If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
McDonald S, Tate R, Togher L, Bornhofen C, Long E, Gertler P, Bowen R. Social skills treatment for people with severe, chronic acquired brain injuries: a multicenter trial.
To determine whether social skills deficits including unskilled, inappropriate behavior, problems reading social cues (social perception), and mood disturbances (such as depression and anxiety) could be remediated after severe traumatic brain injuries.
Randomized controlled trial comparing a social skills program with social activity alone or with waitlist control. Several participants were reassigned after randomization.
Hospital outpatient and community facilities.
Fifty-one outpatients from 3 brain injury units in Sydney, Australia, with severe, chronic acquired brain injuries were recruited. A total of 39 people (13 in skills training, 13 in social activity, 13 in waitlist) completed all phases of the study.
Twelve-week social skills treatment program encompassing weekly 3-hour group sessions focused on shaping social behavior and remediating social perception and 1-hour individual sessions to address psychologic issues with mood, self-esteem, etc.
Main Outcome Measures
Primary outcomes were: (1) social behavior during encounters with a confederate as rated on the Behaviorally Referenced Rating System of Intermediary Social Skills−Revised (BRISS-R), (2) social perception as measured by The Awareness of Social Inference Test, and (3) depression and anxiety as measured by the Depression, Anxiety and Stress Scale. Secondary outcomes were: relative report on social behavior and participation using: the Katz Adjustment Scale−R1; the Social Performance Survey Schedule; the La Trobe Communication Questionnaire; and the Sydney Psychosocial Reintegration Scale (both relative and self-report).
Repeated-measures analysis of variance indicated that social activity alone did not lead to improved performance relative to waitlist (placebo effect) on any outcome variable. On the other hand, the skills training group improved differentially on the Partner Directed Behavior Scale of the BRISS-R, specifically the self-centered behavior and partner involvement behavior subscales. No treatment effects were found for the remaining primary outcomes (social perception, emotional adjustment) or for secondary outcome variables (relative and self-report measures of social function).
This study suggested that treatment effects after social skills training in people with severe, chronic brain injuries are modest and are limited to direct measures of social behavior.
PSYCHOSOCIAL PROBLEMS are common after severe TBI. Deficits include insensitivity to others, socially inappropriate behavior, and egocentricity, or, on the other hand, a general failure to initiate social contact or to respond appropriately in social contexts.
These poor social behaviors lead to unsuccessful social encounters and disinclination on the part of others to initiate or maintain social contact, whether in the workplace or school or during leisure and social activities. The cost of these poor social skills is profound.
Traditional social skills techniques use explicit reinforcement to increase appropriate versus inappropriate behavior and have also been effective in producing desirable behavior changes in people with severe TBI.
Other behavioral techniques such as role plays, video feedback, and cues to assist self-monitoring have both increased the number of positive behaviors, for example, active listening, and eliminated or significantly reduced the frequency of negative behaviors, such as interrupting and making inappropriate sexual remarks, after TBI.
Although there are many such positive findings, relatively few published studies have used properly controlled trials to evaluate the efficacy of social skills training for people with TBI. Indeed, to date, only 2 RCTs have been reported. The first, by Helffenstein and Wechsler,
evaluated the effects of role plays and feedback to target interpersonal and communication skills in 8 people with TBI assigned to treatment versus 8 assigned to “nontherapeutic attention.” Despite the small numbers, findings were encouraging, with improvements noted in anxiety, self-concept, and ratings of social skills. Measures of more general improvements in social participation and behavior were not reported. A more recent study by Dahlberg et al
used an RCT with 52 participants allocated to a treatment versus waitlist (deferred treatment) to evaluate the effectiveness of a manualized approach to social skills training for people with chronic, severe TBI. Post treatment, improvements were reported in social behavior as rated by independent observers using the Profile of Functional Impairment in Communication. Broader measures of social outcome (secondary outcome variables) were taken but revealed no change. This study is the largest and most sophisticated to date, but it was limited by the lack of a control for the general effects of “nontherapeutic” attention. In addition, whereas both previous RCTs focused treatment on social behavior, there are other factors that potentially impact on social skills which may also be targeted in an effective intervention.
First, social competence requires not only a good repertoire of skilled social behavior but also the ability to “read” the social context accurately so as to know when and how to respond.
