Effectiveness of Rehabilitation in Enhancing Community Integration After Acute Traumatic Brain Injury: A Systematic Review

Previous Systematic Reviews of CI After Acquired Brain Injury

In recent years, three systematic reviews assessing CI after acquired brain injury (ABI), which includes TBI as well as brain injury from nontraumatic causes, have been published: Reistetter and Abreu (2005), Carlson et al. (2006) and McCabe et al. (2007). Gordon et al. (2006) also published an extensive literature review on TBI rehabilitation, one small part of which covered CI outcomes. Reistetter and Abreu’s study focused on finding the best CI measures for the ABI population, the predictors for success, and CI’s relationship to performance level. Most of the studies they reviewed were retrospective cohort, cross-sectional, and reliability studies. Their study found strong evidence supporting the Community Integration Questionnaire’s (CIQ’s) utility for CI outcome measurement and revealed major predictors of CI, such as injury severity, age, gender, education, prior work, living environment, cognition, emotional status, functional performance, and disability. However, Reistetter and Abreu did not review intervention studies in relation to CI.

Carlson et al.’s (2006) systematic review aimed to identify best practices and promising new ways to enhance participation in the ABI population. Yet the scope of their review was likely narrowed by their exclusion of both non-English publications and articles on the geriatric population. They classified CI programs as “learning to participate” or “participating to learn” and supported the latter as a useful rehabilitation model.

McCabe et al.’s (2007) systematic review focused on CI using the Downs and Black (1998) checklist and the Physiotherapy Evidence Database (PEDro) scale to assess evidence levels. McCabe et al. concluded that because of a lack of studies with strong methodological quality, only limited evidence existed to support rehabilitation interventions.

To summarize current TBI rehabilitation studies, Gordon et al. (2006) undertook a literature review that focused specifically on TBI and included a CI component. This work, however, was not based on a comprehensive literature search and did not assess each article individually for evidence levels.

All seven studies that Gordon et al. (2006) examined in their review found that the interventions studied improved employment, health-related quality of life, physical well-being, successful living in home communities, self-efficacy, and the ability to cope with depression. The range of outcome variables was great. Their review did not include a comprehensive synthesis of the findings, and not a single randomized controlled trial (RCT) was reviewed.

The three systematic reviews discussed earlier (Carlson et al., 2006; McCabe et al., 2007; Reistetter & Abreu, 2005) were restricted to articles published in English and to ABI populations, including those with stroke and brain tumors. One could argue that because of TBI’s many clinical characteristics, such as the nature of the injury and its neurological symptoms, as well as the different acute and post–acute TBI treatment approaches, TBIs should be researched apart from all other ABIs.

Our systematic review, which synthesizes evidence from previous work, focuses on the CI outcomes of TBI survivors and includes a search for non-English articles. Also, in contrast to the earlier reviews, we expanded the inclusion–exclusion criteria to include all adult populations and made use of specific CI outcome measurements from intervention studies.

Measurements of CI

Although some consensus exists as to what constitutes CI, it is still difficult to assess and accurately measure (Gordon et al., 2006). CI measures have included the Craig Handicap Assessment and Reporting Technique (CHART; Whiteneck, Charlifue, Gerhart, Overholster, & Richardson, 1992), CIM (McColl et al., 2001), CIQ (Willer et al., 1993), RNL (Wood-Dauphinee, Opzoomer, Williams, Marchand, & Spizer, 1988), SPRS (Tate, Hodgkinson, Veerbangsa, & Maggiotto, 1999), Brain Injury Community Rehabilitation Outcome–39 (BICRO–39; Powell, Beckers, & Greenwood, 1998), and AIMS Interview (Berry, 1994), all of which informed our study’s inclusion criteria for outcome measures. These instruments were developed to assess core elements of CI: relationships with others, independence in living situation, and meaningful activities (Salter, Foley, Jutai, Bayley, & Teasell, 2008). The types of subscales that can be found in these assessments include physical and cognitive independence, mobility, occupation, social integration, economic self-sufficiency, home and social integration, productivity, daily functioning, perception of self, occupational activities, interpersonal relationships, and independent living skills, which are relevant outcomes for occupational therapists.

