THE BRAIN INJURY interdisciplinary Special Interest

Transcription

1 REVIEW ARTICLE Evidence-Based Cognitive Rehabilitation: Updated Review of the Literature From 1998 Through 2002 Keith D. Cicerone, PhD, Cynthia Dahlberg, MA, CCC-SLP, James F. Malec, PhD, Donna M. Langenbahn, PhD, Thomas Felicetti, PhD, Sally Kneipp, PhD, Wendy Ellmo, MS, CCC-SLP, Kathleen Kalmar, PhD, Joseph T. Giacino, PhD, J. Preston Harley, PhD, Linda Laatsch, PhD, Philip A. Morse, PhD, Jeanne Catanese, MA, CCC-SLP 1681 ABSTRACT. Cicerone KD, Dahlberg C, Malec JF, Langenbahn DM, Felicetti T, Kneipp S, Ellmo W, Kalmar K, Giacino JT, Harley JP, Laatsch L, Morse PA, Catanese J. Evidencebased cognitive rehabilitation: updated review of the literature from 1998 through Arch Phys Med Rehabil 2005;86: Objective: To update the previous evidence-based recommendations of the Brain Injury Interdisciplinary Special Interest Group of the American Congress of Rehabilitation Medicine for cognitive rehabilitation of people with traumatic brain injury (TBI) and stroke, based on a systematic review of the literature from 1998 through Data Sources: PubMed and Infotrieve literature searches were conducted using the terms attention, awareness, cognition, communication, executive, language, memory, perception, problem solving, and reasoning combined with each of the terms rehabilitation, remediation, and training. Reference lists from identified articles were reviewed and a bibliography listing 312 articles was compiled. Study Selection: One hundred eighteen articles were initially selected for inclusion. Thirty-one studies were excluded after detailed review. Excluded articles included 14 studies without data, 6 duplicate publications or follow-up studies, 5 nontreatment studies, 4 reviews, and 2 case studies involving diagnoses other than TBI or stroke. Data Extraction: Articles were assigned to 1 of 7 categories reflecting the primary area of intervention: attention; visual perception; apraxia; language and communication; memory; executive functioning, problem solving and awareness; and comprehensive-holistic cognitive rehabilitation. Articles were abstracted and levels of evidence determined using specific criteria. Data Synthesis: Of the 87 studies evaluated, 17 were rated as class I, 8 as class II, and 62 as class III. Evidence within each area of intervention was synthesized and recommendations for practice standards, practice guidelines, and practice options were made. From JFK-Johnson Rehabilitation Institute, Edison, NJ (Cicerone, Kalmar, Ellmo, Giacino, Catanese); Craig Hospital, Englewood, CO (Dahlberg); Rusk Institute of Rehabilitation Medicine, New York, NY (Langenbahn); Mayo Medical Center and Medical School, Rochester, MN (Malec); Beechwood Rehabilitation Services, Langhorne, PA (Felicetti); Marianjoy RehabLink, Wheaton, IL (Harley); Community Skills Program, Counseling and Rehabilitation Inc, Marlton, NJ (Kneipp); University of Illinois, Chicago, IL (Laatsch); and Neurobehavioral Services of New England, Portland, ME (Morse). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to Keith D. Cicerone, JFK-Johnson Rehabilitation Institute, 2048 Oak Tree Rd, Edison, NJ 08820, kcicerone@solarishs.org /05/ $30.00/0 doi: /j.apmr Conclusions: There is substantial evidence to support cognitive-linguistic therapies for people with language deficits after left hemisphere stroke. New evidence supports training for apraxia after left hemisphere stroke. The evidence supports visuospatial rehabilitation for deficits associated with visual neglect after right hemisphere stroke. There is substantial evidence to support cognitive rehabilitation for people with TBI, including strategy training for mild memory impairment, strategy training for postacute attention deficits, and interventions for functional communication deficits. The overall analysis of 47 treatment comparisons, based on class I studies included in the current and previous review, reveals a differential benefit in favor of cognitive rehabilitation in 37 of 47 (78.7%) comparisons, with no comparison demonstrating a benefit in favor of the alternative treatment condition. Future research should move beyond the simple question of whether cognitive rehabilitation is effective, and examine the therapy factors and patient characteristics that optimize the clinical outcomes of cognitive rehabilitation. Key Words: Brain injuries; Cognition disorders; Practice guidelines; Rehabilitation; Stroke by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation THE BRAIN INJURY interdisciplinary Special Interest Group (BI-ISIG) of the American Congress of Rehabilitation Medicine (ACRM) previously conducted an evidencebased review of the literature through 1997 about cognitive rehabilitation for people with traumatic brain injury (TBI) or stroke. 1 That review led the BI-ISIG to make several specific recommendations concerning the clinical practice of cognitive rehabilitation and its effectiveness in TBI and stroke patients. Since publication of our initial findings, several additional systematic reviews of cognitive rehabilitation have been published. A task force under the auspices of the European Federation of Neurological Societies 2 used similar methods and reviewed many of the same studies referenced in our initial report. That task force concluded that there is substantial evidence to support attention training in the postacute phase after TBI (but not during the period of acute recovery) and compensatory memory training for subjects with mild memory impairments. Evidence of the effectiveness of pragmatic conversational therapy after TBI was based on a limited number of studies with small samples, and was in need of confirmation. Several methods of rehabilitation for spatial neglect were found to be effective, as was the treatment of apraxia with compensatory strategies. As part of a broader effort by the Academy of Neurologic Communication Disorders and Sciences to develop practice guidelines for treating cognitive-communication disorders after TBI, Sohlberg et al 3 examined the evidence for the effectiveness of direct attention training after TBI. They con-