Consequently, traditional behavioral approaches to social skills training in people with TBI can be enhanced by the inclusion of training in emotion perception, that is, training to discriminate and recognize emotional expressions and to learn what these might mean in a social context. Emotion perception training has proven to be efficacious in schizophrenia,
Another important issue is the presence and influence of reactive and pre-existing psychologic problems. High rates of depression, dysthymia, bipolar disorder, substance abuse disorders, and anxiety disorders have been reported after TBI.
Psychologic disorders may represent a significant impediment to social opportunities, but can also directly impact on levels of social competence. For example, depressed persons without brain damage interact with others less often
Cognitive-behavioral therapy is particularly suitable for those with cognitive impairments because it is structured, focuses on concrete behaviors and thoughts, and allows for the use of written aids to bypass cognitive problems.
Our study was designed to examine the effects of a combined social skills treatment for people with severe, chronic TBI and chronic social difficulties. The treatment used a 3-pronged approach to address (1) social behavior, (2) social perception, and (3) emotional adjustment (anxiety, depression). It was hypothesized that treatment would lead to significant improvement in these 3 areas as directly assessed (primary outcome variables) and in reports of social behavior and participation as rated by self and relatives (secondary outcome variables).
An RCT compared social skills training with social activity alone (placebo: control group 1) or waitlist (control group 2). Previous treatment trials
with TBI have produced significant treatment effects and large effects sizes (.89) with as few as 8 participants per group. Formal calculation of sample size based on estimated power of 0.8 and α of .05 suggests that a design comparing 1 treatment to 2 control groups requires 16 per group (total, 48) so this was the number that we aimed to recruit. In all, we recruited 51 participants from the 3 brain injury units in Sydney, Australia, and randomly allocated to (1) a 12-week skills training program, (2) a 12-week nontherapeutic social group (placebo), or (3) a waitlist control (deferred treatment) condition. We used planned contrasts with our primary outcome variables to determine (1) whether social activity alone produced improvements relative to waitlist and (2) whether treatment was more efficacious than control conditions (waitlist or social group).
Neuropsychologic Baseline Data
We assessed participants for basic neuropsychologic abilities prior to study commencement. Cognitive abilities included premorbid ability and those abilities known to be vulnerable to the effects of brain injury and potentially important for responsiveness to treatment. Tests used were: for premorbid ability: Wechsler Test of Adult Reading
subtests—logical memory I and II, verbal paired associates I and II, and faces I and II; for executive function: higher level abstraction, generativity, and for flexibility: WAIS-III similarities, Wisconsin Card Sorting Test,
Participants were filmed during spontaneous social encounters with a male and a female stranger (professional actors) on 3 occasions prior and 3 occasions post-treatment. Within each session the actor engaged in 4 minutes of social chit-chat in a format of alternating between friendly and disinterested.
After this the second actor entered the room and introduced himself/herself. The first actor ended the conversation politely and left the room. The second actor then repeated the procedure.
All videotaped encounters were re-copied onto master tapes in a predetermined random sequence. The encounters were rated by 2 raters who were blind to the goals of the study and participant allocation. Raters were trained to use 2 scales from the BRISS-R.
The PDBS entails 3 subscales: use of reinforcers, self centered behavior, and partner involvement behavior. The PCSS also contains 3 subscales: use of humor, social manners, and self disclosure. Each subscale produces a single rating for the specified behavior scored from 1 (very inappropriate) to 3 to 5 (normative range) to 7 (very appropriate) and provides specific behavioral referents (eg, “Follow up on partner's remarks most of the time”) to assist in anchoring the ratings. Both scales measure the degree to which the conversant adapts to the social context but the PDBS, in particular, focuses on behaviors requiring awareness and sensitivity to “the other” and to the effect of one's behavior on “the other.” Consequently, it is sensitive to the loss of higher level reasoning, decision making, and self-regulation after TBI.
