Description of studies
See: Characteristics of included studies; Characteristics of excluded studies.
The most current search on electronic databases was conducted in March 2006 and this yielded a total of 3038 references. A total of three RCTs were identified which met the inclusion criteria. An additional 25 trials addressing anxiety after acquired brain injury were identified but these were excluded because they were either not an RCT or did not include a sample consisting of 80% or greater of participants with TBI. Investigation of other specified sources for identifying studies yielded no additional studies meeting inclusion criteria. Two of three identified trials, by Bryant 2003 and Tiersky 2005, met inclusion criteria from information reported. The third trial by Helffenstein 1982 did not explicitly state that participants of the study included people with TBI. After correspondence with the trial author it was confirmed that all study participants had sustained a TBI and the trial was included in the review.
The three included studies examined the corresponding group comparisons as follows.
CBT versus supportive counselling (SC) (Bryant 2003).
Interpersonal process recall (IPR) therapy versus controls with no feedback on interpersonal functioning (Helffenstein 1982).
CBT combined with neurorehabilitation (NR) versus no psychological intervention controls (Tiersky 2005).
This study examined psychological therapy for people with mild TBI diagnosed with a specific anxiety disorder (acute stress disorder) comparing CBT (n = 12; eight females) and SC (n = 12; eight females) groups. The mean (SD) age for the CBT group was 29.4 (13.9) years and for SC group was 33.0 (14.4) years. Table 1 Primary outcome measures were;
Table 1. Characteristics of CBT and SC groups (Bryant 2003)
|Age||29.42 (13.93)||33.00 (14.37)|
|Sex||4 males, 8 females||4 males, 8 females|
|Prior psychiatric diagnosis||1 major depression, 1 bulimia nervosa||1 major depression, 1 attention defict disorder|
|Pre-treatment Acute Stress Disorder Interview||65.42 (10.60)||62.42 (14.58)|
|Pre-treatment IES (Intrusion subscale)||27.83 (5.31)||24.50 (8.20)|
|Pre-treatment IES (Avoidance subscale)||20.58 (5.02)||16.25 (7.24)|
|Pre-treatment BAI||25.58 (11.43)||26.83 (13.90)|
Clinician Administered PTSD Scale (CAPS) frequency and intensity subscales,
Impact of Event Scale (IES) intrusions and avoidance subscales, and
Beck Anxiety Inventory (BAI).
This study examined 13 males and three females with TBI aged between 17 and 35 years. Table 2 Severity of TBI of participants was not indicated. IPR therapy (n = 8) and control (n = 8) groups were examined using the State-Trait Anxiety Inventory (STAI; state and trait forms) as the primary outcome measure.
Table 2. Characteristics of IPR and control groups (Helffenstein 1982)
|Characteristics||IPR group||Control group||Notes|
|Age||Information not provided||Information not provided||17 to 35 years across both groups|
|Sex||Information not provided||Information not provided||13 males, 3 females across both groups|
|Pre-treatment STAI score||Information not provided||Information not provided||Reviewers contacted authors of this paper requesting mean age, sex and pre-treatment STAI scores across groups, however, no response was provided.|
This study focused on mild to moderate TBI comparing CBT and NR (n = 11; five females) and control (n = 9; six females) groups. Table 3 The primary outcome measure was the anxiety subscale of the Symptoms Checklist - 90R (SCL-90R). Mean (SD) age of participants in CBT and NR and control groups were 47.55 (11.78) and 46.00 (9.35) years respectively.
Table 3. Characteristics of CBT/NR and control groups (Tiersky 2005)
|Age||47.55 (11.78)||46.00 (9.35)|
|Sex||5 females, 6 males||6 females, 3 males|
|Pre-treatment SCL-90R anxiety subscale||0.921 (0.85)||1.39 (0.70)|
Risk of bias in included studies
Methodological quality for the RCTs was independently rated by two authors using the PEDro scale (Maher 2003). The mean kappa coefficient between the two raters for total PEDro scores for the three included studies was 0.81. Kappa coefficients for each of the three studies were: 1 (P < 0.001) for Bryant 2003, 0.65 (P < 0.05) for Helffenstein 1982 and 0.79 (P < 0.01) for Tiersky 2005. Scores for each item on the PEDro scale for the three included studies were reported in Table 4. Both studies by Bryant 2003 and Tiersky 2005 achieved a total score of 6/10 and the study by Helffenstein 1982 achieved a total score of 3/10.
Table 4. Methodology quality assessed by the PEDro scale
|Criterion||Bryant 2003||Helffenstein 1982||Tiersky 2005|
|1. Eligibility criteria were specified||Y||Y||Y|
|2. Participants were randomly allocated to interventions (in a crossover study, subjects were randomly allocated an order in which treatments were received)||Y||Y||Y|
|3. Allocation was concealed||N||N||Y|
|4. The intervention groups were similar at baseline regarding the key outcome measure(s) and most important prognostic indicators||Y||N||Y|
|5. There was blinding of all participants||N||N||N|
|6. There was blinding of all therapists who administered the therapy||N||N||N|
|7. There was blinding of all assessors who measured at least one key outcome||Y||Y||Y|
|8. Measures of at least one key outcome were obtained from more than 85% of the participants initially allocated to groups||Y||Y||N|
|9. All participants for whom outcome measures were available received the treatment or control condition as allocated or, where this was not the case, data for at least one key outcome was analysed by "intention to treat"||N||N||N|
|10. The results of between- intervention group statistical comparisons are reported for at least one key outcome||Y||N||Y|
|11. The study provides both point measures and measures of variability for at least one key outcome||Y||N||Y|
|Total (Items 2-11)||6/10||3/10||6/10|
Performance of the three studies on the 11 items of the PEDro scale were as follows.