2 1682 UPDATE ON COGNITIVE REHABILITATION, Cicerone cluded that there was evidence of improvement in attentionbased skills with direct training, although the interpretation of studies was limited by factors such as subject heterogeneity and the lack of replications. Two Cochrane reviews have examined cognitive rehabilitation for attention deficits 4 or memory deficits 5 after stroke, but only class I studies were included. Lincoln et al 4 concluded there was some evidence that training improves alertness and sustained attention but insufficient evidence of improved functional independence after stroke. Majid et al 5 identified a single study that met their criteria for inclusion and found insufficient evidence to support or refute the effectiveness of cognitive rehabilitation for memory problems after stroke. Several systematic reviews 6-8 found evidence that cognitive rehabilitation, including visual scanning training, improves spatial neglect after right hemisphere stroke, but also found that there is limited or insufficient evidence for the duration of treatment effects or relevance to everyday functioning. 6,8 Systematic reviews of treatment for aphasia have reached conflicting conclusions. Robey 9 conducted a meta-analytic review of 55 studies of clinical outcomes after aphasia rehabilitation. These were generally observational studies, rather than randomized controlled trials (RCTs). Outcomes for treated subjects were superior to outcomes for untreated subjects in all stages of recovery, particularly when treatment was begun in the acute stage of recovery, and the extent of treatment effects was positively associated with the amount of treatment. There were too few studies to permit an examination of the differential effects of treatments for different types of aphasia. Cappa et al 2 found some evidence for the effectiveness of aphasia therapy, again based largely on class II and III studies. In contrast, a Cochrane review 10 of aphasia rehabilitation identified only 12 RCTs suitable for the authors review, none of which were considered of adequate methodologic quality to warrant complete description and analysis. The main conclusion reached in that review was that aphasia therapy after stroke has not been shown to be clearly effective or ineffective within an RCT. These systematic reviews have generally found some evidence to support the effectiveness of cognitive rehabilitation after TBI or stroke, but have also recognized the need for better specification of treatment effects and increased methodologic rigor, and have sometimes limited their conclusions on these grounds. We recognize that clinical guideline development is an ongoing process that should include an updated review of the literature within 5 years of the initial recommendations. The BI-ISIG recently completed an updated evidence-based review and made recommendations that incorporate the literature published from 1998 through This article summarizes those findings and recommendations. In this article, we concentrate on describing class I studies. Class II and III studies are discussed when they provide unique or contradictory information. A more complete discussion of these studies is in the full report of the BI-ISIG committee (online at METHODS We followed prior methodology for identifying relevant literature, reviewing, and classifying, and developing recommendations. These methods are described in more detail in our initial publication. 1 For the current review, we searched PubMed and Infotrieve for articles published between 1998 and 2002, using the terms attention, awareness, cognitive, communication, executive, language, memory, perception, problem solving, and reasoning combined with each of the terms rehabilitation, remediation, and training. We elected to include RCTs published before 1998 that were not included in our first review, but not other studies. Two such articles were identified and included in this review. This resulted in the identification of 312 articles. The abstracts or complete articles were reviewed to eliminate articles according to the following exclusion criteria: (1) articles not addressing intervention, (2) theoretical articles or descriptions of treatment approaches, (3) review articles, (4) articles without adequate specification of interventions, (5) articles that did not include participants primarily with a diagnosis of TBI or stroke, (6) studies of pediatric subjects, (7) single-case reports without empirical data, (8) non peer-reviewed articles and book chapters, (9) articles describing pharmacologic interventions, and (10) articles not written in English. One hundred eighteen articles were included this screening process. Thirty-one studies were excluded after further detailed review (14 studies without data, 6 articles representing duplicate publications or follow-up studies, 5 articles that were nontreatment studies or experimental manipulations, 4 reviews, and 2 single-case studies of subjects with diagnoses other than TBI or stroke). Studies providing follow-up to previously reviewed treatment studies were not fully evaluated or classified as new treatment studies, but findings from these studies were reviewed and are discussed here when appropriate. Eighty-seven studies were fully reviewed and evaluated, with the level of evidence determined based on criteria used in our prior review. 