which is comprised of 3 subscales: part 1 assesses the ability to name 7 basic emotions (happiness, surprise, anger, sadness, fear, disgust, neutral) when these are portrayed dynamically by professional actors in videoed vignettes (4 exemplars of each; maximum score, 28). Part 2 assesses the ability to interpret conversational remarks made by speakers (professional actors) during 15 short vignettes of everyday interactions. In the vignettes the speaker is either sincere or sarcastic (ie, inferring the opposite to that asserted). Part 3 comprises 16 similar vignettes but in this case judgments are made as to whether speakers are lying to be kind (in which case they mean their remark to be taken literally) or, again, being sarcastic and inferring the opposite to be true. In both parts 2 and 3 comprehension is assessed via 4 questions about the thoughts, feelings (including subtle emotions such as embarrassment and annoyance), intentions and meaning of the speakers based on their demeanor (facial expression, tone of voice, gesture). The content of the conversation is otherwise ambiguous (part 2 maximum score, 60; part 3 maximum score, 64). TASIT is one of the very few standardized instruments available that measures the ability to read emotional cues in naturalistic contexts. It has good test-retest reliability in people with TBI (.74−.88)
We assessed depression and anxiety via self-report on the DASS (21 items). The DASS is designed to provide relatively pure measures of these constructs without contamination of somatic complaints in a brief format with each item scored from 0 to 3. It has excellent psychometric properties (depression, α=.91; anxiety, α=.84; stress, α=.90) and has convergent validity with other scales discriminating between anxiety and depression.
In this study, scores for the 3 subscales (each with a maximum of 21) were reported.
Secondary Outcome Variables
We used secondary outcome measures to evaluate the extent to which treatment gains were apparent outside the treatment context (generalization). Social behavior was assessed via 3 relative-report scales that have established validity for use in the TBI population. Each assesses related facets of social skills.
the first of the 5 subscales of the KAS, is a questionnaire that focuses on a range of primarily poor social behaviors, for example, “acts childish,” and has 127 items which are each rated from 1 to 4 (maximum score, 508) where a higher score reflects greater impairment. The KAS-R1 has previously been used to evaluate social skills training after TBI
is a questionnaire designed to measure both antisocial (negative) behaviors (eg, “interrupts others”) and prosocial (positive) behaviors (eg, “has eye contact when speaking”) with 50 questions pertaining to both, each scored from 1 to 4. Higher scores on both scales indicate greater social competence. The SPSS has been used with a variety of populations including students,
is designed specifically to address communication problems after TBI (eg, “speaks too slowly?”, “goes over and over the same ground in conversation?”). It has 30 questions each rated 1 to 4, where 4 represents greater impairment (maximum score, 120). The relative form of the LCQ has good internal reliability (α=.85) and stability over time (r=.48) and was used in this study.
Sydney Psychosocial Reintegration Scale
Social participation was assessed using both self- and relative report on the SPRS
with 12 items assessing 3 areas: work and leisure activities, interpersonal relationships, and independent living skills; it ranges from 6 (no change/improvement) to 0 (extreme decline in capabilities) since the brain injury. The 3 subscales have good reliability (test-retest, r range, .77−.93) and validity when used in this population.
In this study, SPRS total scores (maximum, 72) were used.
We recruited 51 outpatients from the 3 brain injury units in Sydney, Australia, who had experienced acquired brain injury. Of these, 47 experienced TBI and 5 experienced brain damage from other causes (eg, anoxia) but presented with similar problems. Inclusion criteria were as follows: (1) a severe acquired brain injury (for participants with TBI this was defined as a length of post-traumatic amnesia exceeding 3 days; for the remaining patients with acquired brain injury, this was defined as a brain injury requiring inpatient rehabilitation), (2) chronic stage of recovery, that is, at least 12 months postinjury and residing in the community, (3) referred to us by a case manager on clinical grounds, that is, being identified as having deficits in social skills, and (4) time available to attend therapy for 12 weeks. Exclusion criteria included (1) severe and/or extensive cognitive impairment that would preclude training based on previous medical records and routine neuropsychologic assessments, (2) insufficient English to be able to follow instructions and converse with group members, (3) significant aphasia, (4) active psychosis, and (5) severe depression.
Recruitment, allocation to groups, and provision of treatment occurred in 3 waves, each occurring at a different brain injury unit. Human subjects research approval was gained from all sites and all participants consented to be involved. Once a group of outpatients who were willing and able to attend social skills training had been recruited and screened, they were assessed on the neuropsychologic tests, primary outcome measures, and the SPRS. A relative or carer was identified who was willing and able to complete the remaining secondary outcome questionnaires. To randomize participants, random numbers were generated by a computer off-site and typed onto inserts in sequentially numbered opaque envelopes corresponding to the sequence of participant recruitment. These were opened after preassessment measures had been completed. Participants who were assigned to skills training were required to attend a hospital outpatient or community facility and participate in a 12-week program that entailed both group and individual sessions totaling 4 hours a week. Subjects assigned to the social group attended 12 weekly sessions of 4 hours focused on group social activities, for example, cooking, craft, and board and ball games which provided companionship but no explicit therapeutic goals. There was no overlap in personnel (all speech pathologists or psychologists) who ran the 2 groups. Those assigned to the waitlist condition were told that they would receive treatment at the end of the study and were otherwise not contacted until the postassessments occurred. At the end of treatment, all participants were reassessed on all outcome measures by personnel who were blind to the goals of the study and to group allocation. At the completion of postassessments those in the waitlist condition were offered treatment.