Study Objectives

The study’s objectives were to assess evidence for post–acute TBI rehabilitation intervention programs for adults that include community integration as an outcome measure, using a systematic literature review, and to describe these program’s characteristics. We also examined occupational therapy’s involvement in these programs and their relevance to occupational therapy.

Method

To formulate research questions and develop search strategies, we adopted an approach from a guidebook for systematic reviews (Khan, Kunz, Kleijnen, & Antes, 2003). For appraisal and interpretation of the methodological quality of the studies, we adopted an approach used in reports from the Evidence-Based Review of Moderate to Severe Acquired Brain Injury Group (Teasell et al., 2005, 2007).

Results

Two reviewers retrieved a total of 559 articles from the six major databases (Medline, CINAHL, AMED, EMBASE, PsycINFO, and the Cochrane Library). Seven more studies were obtained through manual searches of the five journals and back-referencing from previous reviews. Of the 566 articles, 537 did not meet the inclusion criteria on the basis of the titles and abstracts. The excluded articles were case studies, measurement studies, economic analyses or articles that did not use CI measurements for the outcomes studied. Of the 29 remaining studies, 18 were further excluded for similar reasons after the full text was thoroughly examined. Of these, 2 were considered as one complete study (Paniak, Toller-Lobe, Durand, & Nagy, 1998; Paniak, Toller-Lobe, Reynolds, Melink, & Nagy, 2000) because the original study had included a 1-yr follow-up published in a separate article. Therefore, the literature search yielded a total of 10 articles for inclusion in this systematic review. We did not find any non-English language papers.

Discussion

By attempting to systematically assess intervention studies in TBI rehabilitation research and focusing specifically on CI, we took a unique approach in our research review. Our review builds on previous research by providing TBI-specific information and highlighting the contributions of recent studies. Given that the studies included in this review were conducted in various settings and included participants with a wide range of injury severity and different demographic and clinical characteristics, we advise careful interpretation of the findings. Yet, despite the methodological challenges and the variability of the clinical population, we nevertheless believe that our study highlights promising evidence, particularly regarding multidisciplinary rehabilitation and cognitive and vocational rehabilitation in moderately to severely injured adults.

Multidisciplinary rehabilitation interventions designed by Powell et al. (2002), Goranson et al. (2003), Willer et al. (1999), and Hashimoto et al. (2006) showed positive results in enhancing CI. Several of the other designs were also beneficial for CI measurements. Powell et al.’s outreach treatment program in a community setting more effectively increased CI levels than did the less structured and less intensively treated control group. When compared with a home-based program, Goranson et al.’s multidisciplinary rehabilitation program was also more successful in improving CI. Compared with a nonrehabilitation control group, the treatment group that received multidisciplinary home-based services was more likely to maintain CI scores at follow-up. Hashimoto et al.’s study also demonstrated the effectiveness of a day-treatment program using a multidisciplinary team approach. The social integration scale score and productive activity scale score were significantly greater for participants.

With respect to cognitive rehabilitation interventions, Cicerone et al.’s (2004) intensive cognitive rehabilitation program was more effective in improving CI than were standard neurorehabilitation programs. Constantinidou et al.’s (2005) categorization training, a unique approach in cognitive rehabilitation that can be conducted by occupational therapists, produced a significant improvement in the level of CI. A comprehensive integrated program using many occupational therapy approaches improved vocational outcomes and social participation (High et al., 2006).

Not all the studies, however, yielded positive results. Paniak et al.’s (1998, 2000) intervention, which had more intensive assessments, education, and treatment as needed, failed to show any improvement in CI when compared with a control group. Tiersky et al.’s (2005) cognitive rehabilitation program, which used individual cognitive–behavioral psychotherapy and cognitive remediation, did not affect CI scores. A scheduled telephone intervention designed by Bell et al. (2005) also did not have significant positive CI effects.

Concerning injury severity, studies on the mild TBI population (Paniak et al., 1998; Tiersky et al., 2005) showed no CI improvements. These 2 studies also showed little change in mean CIQ scores (−0.52 and 0.29, respectively). In comparison, the other 8 studies had mean change scores ranging from 1.81 to 5.2. Our review found that patients with mild TBI were not likely to benefit significantly from post–acute TBI CI rehabilitation. As mentioned in Paniak et al.’s study, many participants had already returned to the workplace after being discharged from acute care hospitals. One could assume that the mild TBI patients already had higher CIQ scores and that it was not the most relevant outcome for this population. Most studies on moderate and severe TBI, however, found greater scores in treatment groups.