1 Well-designed, prospective RCTs were considered class I evidence; studies using a prospective design with quasirandomized assignment to treatment conditions were designated as class Ia studies. Given the inherent difficulty in blinding rehabilitation interventions, we did not consider this as criterion for class I or Ia studies. Class II studies consisted of prospective, nonrandomized cohort studies; retrospective, nonrandomized case-control studies; or multiplebaseline studies that permitted a direct comparison of treatment conditions. Clinical series without concurrent controls, or single-subject designs with adequate quantification and analysis were considered class III evidence. Articles were reviewed and classified by at least 2 committee members. Disagreements between the 2 primary reviewers (as happened with 7 articles) were first addressed by discussion between reviewers to correct minor sources of disagreement, ant then by obtaining a third review. Of the 87 studies, 17 were rated as class I, 8 as class II, and 62 as class III. The overall evidence within each predefined area of intervention was synthesized and recommendations were derived from the relative strengths of the evidence. The level of evidence required to determine practice standards, practice guidelines, or practice options was based on the decision rules applied in our initial review (appendix 1). All recommendations were reviewed by the entire committee to ensure consensus. Remediation of Attention Deficits We identified 5 studies on remediation of attention deficits after TBI. Two were class I prospective randomized studies 11,12 comparing attention treatment with alternative treatments; 1 was a class II study 13 that compared attention treatment with no treatment; and 2 were class III studies. 14,15 Sohlberg et al 11 used a crossover design to compare the effectiveness of attention process training (APT) brain injury education and support for 14 patients with acquired brain injury. Self-reported changes in attention and memory functioning, as well as improvement on neuropsychologic measures of attention-executive functioning, were greater after APT than after therapeutic support. The second class I study 12 taught 22

3 UPDATE ON COGNITIVE REHABILITATION, Cicerone 1683 patients with severe TBI to compensate for slowed information processing and the experience of information overload in daily tasks. Participants were randomly assigned to receive either time pressure management (TPM) or an alternative treatment of generic concentration training. Participants receiving TPM showed significantly greater use of self-management strategies and greater improvement of attention and memory functioning than did participants who received the alternative treatment. Although the precise nature of the interventions in these 2 class I studies differ, they share a common emphasis on the development of strategies to compensate for residual cognitive deficits ( strategy training ) rather than attempting to directly restore the underlying impaired function ( restitution training ). The results of these 2 studies and of an additional small class II study 13 are therefore consistent with a strategy training model for attention deficits after TBI, and reinforce the conclusions made in our initial review. 1 Clinical recommendations. In the current review, there is evidence from 2 class I studies 11,12 with 36 subjects that supports the effectiveness of attention training for subjects with TBI during the postacute period of rehabilitation. Considering this evidence, along with our previous recommendation based on 2 class I studies with 57 subjects, 1 the committee recommends that strategy training for attention deficits exhibited by subjects with TBI be considered as a practice standard during the postacute period of rehabilitation. Results of studies in this area suggest greater benefits on complex tasks requiring the regulation of attention, rather than on basic aspects of attention (eg, reaction time or vigilance). These results are consistent with the emphasis on strategy training to compensate for attention deficits in functional situations. There is insufficient evidence to support the use of specific interventions for attention deficits during acute rehabilitation. Recommendations for future research. Evidence from 1 class I study 11 suggests that cognitive remediation has differential effects on various components of attention; therefore, more research is needed to clarify the differential effects of interventions. The finding that cognitive interventions influence the regulation of attention can be contrasted with pharmacologic treatment, which may exert its primary effect on basic processing speed. 16 This suggests that a combination of cognitive and pharmacologic interventions may produce the greatest overall improvement in attention deficits after TBI, although this has not been evaluated through controlled research. Remediation of Visuospatial Deficits We reviewed 11 studies in the area of rehabilitation of visuoperceptual deficits. The majority of these involved rehabilitation of unilateral visual neglect, and 1 involved the restitution of visual fields. The 13 studies included 3 class I 17,18 or Ia 19 studies, 3 class II studies, and 5 class III studies Two class I or Ia studies involved the rehabilitation of visual neglect after stroke. The class Ia study 19 was a prospective controlled trial that replicated a prior class II study. 28 Training for neglect produced improvement on standard neglect tests and a functional measure evaluating generalization, compared with general cognitive stimulation. Niemeier 18 found systematic cueing of visual scanning to be superior to conventional rehabilitation (which did not produce improvement). Kasten et al 17 examined computer-based treatment of partial blindness resulting from optic nerve damage or postchiasmal lesions. The training was intended to restore the underlying neurologic and visual functions and reduce the extent of the damaged visual fields (ie, restitution training ). The training resulted in significant enlargement of the visual field in 95% of the subjects, which remained stable 2 years later. 