Social Skills Treatment Program
The social skills treatment program entailed 12 weekly, group sessions of 3 hours involving 3 to 5 participants with 2 therapists. In addition each participant attended a 1 hour weekly individual session with a clinical psychologist.
Two hours per session were devoted to training social behaviors using a manual developed in earlier research.
A different facet of social behavior was addressed each week: greetings, introducing oneself and others, listening, giving compliments, starting a conversation, topic selection, being assertive, and coping with disagreements. Within these broad frameworks, individual goals were identified for each client (eg, to ask more questions, to reduce immature joke-telling, to communicate in full sentences rather than broken phrases, to develop topics).
Further, specific difficulties for each participant were identified, for example, lack of knowledge of appropriate ways to behave; failure to remember, initiate, or regulate behavior; or poor motivation. Interventions were then tailored accordingly. Training each week followed a similar structure: warm-up games, review of homework, introduction of target skill, discussion of potential issues and solutions, therapist modelling of appropriate and inappropriate behavior, and role-playing to develop skills. In all sessions the acronym WSTC (What am I doing? What's the best Strategy? Try it. Check it out) was used to assist planning and monitoring of behavior. High levels of social reinforcement aided confidence and motivation. Immediate feedback and extensive repetition aided learning and memory. The emphasis throughout was on positive feedback with any reference to unsuccessful elements being couched in terms of “what to work on next.” Weekly homework assignments increased the likelihood of generalization of new behaviors to other situations outside the training situation.
One hour per session was devoted to training in the decoding of expressions of emotion in face, voice, and gesture, as well as to understanding social inferences that arose from the demeanor of the speaker (eg, sarcastic, hostile, tentative, friendly). This training also followed a manual developed in previous research.
The overall emphasis was on graduated practice of increasingly complex, guided tasks. Participants were trained to orient to and to recognize specific components of emotional expression including face, voice, posture, and movement.
Training commenced with simple line drawings of basic expressions and progressed to photographs and dynamic presentations (video and role plays) culminating in the interpretation of complex social inferences based on speaker demeanor and situational cues. Discrimination tasks were given repeatedly, as well as rehearsal of different modes of expression through mirror practice, role play of facial, body and verbal cues, and games from the social skills literature.
Material was cumulatively reviewed during and outside sessions via written outlines and explicitly scheduled reviews at home.
Participants also attended weekly individual sessions with a clinical psychologist in order to address personally identified issues in self-esteem, anxiety, and depression as relevant and to reinforce skills learned in the group session. Treatment was not manualized but encompassed standard cognitive-behavioral techniques including relaxation, cognitive strategies, graded exposure, and assertiveness training with modifications to minimize the impact of cognitive impairments
including short sessions and frequent breaks, frequent repetition of information, visual aids, cue cards, session summaries, and audiotapes and simplified cognitive strategies, including self-instructional methods. Summarizing techniques, commonly used to display empathy during sessions, served the additional function of refocussing tangential clients.
Fidelity of Randomization Procedure
Participants were randomly assigned to skills training (n=18), social group (n=17), and waitlist (n=16). After allocation it was necessary to transfer some participants between groups. Five participants assigned to skills training were unavailable on the day scheduled for group sessions and were reallocated to the social group (n=3) or the waitlist (n=2). Two participants allocated to the social group could not attend on the day scheduled for this group and were reassigned to skills training. Two other participants, one from social group and one from waitlist, were randomly selected to transfer to skills training to rebalance numbers. Despite reallocation, subject attrition occurred, with 39 of 51 completing the treatment protocol and post-treatment assessments. In all there were 4 dropouts for each condition. These were all men, which is not unexpected given the predominance of males in the overall sample. Two dropped out due to anger, noncompliance, or alcohol (2 from skills training, one from social group), 2 dropped out because of vocational issues (both from social group), and 7 failed to show up for sessions and/or assessments without explanation (2 in skills training, 1 in social group, 4 in waitlist). Figure 1 summarizes the allocation process.
