The Efficacy of Cognitive Behavior Therapy in the Treatment of Emotional Distress After Acquired Brain Injury

Participants 

The study protocol was approved by the research ethics board at the institution at which the study was conducted, and the procedures of the study were in accordance with the standards of the research ethics board.

A total of 20 patients were initially recruited from a community-based ABI treatment organization for patients with the chronic effects of ABI. Inclusion criteria were as follows: (1) age between 18 and 65 years, (2) greater than 1 year postinjury, (3) able to provide informed consent, (4) on stable dosage of psychoactive medication and being monitored by a physician or psychiatrist, and (5) scored at least 1 SD above the published mean on the SCL-90-R36 GSI.

Individuals were excluded from the study if they (1) were endorsing significant suicidal ideation at the time of evaluation, (2) were currently engaged in another CBT or other psychotherapeutic intervention (not including sporadic support from social work, psychiatry, case manager, or other counselor), (3) had a concurrent neurological or psychotic disorder, or (4) were diagnosed with a communication disorder that would preclude participation.

Twenty-nine patients were approached to participate in the study. Of these, 7 declined. Of the 22 who agreed to participate, 20 met the SCL-90-R cutoff. From these patients, 2 groups were created who were matched on SCL-90-R scores, age, sex, and years of education. All participants in the groups had sustained significant brain injuries. Patients with TBI had initial GCS scores in the moderate or severe range. All patients with nontraumatic injuries were currently operating in the moderate to severe range of cognitive impairment in at least 1 cognitive domain and/or had remained in inpatient treatment for more than double the provincial average length of stay (27 days).

Within each group, half of the participants were assigned to a conventional group format for their sessions, and half were assigned to a telephone-administered format; this latter assignment was based on logistical considerations, with patients traveling the farthest distances or without reliable transportation to the treatment facility receiving the telephone administration. (After the study commenced, 1 participant in the telephone-administered CBT condition withdrew and was subsequently replaced by a new recruit.) Table 1 shows the demographic, medical, and injury variables of the 2 groups. The CBT and education control groups did not differ significantly across any of the following parameters: age, years of education, time since injury, estimated premorbid intelligence quotient as measured by the WTAR,37 or current cognitive functioning as measured by the RBANS.38

Table 1. Demographic, Medical, and Injury Variables

		CBT Group (n=10)	Education Control (n=10)
	Age, mean ± SD (y)	39.8±10.44	42.50±13.01
	Sex, n (%)
	Female	5(50)	5(50)
	Male	5(50)	5(50)
	Years postinjury, mean ± SD	7.00±6.15	11.40±9.42
	Injury type
	Trauma, n (%)	7(70)	3(30)
	MVC	5	1
	Pedestrian MVC	1	1
	Fall	1	1
	Nontrauma, n (%)	3(30)	7(70)
	Aneurysm/AVM/stroke	2	3
	Hypoxia/anoxia	1	1
	Anaphylaxis	0	1
	Tumor	0	2
	Psychotropic medication, n (%)
	No medication	4(40)	3(30)
	Antidepressant	5(50)	6(60)
	Antipsychotic/mood stabilizer	3(30)	6(60)
	Benzodiazepine	2(20)	2(20)
	Stimulant	0(0)	2(20)
	Marital status, n (%)
	Single	8(80)	4(40)
	Married/common law	1(10)	4(40)
	Divorced/separated	1(10)	2(20)
	Years of education, mean ± SD	13.10±1.66	14.00±1.89
	Preinjury employment, n (%)
	Employed	8(80)	6(60)
	Student	2(20)	2(20)
	Unemployed	0(0)	2(20)
	Postinjury employment, n (%)
	Employed	1(10)	0(0)
	Student	0(0)	1(10)
	Unemployed	9(90)	9(90)
	Estimated premorbid IQ, mean ± SD
	WTAR standard score	95.40±11.58	102.8±13.47
	Cognitive functioning, mean ± SD
	RBANS total index	66.10±16.37	73.20±15.62
	Immediate memory	71.10±20.96	75.80±17.64
	Attention	68.50±21.10	73.00±20.36
	Language	79.50±13.06	83.00±16.55
	Visuospatial	76.50±19.80	86.70±16.67
	Delayed memory	64.30±25.29	75.40±20.38

Abbreviations: AVM, arteriovenous malformation; MVC, motor vehicle collision; IQ, intelligence quotient.