29 One class II 22 and 1 class III 25 study replicated class I demonstrations that visual scanning is an effective treatment for visual neglect. Additional class II 20 and class III 23,24,26,27 studies introduced forced activation of the affected limb in conjunction with visual scanning training as a treatment for visual, and possibly sensory, neglect. Clinical recommendations. Two class I 18 or Ia 19 studies of 41 subjects with cerebral stroke replicated the effectiveness of visual scanning in treating unilateral neglect, consistent with the 6 class I studies involving 286 subjects from our prior review. 1 These class I studies demonstrating the efficacy of visual scanning training for visual neglect strongly support that this type of intervention as a practice standard. Inclusion of limb activation or electronic technologies for visual scanning training is recommended as a new practice option, pending confirmation of efficacy through more rigorous investigation. A class I study 17 provided evidence that computer-based restitution training can reduce the extent of damaged visual fields due to postchiasmal lesions, and produce subjective improvements in visual functioning. We previously noted 1 that the observed reductions in visual field defects were insufficient to explain the associated reduction in functional impairments, and that functional improvement was associated with increased compensation rather than change in the underlying visual field deficit. We recommend that interventions intended to reduce the extent of damaged visual fields be considered a practice option, pending replication by independent investigators. Recommendations for future research. Class I studies verifying the usefulness of technologies that might increase the availability of visual scanning treatment, such as computer projection 22 and the Useful Field of View test 25 introduced in class II and III studies, should be conducted. Class II and III studies demonstrating positive results for limb activation in conjunction with visual scanning treatment for neglect merit verification through class I research. The evidence that visual restitution training may actually result in regeneration of the visual fields makes it increasingly difficult to dismiss this possibility, even though it runs counter to conventional neurologic wisdom. This method merits further independent and rigorous investigation in other centers. Such studies should attempt to differentiate between the contribution of visual restitution training intended to have an impact directly on visual field restrictions and the contribution of compensatory mechanisms (eg, improved scanning) to improved functioning. Prospective, controlled (class I) studies of interventions to improve more complex visuospatial abilities required for functional activities (eg, meal preparation, driving) are recommended. Such interventions were explored through class II and III studies with positive results in our previous review, but further investigation of such interventions is noticeably absent in this update. Remediation of Apraxia There was a new area of focus with 2 class I 30 or Ia 31 studies of rehabilitation for apraxia. Apraxia, the inability to do learned and purposeful activities such as dressing, can result in selfcare deficits and dependence on caregivers. Smania et al 31 compared the effectiveness of gestural training and object use in 13 subjects with limb apraxia after left hemisphere stroke with conventional aphasia therapy. Despite an equal amount of treatment, the group receiving conventional therapies showed no improvement, while the apraxia treatment produced improvement in both ideational and ideomotor apraxia. Donker-

4 1684 UPDATE ON COGNITIVE REHABILITATION, Cicerone voort et al 30 conducted a multisite study of subjects with apraxia from left hemisphere stroke. Subjects were randomized to receive either strategy training integrated into occupational therapy (OT) or conventional OT. The main principle of the strategy training program was the use of internal or external compensations for the apraxic impairment during the performance of activities of daily living (ADLs). 32 The emphasis of conventional treatment was on sensorimotor impairments. During the first 8 weeks of their inpatient stay, all subjects received 15 to 19 hours of treatment, with no difference in the amount of treatment provided in the 2 conditions. Strategy training was more effective in improving ADL function than the conventional OT at the conclusion of the 8 weeks of training. At 5-month follow-up, there were no differences between the 2 groups; apparently because of the continued improvement in the subjects who had received (and in some cases were still receiving) conventional treatment. The investigators suggested that patients receiving conventional therapy needed more therapy to improve their ADL functioning. Clinical recommendations. Two class I 30 or class Ia 31 studies of 126 subjects with left hemisphere stroke provide evidence that apraxia can be treated effectively and may facilitate improvement in functional ADLs during the initial period of inpatient rehabilitation, compared with conventional sensorimotor or aphasia therapies. Specific gestural or strategy training for apraxia after left hemisphere stroke is therefore recommended as a new practice standard during acute rehabilitation. Remediation of Language and Communication Deficits The remediation of neurogenic communication disorders is an active area of research, with 40 studies identified in the current review that addressed a range of language-related impairments. There were 3 class I studies, class Ia study, 36 1 class II study, 37 and 35 class III studies Most of this research