Mean scores for demographic and clinical variables are depicted in table 1 for both the original groups as randomly allocated and after transfer and dropouts. Neuropsychologic test scores are detailed in table 2 for the 3 groups who proceeded to the end of the study. Between-group differences on all variables were examined via a series of 1-way ANOVA. There were no significant differences between groups on demographic and clinical variables either before or after reallocation. There were also no differences between the groups on neuropsychologic measures. Nor was there any difference between the dropouts and those who remained in treatment on neuropsychologic test measures. Finally, a multivariate ANOVA was conducted comparing the 3 final groups on all neuropsychologic test scores in combination. This again indicated a lack of difference between groups (F30,22=1.00, P=.51). It should be noted that as numbers in the final groups dropped below ideal, statistical difference between groups may reflect limited power rather than genuine equivalence.
Table 1Basic Demographic Data for Participants as Initially Randomized to Groups and in Their Final Groups After Reallocation, Including Dropouts
were used to measure interrater reliability for the 2 raters rating participants' behavior on the BRISS-R scales, PDBS, and the PCSS separately. ICCs were calculated for each of the 3 subscales of the 2 scales for each of the 6 test occasions. According to the guidelines of Cicchetti
these indicated good to excellent interrater reliability for the self centered behavior and partner involvement behavior subscales of the PDBS (self centered behavior subscale, .664−.825; partner involvement behavior subscale, .608−.815), fair to good ratings for its use of reinforcers subscale (.428−.679), and 2 of the PCSS subscales (self disclosure subscale, .419−.619; social manners subscale, .097−.682) but only poor to fair ratings for the remaining subscale (use of humor subscale, .24−.55).
Analysis of Treatment Effects
To compare pre- and post-treatment scores on the BRISS-R scales, participants' scores on each subscale for each session were averaged across raters. Pretreatment was defined as the mean score across the 3 preassessment sessions and post-treatment as the mean of the 3 postassessments.
Treatment effects were examined using repeated-measures analyses. For each primary outcome measure, we were interested to determine whether there was an improvement as a result of treatment, both overall and at the subscale level (eg, on the PDBS, was improvement due to improved use of reinforcers versus less egocentric behavior?). Consequently, scores on each primary variable: BRISS-R (PDBS, PCSS), TASIT, and DASS were analyzed via ANOVAs using PSY
with orthogonal planned contrasts controlling the α error rate at .05. These examined interactions involving (1) the 2 control conditions (to determine if social activity alone, ie, placebo, improved performance) and (2) skills training compared with the 2 control conditions (social group, waitlist) (to determine if treatment improved performance). Both placebo and treatment effects were defined as significant group by time (pre vs post) interactions for overall scores on each primary outcome variable. Significant interactions were followed up with post hoc examination of confidence intervals surrounding standardized mean differences for interactions involving individual subscales. Complete data were available for TASIT (ie, n=13 for each group) but not all participants attended all BRISS-R assessments so data analysis of these scales was restricted to those with a full data set (10 in skills training and 11 in each of the social group and waitlist conditions). Similarly, not all self-report questionnaires were returned, so DASS (and SPRS-self) data were limited to 13 in skills training, 11 in social group, and 12 in waitlist.
Secondary outcome variables
KAZ-R, SPSS-P, SPSS-N, LCQ, and SPRS tapped overlapping facets of social behavior so a repeated-measures multivariate ANOVA was used to examine treatment effects (defined for this analysis as a significant group: [skills training vs social group vs waitlist] × time [pre vs post] interaction). Not all post-test relative questionnaires were completed. Analyses were restricted to 9 in skills training, 10 in social group, and 8 in waitlist conditions. A separate ANOVA using PSY and planned contrasts (as above) was used to assess SPRS self-report.
Scores for subscales of the PDBS and PCSS are depicted in Fig 2, Fig 3, respectively, with more specific details in table 3. Mean scores for the 3 groups at pre- and post-test on all primary and secondary outcome variables are detailed in table 3. Inferential statistics for the planned contrasts are summarized in the right hand columns of table 3.