There were 4 subgroups in total. The treatment group comprised a telephone- administered subgroup (telephone CBT; n=5) and a face-to-face, group format subgroup (group format CBT; n=5). The education control group was composed of a telephone-administered subgroup (telephone education control; n=5) and a group-administered subgroup (group education control; n=5).

Materials 

Prior to treatment, all participants were administered the following neuropsychological tests to confirm similar levels of premorbid intellectual functioning and current cognitive functioning, and to provide clinical information with which to tailor therapy: (1) the WTAR,37 an estimate of premorbid intelligence; and (2) the RBANS,38 a brief, valid, and reliable measure of neuropsychological functioning that evaluates performance across a range of cognitive domains, providing summary index scores for each.

Primary Outcome Measures 

Secondary Outcome Measures 

Design and Procedures 

The design was a pre/posttreatment, education controlled design, comparing the effects of CBT to education both within and between groups. The modalities were commensurate on pretreatment SCL-90-R means, with no significant differences, similar distributions, and minimal effect sizes. Prior to treatment commencing, neuropsychological tests and baseline psychologic outcome measures were administered. Then, the first session for all participants took place in person, 1-on-1, with their therapist or educator. At this time, all procedures were explained and materials were distributed.

The subsequent 10 treatment/education sessions were conducted either over the telephone (1-on-1 format) or in the face-to-face, group format. The sessions took place on a weekly basis with some sessions occurring twice weekly because of time constraints and scheduling. All sessions were completed within a 9-week time frame, with sessions ranging from 45 to 75 minutes as required. All participants completed the DASS-21 at the beginning of each weekly session to monitor progress and suicide risk. After completion of the eleventh session, all participants completed the posttreatment outcome measures. Participants in both groups also received 1-month follow-up sessions, at which time only the primary outcome measures (SCL-90-R and DASS-21) were administered.

Statistical Analyses 

Hypothesis 1 

A significant group by time interaction was observed for the SCL-90-R (F1,18=7.03, P<.05), with planned comparisons showing no significant pretreatment differences (t18=–0.40, P=.69, 1-tailed, Cohen's d=–0.18), but significant posttreatment differences (t18=–2.95, P<.01, 1-tailed Cohen's d=−1.32). Further, the CBT group showed a significant reduction in distress (t9=4.11, P<.01) with a large effect size (Cohen's d=1.30). Similarly, with the DASS-21 total score as a dependent variable, a significant group by time interaction was observed (F1,18=7.56, P<.05), with nonsignificant pretreatment differences between the 2 groups (t18=–1.01, P=.33, although a moderate effect size was observed, Cohen's d=–0.48), but significant posttreatment differences (t18=–0.93, P<.001, with a very large effect size Cohen's d=–1.76). Significant improvement on the DASS-21 was also observed from pretreatment to posttreatment in the CBT group (t9=6.27, P<.001, Cohen's d=1.79). Moreover, although the mean pretreatment DASS-21 scores indicated a moderate range of emotional disturbance, relative to normative data, posttreatment DASS-21 total scores for the CBT group fell within the normative range for emotional disturbance. See figure 1 for group differences on DASS-21 scores across all sessions.

View Large Image Download to PowerPoint

Fig 1. Group differences on DASS-21 total score over time. Performances in each of the 4 groups are presented across the 10 treatment/education sessions. A lower score represents less emotional distress. G-CBT, Group CBT; T-CBT, telephone CBT; G-Ctrl, group education control; T-Ctrl, telephone education control; Pre-TX, pretreatment; Post-TX, posttreatment.

Table 2 displays the descriptive statistics for the subgroups on the primary outcome measures. On the SCL-90-R GSI, the group format CBT patients showed significant improvement over time as predicted (t4=3.67, P<.01, 1-tailed, with a large Cohen's d effect size: Cohen's d=1.45) as did the telephone CBT patients (t4=2.20, P<.05, 1-tailed, with a large effect size: Cohen's d=1.06). On the DASS-21 total score, significant reductions in emotional distress with large effect sizes were observed for both the group format CBT subgroup (t4=5.03, P<.01, 1-tailed, Cohen's d=1.91) and the telephone CBT subgroup (t4=3.07, P<.05, 1-tailed Cohen's d=1.46). Further, there were no significant between-group differences at posttreatment, when the group format CBT and telephone CBT groups were compared on either the SCL-90-R (t8= 0.19, P=.85, Cohen's d=0.12) or the DASS-21 (t8=–0.78, P=.46, Cohen's d=–0.49). Last, there were no significant improvements, or trends toward improvement, after treatment in either of the education control groups on either test.