involved subjects with left hemisphere stroke (35 articles with a total of 253 stroke subjects). Three studies in which the subjects were primarily people with stroke also included subjects with TBI There were also 4 singlesubject studies of people with TBI There was 1 study 41 of 6 subjects with TBI for a total of 16 subjects with TBI in 8 research articles during this review period. Denes et al 33 evaluated the effect of intensity of treatment for 17 subjects with global aphasia at an average of 3 months postonset, using a functional stimulation approach to therapy. More subjects with intensive treatment reached significant improvement in all language modalities. This concept was supported by 3 class III articles reporting improved functional communication after chronic stroke 46 and improved naming ability of subjects with stroke or TBI 40,43 with increased intensity of treatment. Constraint-induced movement therapy (CIMT) to improve language skills after left hemisphere stroke was evaluated in a class I study of 17 subjects with chronic aphasia. 35 CIMT was described as massed practice of verbal responses that was designed to constrain patients to the systematic practice of speech acts with which they had difficulty. The results demonstrated significant benefit of CIMT compared with conventional aphasia therapy. Elman and Bernstein-Ellis 34 evaluated the effectiveness of group communication treatment after left hemisphere stroke. All participants were at least 6 months postinjury and were stratified according to their initial aphasia severity. They were randomly assigned to receive either group communication treatment or deferred treatment. Group communication treatment focused on initiating conversations and conveying a message, understanding the communication disorder, being aware of personal goals and progress, and having confidence in being able to communicate in personally relevant situations. To control for the effect of social contact, the control subjects participated in social activities (eg, movement classes, church activities) while their group communication treatment was deferred. Participants who received the group communication treatment showed significantly more improvement in functional communication. In an effort to evaluate a cost-effective and functional treatment program, Worrall and Yiu 36 compared a home-based, functional communication program provided by trained volunteers with non-language-based recreational treatment and no treatment conditions, for people with chronic aphasia. Between-group comparisons revealed significant improvement from the functional communication program compared with no treatment, but not compared with recreational activities. There is evidence supporting the use of cuing techniques and semantic analysis to improve naming ability by people with aphasia and TBI, 43,44,46,52,53,56,57,65,69,71 to improve writing skills in patients with chronic aphasia, 46-51,61,67,68 and to improve sentence production in patients with agrammatic aphasia. 39,59,64,70 These studies, along with the studies from the last review, 1 support the efficacy of treating naming disorders in people with chronic aphasia, but there is still no clear evidence that one method is more effective than another. A class III study by Wiseman-Hakes et al 41 supported the effectiveness of group treatment to improve pragmatic communication skills for 6 subjects with TBI. However, 1 class III single-subject study 42 failed to demonstrate the effectiveness of pragmatic training, primarily because of the limited contextual aspects of the treatment. Clinical recommendations. The results from 3 class I studies of treatment of language deficits in 58 subjects after left hemisphere stroke are consistent with the results from 3 class I studies with 169 subjects from our prior review, 1 and provide additional support for the practice standard recommending cognitive linguistic therapies during acute and postacute rehabilitation for such subjects. There is evidence that group communication treatment can produce clinically meaningful improvements in language functioning, including improved functional communication, beyond the effects of social contact alone. 34,41 Two class I studies with 34 subjects 33,35 and 3 class III studies 40,43,46 with 44 subjects support the concept that increased intensity of treatment for subjects with aphasia results in improved communication skills. We recommend that treatment intensity be considered as a key factor in the rehabilitation of language skills after left hemisphere stroke, as a new practice guideline. Several class III studies provide additional support for the established practice guideline that interventions for specific language impairments, such as reading comprehension and language formulation, are effective after left hemisphere stroke or TBI. Two class III studies 53,65 suggest that independent computer use may serve as an adjunct to clinician-directed treatment of word retrieval. Based on these findings, the committee recommends as a practice option that computer-based interventions be considered as an adjunct to treatment when there is therapist involvement; sole reliance on repeated exposure and practice on computer-based tasks without some involvement and intervention by a therapist is not recommended. Recommendations for future research. Given the overall evidence to support the effectiveness of language interventions after left hemisphere stroke, additional research should be directed at specific parameters of treatment related to effectiveness. Several studies suggest that more intense treatment, in-