Table 3Scores at Pre- and Post-Treatment on Primary Outcome Variables for the 3 Groups: Skills Training, Social Group, and Waitlist
There were no interaction effects for social group relative to waitlist on any primary outcome variable; that is, social activity alone did not lead to differential improvement. On the other hand, the skills training group improved significantly in social behavior as measured by the PDBS relative to the other 2 groups. Examination of confidence intervals revealed that this was true for both the self centered behavior and partner involvement behavior subscales but no real change was seen on the use of reinforcers subscale. There was no evidence of treatment effects on the PCCS. Nor was there any improvement as a result of treatment on social perception (TASIT). The groups were, on average, in the mild-to-moderate range for anxiety, depression, and stress on DASS scales at pretest. This changed little at post-test and there was no overall treatment effect. Nor were there treatment effects for any of the secondary variables overall. In order to increase power, analyses were rerun comparing the skills training group with the control participants as 1 group (i.e. the social group and waitlist group combined). A similar pattern emerged, that is, an overall treatment effect for the PDBS (F1,30=9.87, P=.004; standardized mean differences, .71) alone. Finally, in order to adhere to intent-to-treat methodology, all analyses were re-conducted comparing participants in the groups to which they were originally allocated. Unsurprisingly, no treatment effects emerged.
Overall, this study suggested that social skills treatment for people with chronic, severe brain injuries produces circumscribed improvements. Replicating Helffenstein and Wechsler
the results documented improvements in a direct measure of social behavior (PDBS). As with Dahlberg, no change was found post-treatment in secondary outcome measures of social functioning and social participation (although anecdotal reports from relatives and participants suggested improved participation and confidence for some at least).
Improvement in social behavior appeared to be specific to partner directed behavior as measured by the PDBS. This scale has been shown to be sensitive to TBI
The scale focuses on the ability to adapt to the social requirements of others. The self centered behavior and partner involvement behavior subscales, in particular, showed improvement, indicating that participants with social skills training were less inclined to talk about themselves and more inclined to encourage their conversational partner to contribute to the conversation. Frontal deficits, prevalent after TBI, are thought to underpin egocentric, insensitive behavior, and a failure to adapt to changing circumstances. These have clear ramifications for poor partner-directed behavior, particularly in situations where the speaker must meet and appraise a stranger, as occurred in our in vivo ratings. Consequently, it was encouraging that poor social skills that clearly relate to poor executive function did appear amenable to treatment. Nonetheless it is important to note that the PDBS had greater reliability compared with the PCSS, a finding also reported elsewhere.
This may have made it a more reliable measure of within-subject change. Unfortunately, all scales of the BRISS-R require intensive rater training and even then, interrater reliability across scales is variable. Given that direct assessment of behavioral skills appeared to be the most effective means of measuring change, further research in this field may benefit by tying outcome measures to even more specific measures of social skills.
The lack of significant improvement in the other 2 primary outcome variables, social perception and mood, is not entirely unexpected. Although emotion perception, depression and anxiety in people with TBI have been shown to be amenable to remediation in other randomized controlled studies
those trials selected participants who showed clear difficulties in these areas. In contrast, participants in the present trial were selected specifically based on clinical judgments regarding behavior. It was anticipated that many would have deficits in social perception because these problems are prevalent after TBI and they interfere with social skills. But participants were variable in the kinds of deficits displayed and not all had significant difficulties in social perception. Consequently not all would benefit from specific remediation. Similarly, the absence of discernible improvement in mood as indicated on the DASS possibly reflects the fact that the group was not selected primarily for the presence of depression, stress, and anxiety. As the pretest scores attest, on average, the groups were generally in the mildly affected to moderate range. The individual sessions were included in the treatment package in order to address issues that may provide an obstacle to deriving benefit from the group program, but it was not always the case that such issues were present. Future research should include other measures more directly related to the impact of better social skills, such as self-esteem, confidence, and self-concept.
The questionnaire-based measures of functioning yielded no indication of benefit as a result of intervention. In hindsight, this is also not surprising. Regardless of any real benefits of social skills training, unrelated life events, for example, illness, legal battles, and weddings that occurred over the 6-month period of participation, influenced actual and perceived psychologic well-being, social behavior, and social participation. For this reason, direct measures of social skills such as those encompassed in the primary outcome variables, are probably the most suitable for assessing treatment effects.