all three studies specified eligibility criteria for study inclusion,
all three studies randomly allocated participants to interventions,
only the study by Tiersky 2005 stated that allocation to interventions was concealed,
studies by Bryant 2003 and Tiersky 2005 indicated that intervention groups were similar at baseline regarding the key outcome measures and most important prognostic indicators,
there was no blinding of subjects in any of the three studies,
there was no blinding of therapists who administered the therapy in any of the three studies,
all three studies blinded assessors who measured at least one key outcome,
studies by Bryant 2003 and Helffenstein 1982 indicated adequate follow up with at least one key outcome obtained from more than 85% of participants initially allocated to groups, whereas Tiersky 2005 followed up only 69% of participants,
none of the three trials stated that all participants received the treatment or control condition as allocated or that data for at least one key outcome were analysed by intention-to-treat principles,
and 11. studies by Bryant 2003 and Tiersky 2005 reported results of between-intervention group statistical comparisons, as well as point measures and measures of variability for at least one key outcome.
Effects of interventions
There was too much heterogeneity among the studies to permit meta-analysis. Specifically, the study populations and interventions of the three identified studies were judged to be sufficiently inconsistent to allow pooling of the data. The study by Bryant 2003 was examined separately because it included participants diagnosed with acute stress disorder, whereas the studies by Helffenstein 1982 and Tiersky 2005 did not include participants with a diagnosed anxiety disorder. Rather, these two studies targeted participants with anxiety symptomatology. The interventions examined by Helffenstein 1982 and Tiersky 2005, however, were also too dissimilar to allow pooling of data. The intervention in the study by Helffenstein 1982 focused on interpersonal communication and interaction whereas the intervention examined by Tiersky 2005 was broader, encompassing both CBT and neurorehabilitation. (Table 5; Table 6)
Table 5. Summary of timing of assessments
|Study||Time 1||Time 2||Time 3||Time 4||Comments|
|Bryant 2003||baseline||post-treatment||6-month follow up|| || |
|Helffenstein 1982||baseline||post-treatment||1-month follow up|| ||Data from 1-month follow up were not analysed due to small sample (n=6)|
|Tiersky 2005||baseline||post-treatment||1-month follow up||3-month follow up||Post-treatment, 1-month and 3-month follow-up assessments were combined and an averaged score was used as outcome score|
Table 6. Summary of secondary outcome variables
|Study||Affective/self||Neuropsychological||Psychosocial||Daily funct/particip||Med/service usage||Treatment compliance|
|Bryant 2003||BDI||N/A||N/A||N/A||N/A||No dropouts|
|Helffenstein 1982||TSCS||N/A||ICI, IRRS, independent observer report scale of interpersonal interaction, videotape analysis of interpersonal interaction||N/A||N/A||No dropouts|
|Tiersky 2005||SCL-90R Global Severity Index, SCL-90R depression subscale, CRI||PASAT, RAVLT, ACFI, AQ||N/A||CIQ||N/A||3 dropouts from treatment group, 6 from the control group|
This study examined psychological therapy for people with mild TBI diagnosed with a specific anxiety disorder, comparing CBT and SC groups which were similar at baseline in terms of anxiety symptomatology and presence of acute stress disorder. Post-treatment and six-month follow-up scores on anxiety outcome measures, controlling for pre-treatment symptom severity, were compared across groups. Results indicated that fewer participants met the criteria for post-traumatic stress disorder (PTSD) in the CBT group than the SC group, both at post-treatment (χ2 = 6.75, df = 1, P < 0.05) and six-month follow up (χ2 = 4.44, df = 1, P < 0.05). Additionally, significant differences were found between groups at post-treatment on the CAPS frequency (F = 12.88, df = 1, 23, P < 0.01) and intensity (F = 11.54, df = 1, 23, P < 0.01) subscales, IES intrusions (F = 9.44, df = 1, 23, P < 0.01) and avoidance (F = 47.49, df = 1, 23, P < 0.01) subscales and the BAI (F = 3.99, df =1, 23, P = 0.05) with the CBT group achieving lower (better functioning) scores compared to the SC group on all of these measures. Maintenance of treatment gains were reported at six-months follow-up, particularly in relation to PTSD symptoms, with the CBT group achieving significantly lower scores on the frequency (F = 5.40, df = 1, 23, P < 0.05) and intensity ( F = 6.75, df = 1, 23, P < 0.05) subscales of the CAPS and the intrusions (F = 6.26, df = 1, 23, P < 0.05) and avoidance (F = 9.58, df = 1, 23, P < 0.01) subscales of the IES. There was, however, no evidence of maintenance of gains in terms of anxiety symptomatology, with no significant difference between groups on BAI scores at the six-month follow-up assessment.
Comparison between IPR therapy and control groups in the second study showed a greater reduction in reported anxiety as measured by the STAI-trait form in the treatment group (F = 7.55, df = 1, 14, P = 0.02). No statistically significant differences in the reduction of anxiety were found on the STAI-state form. This study, however, did not appear to perform a between-group statistical comparison on the anxiety measures, examining only within group comparison across time.
The third study compared CBT with NR and control groups using univariate analyses of covariance with baseline average as a covariate. Significant effects were found for anxiety as indexed by the SCL-90R anxiety subscale with lower levels of anxiety in the treatment group (P < 0.05).