5 UPDATE ON COGNITIVE REHABILITATION, Cicerone 1685 cluding massed practice, produces significant benefits; this area merits further controlled investigation. Several studies from this and our prior review suggest that intervention provided in the home or community by trained volunteers can improve language functioning. 36,73,74 We identified several additional studies that addressed training for volunteers 75 or caretakers 76 to support the conversation of partners with aphasia; however, we excluded those studied from formal review because they did not directly treat the people with aphasia. This may represent a valuable adjunct or alternative to traditional treatments, particularly for chronic aphasia, that merits further controlled investigation. There remains a compelling need for controlled studies of interventions to improve pragmatic communication in people with TBI. Remediation of Memory Deficits Studies of the remediation of memory deficits have continued to address the use of compensatory strategies, as well as a growing interest in evaluating the application of assistive technology. We reviewed 13 studies in this area, including 3 class I 77,78 or Ia 79 studies, 2 class II studies, 80,81 and 8 class III studies Two class I studies 77,78 addressed the effectiveness of training strategies in memory rehabilitation. Kaschel et al 77 evaluated the use of a simple visual imagery technique for the rehabilitation of participants with mild memory impairment after acquired brain injury. Visual imagery was compared with the standard approach to memory treatment in 7 different rehabilitation centers (eg, practical guidelines to improve memory, use of notebooks and calendars). Both the visual imagery and standard conditions were preceded by 3 months of notreatment baseline; this was followed by memory training for 30 sessions over 10 weeks, with follow-up assessment 3 months later. Significant improvement was apparent for the imagery condition, and was restricted to the therapeutic interval and recall of verbal material, consistent with predictions. The improvements associated with visual imagery training were paralleled by positive changes in relatives ratings of patients memory functioning and were maintained at 3-month follow-up. Ownsworth and McFarland 78 investigated the remediation of everyday memory impairment using a diary combined with self-instructional training. The addition of the self-management strategy to diary use was associated with better maintenance of strategy use and greater decline in memory problems. The results are consistent with a previous single-subject study 90 and also suggest that some forms of compensatory strategy training may be beneficial to patients who are many years postinjury. These class I studies from this review are consistent with the conclusions from our prior review, indicating that compensatory memory strategy training is effective for patients with relatively mild memory impairment. One class Ia study 79 extended the results of a previous class II study 91 investigating the effectiveness of a portable pager (NeuroPage) to improve independence in people with memory and planning problems. During the intervention period, the pager was used to address specific problems in daily functioning that were identified by patients or relatives. Significant improvements in participants completion of everyday tasks resulted, compared with no-treatment and baseline conditions. Use of a pager appears to be particularly beneficial for people who must complete certain tasks on a regular basis, and is facilitated by its ease of use and relevance to patients selfidentified needs. 87 Five additional class III studies 84-86,88,89 investigated the application of some form of external compensation for memory impairments using assistive technology (eg, voice organizers). 84,85,89 Clinical recommendations. Two class I studies 77,78 with 41 participants demonstrated the effectiveness of memory strategies for subjects with mild memory impairments after TBI or stroke, including reducing memory failures. These findings are consistent with the previous 4 class I studies 1 with 91 subjects and support the use of strategy training for subjects with mild memory impairment as a practice standard. These interventions may consist of internal strategies (eg, visual imagery) as well as compensation through the use of notebooks or diaries. The benefits of specific intervention strategies may be evident in discrete aspects of memory performance, for example, visual imagery to facilitate verbal recall. 77 There is evidence that the incorporation of self-management techniques enhances the use, maintenance, and perceived helpfulness of a memory notebook. 78 There is also evidence that interventions to promote the use of external memory aids can benefit people with TBI who are many years postinjury. 78 The use of externally directed assistive devices such as pagers and voice organizers appears to benefit people with moderate to severe memory impairments, including evidence from 1 class Ia 79 study that included subjects for whom previous interventions were ineffective. These interventions facilitate completion of everyday activities that subjects have selfidentified as relevant. The use of external memory aids and assistive devices may require extensive training, 1 or may need to remain under the direction of someone other than the person with memory impairment. 