In this study, we treated people who were chronically injured. This restriction circumvented confounds associated with spontaneous recovery and ensured that participants had opportunities to experience social life (or lack thereof) postinjury, thereby providing some level of insight and motivation. It also meant participants were able to practice skills in their own social contexts, increasing the likelihood of generalization and transfer. Further, provision of social skills training post rehabilitation avoided the need to coordinate with other rehabilitation commitments that might otherwise compete for time and potentially confound any treatment effects. However, the choice of people with chronic injuries yielded its own limitations. The requirement that participants must be available on a weekly basis for an extended period was difficult for many potential participants to meet. Indeed, this requirement not only limited numbers but also skewed the sample toward those who were chronically underoccupied and, by implication, less able. Most of those who had full- or part-time jobs or avocational pursuits found the treatment schedule untenable and therefore could not commit. Yet another consequence was that the participants included in this study were unlikely to have a range of social opportunities in which to practice (and demonstrate) improved social skills. One way in which this latter problem could be overcome would be to enhance social skills training by enlisting community members to provide structured opportunities for leisure and social activities. Such programs have proven effective for increasing participation for both adults and children with TBI.
There were several limitations associated with this study. Despite the fact that it was carried out in a large city, with the full cooperation of all metropolitan brain injury units, there was an inherent limit to the number of participants suitable and available for an extensive treatment program. Although we initially recruited sufficient numbers according to power requirements, dropouts led to attrition rates that dropped numbers below optimum. Furthermore, effect sizes were smaller than anticipated. Even the primary outcome variable PDBS in which treatment effects were clearly evident yielded only a medium effect size of 0.7. Consequently, the number of participants was small, relative to the effects achieved and the heterogeneity they displayed in basic neuropsychologic profiles (see table 2). This limited power to see real change as a result of treatment across groups. Logistical constraints made it impossible to retain the initial randomization sequence which further reduced the validity of the findings. The small numbers meant that analysis based on reassignment of participants to their initial groups, according to conservative intention-to-treat methodology, guaranteed that any treatment effects would be washed out. Other statistical methods for dealing with dropouts were not appropriate given the small numbers. Future work needs to plan ways to overcome these problems, for example, as with Dahlberg et al,
the use of only 1 control group and the inclusion of deferred treatment (waitlist) post assessment data is a means to increase numbers and improve power.
This trial, evaluating a 3-pronged approach to social skills training, produced modest evidence for improvements in social behavior as measured directly. Violation of the initial randomization sequence and small numbers limit its validity as a true RCT. Nonetheless, despite the small numbers and the severe, chronic nature of disability experienced by the participants, improvements in social behavior were apparent especially in a reduction in self-centered behavior and greater effort to involve the conversational partner. The impact of this change on the lives and social relationships of these participants was not, however, seen in relative or self-report of social function or social participation. Partly, this was a reflection of the problems inherent in the use of relative and self-report scales combined with lack of power. Partly also, it may reflect the social realities of people with long-term chronic disability, whose opportunities to practice and benefit from improved social skills are curtailed.
Despite targeting social perception and mood, no improvements were seen in these areas as a result of this program, possibly because these represented difficulties for only a proportion of research participants. If so, these aspects of the training program may be better used as modules to target people for whom these issues are major obstacles to social functioning. It could be argued that this trial evaluated treatment effectiveness at the most difficult end of the spectrum. The people we recruited had very severe injuries. And even within this population, our sample consisted of those who were chronically underoccupied, suggesting they represented those with greater disability. Arguably, there may be greater scope to see treatment effects in people with less severe injuries and who are in the acute or early chronic phase of recovery. Certainly the work of Helffenstein and Wechsler
suggests so. Psychosocial deficits after brain injury are clearly a major impediment to living a socially fulfilling life, to finding and keeping work, friends, and families. We hope this study marks a small step in a sustained research effort to establish good evidence-based techniques for maximizing social skills remediation.
We thank the staff of the Liverpool Brain Injury Unit, Royal Rehabilitation Centre Sydney, and the Westmead Brain Injury Unit who assisted with recruitment and logistical issues and enabled this multicenter trial. We also thank the many postgraduate psychologists who worked on this project including Jo Speyer, MPsychol (Clin), Amanda Lane-Brown, MA, Regina Shultz, MClinNeuroPsy, Maureen Kong, MPsychol (Clin), Maria Haros, MPsychol (Clin), Fiona McGregor, MPsychol (Clin), Chris Hunt, MPsychol (Clin), and Kay Pegg, MPsychol (Clin).
Communication disorders following traumatic brain injury.
in: McDonald S. Togher L. Code C. Communication disorders after traumatic brain injury. Psychology Pr,
Supported by the Australian National Health and Medical Research Council (project grant no. 222754).
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. McDonald is an author of The Awareness of Social Inference Test, which is used as an outcome measure in this study, and receives royalties for its sale.