain: visual analogue pain scale


Outcome assessed at 6 months:

  • Health related function: SF-36


NotesThe authors state that "The mean hours of specific psychological intervention (other than IPC) and non-specific therapy did not differ between the groups", the meaning of this is unclear, especially as regards the nature of other types of "specific psychological intervention" that were used as part of treatment as usual.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low riskComputer generated. Intervention allocated in a ratio of 5:4 in expectation of greater losses to follow up in intervention group.

Allocation concealment?Low riskResearch officer made blinded selection from a box of sealed envelopes.

Incomplete outcome data addressed?
All outcomes
High riskDifferential losses to follow up may have affected results with only 53% (27/51) of intervention group completing IPC therapy and available for 6-month follow up compared to 80% (31/39) of the control group. No significant differences between those lost to follow up and those with complete follow up, and those who did and did not complete the intervention.

Free of selective reporting?Low risk

Free of other bias?High risk1. The intervention group received less ‘intervention time’ in terms of psychological support than the control group (mean = 5.9 hours versus 22.6 hours).

2. Participants who commenced but failed to complete therapy had significantly higher alcohol use than those who completed.

3. Low power due to small sample size and high rate of drop- out in intervention group.

Intention to Treat analysis?High risk

Blinding of outcome assessmentUnclear riskPsychiatric diagnosis (primary outcome) assessor blinded at 6 months follow-up. other outcomes not blinded.

Pirente 2007

MethodsRandomised controlled trial.


Participants171 patients admitted to 2 trauma centres in Cologne, Germany with at least 2 injuries with a combined Abbreviated Injury Scale score of >= 6, aged between 18 and 70 years, well orientated in time/person/location at time of contact. 171/184 eligible patients randomised (83 = intervention, 88 = control). Complete outcome data available for 92 patients (45 = intervention, 47 = control). 70.7% male, mean age 38 years.


InterventionsCBT of up to 8 sessions given by a research psychologist trained in CBT with a maximum of 3 sessions per week. Compared to a treatment as usual group (standard hospital care without formal psychosocial intervention). Control groups told they would not receive CBT intervention.


OutcomesPrimary outcome measured at discharge, and at 6 and 12 months:

  • Health Related Quality of Life (HLQOL) composite sum score comprised of parts of several questionnaires: Short Form Health Survey Questionnaire (SF-36); Symptom Check-list-90; Beck Depression Inventory (BDI); State-Trait Anxiety Inventory (STAI) and the social support questionnaire (F-SOZU)


Secondary outcomes measured at discharge and at 6 and 12 months:

  • Depression and anxiety: BDI, SCL90-R and STAI


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low riskComputer generated, stratified for hospital and for mild or no brain injury.

Allocation concealment?Low riskCentral allocation.

Incomplete outcome data addressed?
All outcomes
High riskSubstantial losses to follow up with complete outcome data only available for 92/171 patients randomised. Reasons for losses explained and no significant differences in demographic characteristics were found between those lost to follow up and those with complete follow up.

Free of selective reporting?High riskEffect estimates for social aspects, pain and physical functioning not reported.

Free of other bias?Unclear risk1. Levels of depression and anxiety at baseline were significantly higher for the intervention compared to the control group. Both intervention and control groups improved, but between group differences were not significant. It may be that the intervention is more effective in treating more severe psychological problems than is the control.

2. Rationale for selection of composite HRQOL measure not given. Use of individual scales may be more valid.

Intention to Treat analysis?High riskAnalyses only performed on the 92 patients with complete data rather than on the 171 patients who were randomised.

Blinding of outcome assessmentLow riskOutcomes were assessed using a self-completion postal questionnaire.

Turpin 2005

MethodsRandomised controlled trial.


Participants291 patients 16 to 65 years old who attended Accident and Emergency (A&E) in Sheffield UK who had sustained an injury due to a road traffic accident, occupational injury or assault. 54 in intervention group (34 female, mean age 40), 46 in control group (36, female, mean age 37).

Baseline data collected within 2 weeks of A&E admission after which patients were randomised.


InterventionsSelf-help information booklet (8 pages, 550 words) entitled to 'Response to Traumatic Injury', describing and normalising physiological, psychological and behavioural reactions to traumatic injury. Intervention group patients sent a self-help booklet within 6 to 8 weeks of attendance; control group sent letter without information booklet.