Table 2. Individual Group Analyses for Primary Outcome Measures

	Outcome Measures	G-CBT (n=5) Mean ± SD	T-CBT (n=5) Mean ± SD	G-Ctrl (n=5) Mean ± SD	T-Ctrl (n=5) Mean ± SD
	SCL-90-R, GSI
	Pre	71.00±7.84	68.40±8.26	72.40±9.15	69.80±6.98
	Post	59.20±8.41	58.00±11.11	68.00±7.97	72.80±9.28
	DASS-21
	Total score
	Pre	42.80±21.15	46.00±13.41	60.40±36.89	49.20±17.81
	Post	12.00±8.60	19.60±19.97	42.80±25.00	62.80±24.68
	Depression
	Pre	21.20±14.87	13.60±7.67	23.20±10.35	18.40±7.80
	Post	6.80±6.57	5.20±5.76	16.00±11.31	22.80±10.26
	Anxiety
	Pre	8.00±3.74	11.20±8.07	16.00±14.63	14.80±5.93
	Post	2.00±3.46	3.60±3.58	7.60±6.69	16.80±8.44
	Stress
	Pre	13.60±9.94	21.20±3.03	21.20±13.31	16.00±6.78
	Post	3.20±3.35	10.80±11.28	19.20±11.19	23.20±8.44

Abbreviations: G-CBT, Group CBT; T-CBT, telephone CBT; G-Ctrl, group education control; T-Ctrl, telephone education control.

Scores on the secondary outcome measures for community integration, emotion-focused coping, and problem-focused coping, all changed in the expected directions—that is, improvements were noted for the most part in the CBT group, but not the education control group. Most of these results were trends toward significance. On total CIQ score, results revealed a trend toward a group by time interaction (F1,18=4.34, P=.052), and similar results were revealed on the Ways of Coping Scale, Revised, Escape Avoidance subscale (F1,18=3.94, P=.063). No group by time interaction effect was revealed on the Ways of Coping Scale, Revised, Planful Problem Focused subscale (F1,18=0.22, P=.65), but a main effect of time was shown (F1,18=7.81, P<.05). Therefore, both groups seemed to improve their problem-solving approach by being involved in CBT or learning about brain injury. However, only the CBT group showed trends toward declines in emotion-focused coping.

Hypothesis 2 

As shown in figure 2, when pretreatment and 1-month follow-up outcomes on the SCL-90-R GSI were examined, a significant group by time interaction was observed (F1,18=18.60, P<.001). As previously stated, the groups had commensurate pretreatment scores. However, significant differences were revealed at follow-up on a 1-tailed t test (t18=–3.38, P<.01, Cohen's d=–1.51), with the treatment group showing significantly better functioning than the control group. As well, significant improvement within the CBT group from pretreatment to follow-up was observed (t9=5.17, P<.001, 1-tailed, Cohen's d=1.30).

View Large Image Download to PowerPoint

Fig 2. Differences on SCL-90-R GSI from pretreatment to follow-up across collapsed CBT groups and education control groups.

As shown in figure 3, similar results were revealed for the DASS-21, with a significant group by time interaction observed (F1,18=12.65, P<.01). The CBT group demonstrated significant reductions in distress (DASS-21 total score) compared with the control group (t18=–4.65, P<.001, Cohen's d=–2.08), and significant improvement within the CBT group from pretreatment to 1-month follow-up was also found (t9=7.01, P<.001, 1-tailed, Cohen's d=1.85).

View Large Image Download to PowerPoint

Fig 3. Differences on DASS-21 total score from pretreatment to follow-up across collapsed CBT groups and education control groups.