79,86 As with our previous recommendation on the use of interventions directed at the acquisition of specific skills, these interventions appear effective for increasing specific behaviors rather than improving memory function. Based on current evidence and in conjunction with the evidence from our previous review, training in the use of external compensations (including assistive technology) with direct application to functional activities is recommended as a practice guideline in subjects with moderate or severe memory impairment after TBI or stroke. Recommendations for future research. Findings from 1 class I study 77 are notable for suggesting that a specific intervention has a differential impact on different aspects of memory functioning. This finding requires replication. The effectiveness of various assistive technologies to compensate for severe memory impairment should be investigated through additional prospective, controlled studies. Remediation of Executive Functioning, Problem-Solving, and Awareness The area of executive functioning includes several integrative cognitive processes by which people monitor, manage, and regulate the orderly execution of goal-directed ADLs. We reviewed 9 studies concerned with executive functioning, problem solving, and awareness, including 1 class I study, 92 1 class Ia study, 93 and 7 class III studies The class I study 92 evaluated the effectiveness of a problemsolving intervention, goal management training (GMT), on successful task completion. Participants received either 1 hour of GMT or 1 hour of motor skills training. GMT was associated with improved performance on paper-and-pencil measures intended to simulate everyday activities. The fact that the entire treatment in this trial was limited to a single hour of instruction limits the translation of these findings in terms of the clinical application and effectiveness of the intervention. One class Ia study addressed the internalization of selfregulation strategies. Medd and Tate 93 examined the effectiveness of a cognitive-behavioral program of anger management

6 1686 UPDATE ON COGNITIVE REHABILITATION, Cicerone that involved self-awareness and self-instructional training. Efforts were made to increase participants awareness of anger via training to recognize their cognitive, physical, and emotional reactions. Results showed a significant decrease in the outward expression of anger for the treatment group compared with controls, suggesting improved emotional self-regulation as a result of treatment. One class III study 95 evaluated a group treatment program developed to enhance self-regulation skills and self-awareness in 21 patients with acquired brain injury. After treatment, participants reported greater awareness of deficits and increased use and effectiveness of strategies in their daily life. In the Medd and Tate study, 93 there was no change in participants awareness of anger problems with the treatment. Three class III case studies specifically addressed techniques for improving awareness by promoting internal control of behavior: one via self-monitoring techniques 96 and the others using observer feedback and self-evaluation. 97,98 Several of these studies suggest that behavioral improvement is not contingent on increased self-awareness. 93,99,100 Clinical recommendations. One class I study with 30 participants supports the effectiveness of interventions for problem-solving deficits, 92 although the direct application to clinical practice is constrained by the limited extent of the intervention. A class III study 95 demonstrated improved strategy application and psychosocial functioning after training in problem solving. These findings are consistent with a class Ia study with 37 participants from our prior review, 1 and reinforce the practice guideline recommending the training of formal problem-solving strategies and their application to everyday situations and functional activities of people with TBI. Our previous review recommended that cognitive interventions that promote internalization of self-regulation strategies through self-instruction and self-monitoring, be considered a practice option for the remediation of deficits in executive functioning. 1 Two class III studies 94,95 support this recommendation and 1 class Ia study 93 suggests that these techniques may be used to improve emotional self-regulation in patients with TBI. In addition, 3 class I studies indicate that self-instructional training is an effective component of interventions for the remediation of deficits in attention, 12 visual neglect, 18 or memory. 78 The effectiveness of interventions directed at patients poor awareness of deficits has been addressed by 1 class Ia study 93 and 6 class III studies Given the small number of uncontrolled studies and inconsistency of methods and results in this area, there is insufficient evidence to make specific recommendations regarding interventions to improve self-awareness after TBI or stroke. Recommendations for future research. Future studies may incorporate treatment for problems of emotional control and psychosocial skills, particularly as these reflect components of problem-solving and self-regulation interventions. The effectiveness of interventions that attempt to promote the internalization of strategies needs to be addressed through prospective, controlled studies. Our previous review noted the potential use of external strategies for the rehabilitation of executive problems, 101 and this approach might have promise in the treatment of patients with marked difficulties in their planning and organization of everyday activities. 79 Despite the importance that is commonly attributed to awareness as a mediator of rehabilitation outcomes, there continue to be few studies of interventions in this area. Comprehensive-Holistic Cognitive Rehabilitation In comprehensive-holistic cognitive rehabilitation, we included studies that investigated interventions directed at multiple aspects of dysfunction, often addressing a combination of cognitive, emotional, motivational, and interpersonal impairments, in the context of an integrated and programmatic treatment approach. We reviewed 1 class I, class II, 103 and 5 class III studies of comprehensive-holistic programs of cognitive rehabilitation. The class I study by Salazar et al 102 evaluated the efficacy of cognitive rehabilitation for 120 people with moderate to severe TBI within a single, military medical referral center. Of 273 consecutively hospitalized patients, 