OutcomesOutcomes assessed at 3 and 6 months:

  • PTSD: Post-Traumatic Diagnostic Scale
  • Depression and anxiety: Hospital Anxiety and Depression Scale


NotesThere is no indication of what the control group were told, but it is stated that "control participants were offered a copy of the self-help booklet at the end of the study". Given that the intervention had no positive effect, but did have some negative effects, the rationale for offering the booklets is unclear.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low riskRandom number tables.

Allocation concealment?Low riskMasked independent investigator.

Incomplete outcome data addressed?
All outcomes
High riskReasons for losses to follow up stated. Significant differences in baseline characteristics between responders and non-responders and in the demographic characteristics (age and sex) of those lost to follow up and those with complete data.

Free of selective reporting?Low risk

Free of other bias?High risk1. There was a significantly higher proportion of assaults and occupational injuries in non-responders, and of road traffic accidents in responders.

2. Non-responders were significantly younger and more likely to be male.

3. Only 10% of those eligible agreed to participate (291/2818).

4. An administrative error required 66 of 291 participants who had completed baseline measures, to be removed from the analysis.

Intention to Treat analysis?Low riskMain results reported are not from the intention-to-treat analysis.

Blinding of outcome assessmentLow riskOutcomes were assessed using a self-completion postal questionnaire.

Zatzick 2001

MethodsRandomised controlled trial.


Participants34 road traffic accident or assault related injured patients admitted to a trauma centre in California USA aged between 14 and 65 years and English speaking. 16 in intervention group (8 female, mean age 35.3), 18 in control (33% female, mean age 32.5). 


InterventionsCollaborative care intervention comprising a personally assigned trauma support specialist (1 of 2 psychiatrists or a clinical nurse specialist) who provided support to participants as inpatients and subsequently as outpatients during community rehabilitation. Their role was to facilitate patient-provider treatment planning, and to elicit and track patients post-traumatic concerns. In addition the intervention group received a brief psycho-educational intervention targeting PTSD and substance use. Trauma support specialists spent on average 91 minutes over 4 months with each patient. Control participants received treatment as usual.


OutcomesPrimary outcomes measured at 1 and 4 months post-injury:

  • PTSD using a modified form of the Post-Traumatic Stress Disorder Checklist (PCL-C)
  • Depressive symptoms using a modified form of the Centre for Epidemiological Studies Depression Scale (CES-D)
  • At-risk drinking using a single question from the Addiction Severity Index (ASI)
  • Physical functioning using a modified form of the Physical Components Summary (PCS)


Notes1. Pilot study.

2. Both the intervention and control participants demonstrated high levels of PTSD and depressive symptoms while in hospital.

3. "There were difficulties in implementing the collaborative care principles of continuous case management and active sustained follow-up" and the success of doing so seems to have varied, with those without insurance receiving less integrated care.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low riskComputer generated algorithm using block randomisation with block size of 6.

Allocation concealment?Low riskIndependent project co-ordinator conducted randomisation.

Incomplete outcome data addressed?
All outcomes
Low riskReasons for losses to follow up stated. No significant differences between responders and non-responders.

Free of selective reporting?High riskEffect estimates for at-risk drinking and functional limitations not reported (stated as non significant).

Free of other bias?High riskInpatient length of stay was significantly longer for the intervention group than the control groups (10.6 versus 5.6 days). It is not stated that this was entered as a covariate in the analysis.

Intention to Treat analysis?Low risk

Blinding of outcome assessmentLow riskResearch associates conducting follow-up outcome assessment interviews were blinded to intervention status.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Allen 1996Treatment not prevention study as patients had not suffered from a recent injury

Andersson 2005Not RCT (randomisation of 2 hospitals)

Bisson 2004Treatment of PTSD

Bordow 1979Not RCT (sequential allocation to groups). Same study as Porritt 1979.

Bryant 1998Treatment of acute stress disorder

Bugg (in press)Patients only selected for trial if suffering from acute stress disorder

Castillo 2002Only around 20% of the patients have suffered an injury (patients in intensive care)

Christakou 2007Treatment of ankle sprain, not prevention of disability arising from injury

Corey 1996Complex intervention including physiotherapy, work conditioning and counselling. Unable to disaggregate effects of psychosocial component. Update of Mitchell 1994.