Importantly, no change was found between posttreatment and 1-month follow-up measures for the CBT group on either the SCL-90-R GSI (t9=–0.89, P=.20, Cohen's d=0.25) or the DASS-21 total score (t9=–0.06, P=.48, Cohen's d=–0.02), demonstrating the stability of group effects over time. The stability of positive change was also upheld, on both outcome measures, for subgroup analyses. The group format CBT participants showed commensurate posttreatment and follow-up scores on the SCL-90-R GSI (t4=0.44, P=.34, Cohen's d=0.19), and the DASS-21 total score (t4=–0.12, P=.45, Cohen's d=–0.05). The telephone CBT participants showed similar results on the SCL-90-R GSI (t4=0.75, P=.25, Cohen's d=0.27) and DASS-21 (t4=0.00, P=.50, Cohen's d=0.00).

Importantly, there were no significant differences between the group format CBT and telephone CBT groups at follow-up on either the SCL-90-R GSI (t8=.32, P=.76, Cohen's d=–0.20), or the DASS-21 (t8=–.82, P=.44, Cohen's d=–0.52). See table 3 for individual change scores and clinical impact for individual participants.

Table 3. Individual Scores on SCL-90-R GSI and DASS-21, Pretreatment Versus Posttreatment Versus 1-Month Follow-Up

	Patient	GSI Pre	GSI Post	GSI Follow-Up	DASS-21 Pre, Total Score	DASS-21 Post, Total Score	DASS-21 Follow-Up
	CBT treatment group
	1⁎	63	55	41	16	0	8
	2⁎	81	58	71	62	22	14
	3⁎	74	68	64	56	18	26
	4⁎	63	48	50	24	8	4
	5⁎	74	67	60	56	12	10
	6	65	66	58	44	38	18
	7	72	71	72	62	44	44
	8	81	59	60	56	10	30
	9	60	50	30	40	0	0
	10	64	44	52	28	6	6
	Education control group
	1⁎	65	60	58	54	20	12
	2⁎	74	66	68	24	26	48
	3⁎	81	81	81	88	82	90
	4⁎	81	69	73	108	52	70
	5⁎	61	64	67	28	34	50
	6	72	61	71	58	82	76
	7	65	65	74	38	32	32
	8	64	81	72	24	40	44
	9	67	76	70	68	78	86
	10	81	81	81	58	82	84

⁎ Group-administered subgroup patients for group cognitive behavior therapy and group education control.

Study Limitations 

There are obvious limitations to the current study. First, the current study was an education-matched study, not a randomized controlled trial; therefore, caution should be applied in the extrapolation and generalization of the findings outside the context of the current study. Another very pertinent limitation was the small sample size, again making it difficult to interpret the current findings confidently. Results were nevertheless compelling, and given the current strength of the findings, additional investigations are warranted.

An additional limitation of this study was the representativeness of the sample. All individuals who participated in the current project were connected to a community ABI program and lived in urban areas with access to high-quality health care. Therefore, the sample tested is not representative of all individuals with ABI. Finally, sample bias is a potential limitation. Individuals who chose to participate in the current study may have been more psychologically minded or at a stage in their recovery at which they were more receptive to psychologic intervention. Results suggest that for individuals motivated to participate, the intervention has the potential to be efficacious and that the results can be lasting.

Clinical Implications 

Overall, these findings are very promising. Efficacy of the adapted CBT protocol for people with ABI was observed on the primary outcome measures even with very small numbers. Although there was diminution of emotional dysfunction in the CBT subgroups and the CBT collapsed group, there was no significant change to symptoms in the education control group. These preliminary results strongly support the value of a larger randomized controlled trial, especially given that the current protocol is unique in that it is universally applicable to all patients with ABI regardless of accessibility restrictions that may be present. Moreover, research has shown that patients with ABI do not necessarily respond as well to traditional therapeutic interventions when their unique cognitive and rehabilitative needs are not formally recognized and addressed.2 The current approach adheres to proven CBT treatment protocols aimed at reducing emotional distress and improving adjustment and coping, but critically, it has been adapted to address the specific needs of the ABI population. An additional potential benefit of this research is that CBT has the potential to promote cognitive and physical functioning, either by enhancing mood (which when significantly disrupted can reduce cognitive functioning) or by promoting greater motivation for and engagement in rehabilitation. This in turn may further enhance coping and adjustment after ABI. The current line of research has the potential to influence standard rehabilitation practices positively by demonstrating the benefits of allocating appropriate resources to emotional recovery after ABI.