Although these studies showed little to no improvement in CI scores, positive changes were observed in other areas, such as functional status and cognitive or psychological outcomes (Bell et al., 2005; Paniak et al., 1998; Tiersky et al., 2005). Statements made about the effectiveness of CI without reference to other outcomes, therefore, require careful interpretation. In addition, the follow-up period studied may not capture CI as the longer term goal of some interventions.

Because the intervention studies differed significantly on treatment and outcomes, a quantitative analysis (i.e., a meta-analysis) was not possible. The lack of consistent treatment interventions and treatment goals and the inability to conduct a meta-analysis reflects the broad scope and nonspecific nature of TBI rehabilitation research. As a result, we are limited to general statements about the effectiveness of any intervention.

Within the scope of our review, occupational therapists participated in six CI programs for TBI as rehabilitation team members (Cicerone et al., 2004; Goranson et al., 2003; Hashimoto et al., 2006; High et al., 2006; Powell et al., 2002; Willer et al., 1999). All of these programs showed significant CI improvements. This finding indicates that rehabilitation teams that include occupational therapists are critical in promoting CI; however, it is difficult to assess the occupational therapists’ precise influence within rehabilitation teams.

Limitations of This Review

Our systematic review includes relatively high-quality studies. Not many RCTs are reviewed in CI studies, so to pool the data we broadened our inclusion criteria to cover high-quality observational studies. Many systematic review articles, even in the Cochrane Library, allow non-RCT studies when synthesizing findings. Checklists and scales have been developed to assess RCT and non-RCT evidence levels. We should also note that the systematic literature approach in relation to CI has been criticized. Some experts feel that an RCT may not be the optimal design by which to study successful CI. Greater weighting of high-quality controlled studies, for instance, would have resulted in stronger statements about CI effectiveness. Although we acknowledge this perspective, the RCTs in our review have produced interesting and useful findings.

Gray literature, such as dissertations and non–peer-reviewed papers, are not mandated sources of information and thus are often not included in systematic reviews. We acknowledge, however, that they might be explored in future reviews, and we are aware of publication bias in the inclusion of only peer-reviewed published studies.

We also recognize the potential overlap of interventions described in the studies. Many multidisciplinary interventions use combinations of cognitive rehabilitation programs. The rigid categorizations used in our review may be limiting, but they were created to help readers better understand the interventions’ characteristics. Otherwise, these categories do not need to be organized as strictly as we have done.

Some of the studies, mostly RCTs, did not focus on or aim to improve CI and may have treated it only as a global outcome measure, similar to the Disability Rating or the Glasgow Outcome Scales. Concerns also exist regarding the CIQ’s quality. Some researchers have argued that because of the high probability of the ceiling effect, it is best not to use it in rehabilitation studies.

The studies included people with mild TBI. Research has shown that patients with mild TBI have diverse outcomes and require outcome measures different from those used with more severe injuries. Nonrobust findings also could be the result of the variability among TBI patients; depending on the brain area affected, great differences can be found in patients’ functional deficits. These differences may have reduced the beneficial CI effects observed after intervention for patients with moderate to severe brain injury (Barrett, Levy, & Gonzalez Rothi, 2007).

Conclusions

Systematic reviews allow health care professionals, clients, researchers, and policymakers to make decisions on the basis of all the available evidence, which has also encouraged advances in rehabilitation (Law & Philip, 2002). To our knowledge, our study is the first of its kind that uses a rigorous, detailed approach to focus exclusively on intervention studies assessing CI outcomes in a post–acute TBI adult population. The promising effects of programs designed by Powell et al. (2002), Goranson et al. (2003), Willer et al. (1999), Hashimoto et al. (2006), Cicerone et al. (2004), Constantinidou et al. (2005), and High et al. (2006) could be considered in future decision making and discharge planning in post–acute TBI services for adult TBI clients to enhance community integration.