Craig 1998Not RCT

Cramer 2007Not RCT

Cupal 2001Treatment of anterior cruciate ligament injuries

Drechsel-Schlund 2003Patients only included in trial if at high risk of Post Traumatic Stress Disorder (PTSD)

Dunn 2003Not RCT

Dunn 2004Outcome is behaviour change (use of seat belts) rather than disability

Ehlers 2003Treatment of PTSD

Evans 1998Treatment rather than prevention of disability and cannot disaggregate those who suffered a traumatic injury from participants with other disabling conditions

Evans 2001Update of Evans 1998 trial. Treatment rather than prevention of disability and cannot disaggregate those who suffered a traumatic injury from participants with other disabling conditions.

Fauerbach 2002Outcome is not disability but reduction in pain during dressing change for a burns wound

Fecteau 1999Treatment of PTSD following a motor vehicle accident

Foa 1995Not RCT

Fronek 2005Participants have not suffered a traumatic physical injury (staff training intervention)

Girolami 2005Not RCT

Hagglund 2007Outcome is not disability but rate of re-injury

Hagsten 2006Intervention is physiotherapy, not psychosocial

Hazard 2000Participants have not suffered a traumatic physical injury

Jensen 2001Patients have not suffered a recent injury and the intervention is complex and the psychosocial component cannot be disaggregated.

Kennedy 2003Not RCT (control group data taken from an existing database)

King 1999Not RCT (matched controls)

Kwon 2006Not RCT (controls and intervention group matched)

Latimer 2006Patients have not suffered a recent injury

Lindstrom 1992The intervention is complex and the psychosocial component cannot be disaggregated

McFarlane 2006Outcome is not disability but prevention of domestic violence

Melnyk 2004Unable to disaggregate those who suffered a traumatic injury from participants with other disabling conditions.

Menzel 2006Participants have not suffered a traumatic physical injury

Miller 1975Not RCT (patient choice as to which group they were assigned)

Mitchell 1994Complex intervention including physiotherapy, work conditioning and counselling. Unable to disaggregate effects of psychosocial component. First report of Corey 1996 study.

Moore 1983Treatment of burn wounds rather than prevention of disability

Norman 2004Patients have not suffered a recent injury (chronic pelvic pain)

Oliveira 2006Not RCT (alternate allocation)

Ottosson 2007The intervention is complex and the psychosocial component cannot be disaggregated

Pain 2007The intervention is complex and the psychosocial component cannot be disaggregated

Phillips 2001The intervention is complex and the psychosocial component cannot be disaggregated

Porritt 1979Not RCT (sequential allocation). Same study as Bordow 1979

Ross 1996Not RCT (alternate allocation)

Rotem-Lehrer 2007Treatment of ankle sprain rather than prevention of disability

Rottkamp 1976Intervention is physiotherapy rather than psychosocial

Rowland 2006The intervention is complex and the psychosocial component cannot be disaggregated

Scholes 2007Treatment rather than prevention as patients screened for acute stress disorder

Scholten-Peeters 2006Treatment for the symptoms of whiplash

Sirles 1991Participants have not suffered a traumatic physical injury

Smith 1984Outcome is not disability but prevention of child abuse

Soderstrom 2007Outcome is not disability but prevention of at-risk drinking

Söderlund 2007Treatment for the symptoms of whiplash

Ventegodt 2004The intervention is complex and the psychosocial component cannot be disaggregated

Vick 2001Not RCT

Vick 2004Not RCT

Wagner 2007Treatment for PTSD and depression

Wise 2002The intervention is complex and the psychosocial component cannot be disaggregated.

Yates 2000Not RCT

Zatzick 2004Treatment of PTSD

Zemper 2003The intervention is complex and the psychosocial component cannot be disaggregated. Not all patients suffered a recent physical injury.

 
Characteristics of studies awaiting assessment [ordered by study ID]
Humphreys 2003

Methods

Participants

Interventions

Outcomes

NotesUnable to locate study.

McKinlay 2003

Methods

Participants

Interventions

Outcomes

NotesUnable to locate study.

Tecic unpublished

MethodsRCT

Participants113 severely injured trauma patients from 4 German trauma centres

InterventionsShort and long-term (up to 6 months post-discharge) psychotherapy compared to short-term (in hospital) psychotherapy

OutcomesDepression, anxiety, PTSD

NotesUnpublished study