120 patients met eligibility criteria and participated in the study. They were, on average, about 38 days postinjury. Patients were randomly assigned to receive either multidisciplinary, in-hospital cognitive rehabilitation 109 or a limited home program 110 consisting of individual education and encouragement from a psychiatric nurse. Return to work rates were 90% for the cognitive rehabilitation group and 94% for the home group; the authors noted these extraordinarily high return-to-work rates and suggested that participants high preinjury education and level of functioning, significant degree of spontaneous recovery, and ready availability of (military) employment after injury might have limited the ability to detect any differential benefits from the cognitive rehabilitation program. A subgroup analysis of 75 participants with more severe injuries (those unconscious for 1h) showed a significant benefit from the cognitive rehabilitation program. While this study does not provide strong support for comprehensive-holistic cognitive rehabilitation, the significant subject selection bias and restricted setting markedly limit the ability to generalize these findings to most areas of rehabilitation practice. A principle contribution of the class III studies is their relevance to understanding the impact of comprehensive-holistic cognitive rehabilitation on social participation and community integration after TBI. Four studies with a total of 270 patients evaluated programs based on the principles of neuropsychologic rehabilitation. Most of the participants were people with severe TBI who had received clinical treatment for at least 3 months-in many cases, several years-after injury. Between 39% 104 and 62% 107 of patients were engaged in community-based employment after treatment, and 49% made clinically significant gains in community integration. 106 Clinical recommendations. The single class I study 102 of 120 subjects with TBI did not provide support for comprehensive cognitive rehabilitation compared with basic education and reassurance in the early stages of recovery from TBI; patients with more severe injuries, however, did show greater benefit with the more intense, structured treatment program. However, the ability to generalize from the results of this study is severely constrained by the restricted nature of the population, unique (military) setting, and limited course of treatment. Four class III studies with 265 subjects support the clinical effectiveness of comprehensive-holistic programs of cognitive rehabilitation for improving community integration, social participation, and productivity after TBI or stroke. There is also evidence that gains in community functioning can be achieved by patients 1 or more years postinjury, 106 and that gains from treatment are maintained for several years after rehabilitation. 105,111,112 These studies support our previous conclusion 1 that treatment in postacute programs of comprehensive-holistic cognitive rehabilitation is recommended as a practice guideline in treating people with moderate to severe TBI. The integration of cognitive and interpersonal interventions is characteristic of compre-

7 UPDATE ON COGNITIVE REHABILITATION, Cicerone 1687 Table 1: Differential Treatment Effects of Cognitive Rehabilitation (CR) Compared With Alternative Treatment or Control Conditions Based on All Class I Studies Nature of Treatment Comparison No. of Comparisons No. of Patients Percentage of Comparisons Showing Differential Benefit in Favor of CR CR vs other cognitive or psychosocial treatment CR vs pseudotreatment CR vs conventional rehabil CR vs no treatment Total hensive-holistic cognitive rehabilitation programs. There is also evidence that psychosocial interventions may facilitate the effectiveness of treatments directed at specific cognitive impairments after TBI 11 or stroke. 64 These findings are consistent with our prior recommendation that the integration of individualized cognitive and interpersonal therapies be considered a practice option. Recommendations for future research. There is a particular need in this area to confirm the positive findings from class III studies using controlled trials. This might best be addressed through practical clinical trials that select clinically relevant, alternative interventions for comparison, represent a diverse population of study participants, and use a broad range of relevant health outcomes. 113 The continued use of observational methods is also encouraged, particularly to identify the types of patients who are most appropriate for, and likely to benefit from, comprehensive-integrated cognitive rehabilitation. 114 Comparison of Cognitive Rehabilitation and Alternative Interventions The question of effectiveness of cognitive rehabilitation must be answered in relation to available alternative treatments. For each of the 46 class I studies included in this and our prior review, we examined the nature of the alternative treatment conditions and classified them as no treatment, conventional rehabilitation, pseudotreatment, a psychosocial intervention, or an alternative cognitive intervention. We defined conventional rehabilitation as the provision of standard physical, OT, and/or speech therapy in accordance with routine procedures in a traditional rehabilitation setting. We defined pseudotreatment as providing participants with mental or social stimulation without specific therapeutic intent (eg, recreational computing, leisure activities, repetitive mental exercises). Psychosocial interventions were defined as individual or group psychotherapy or emotional support. Studies that compared different types of specific cognitive interventions (eg, attention training compared with memory training) were considered to provide an alternative cognitive intervention in comparison with the intervention under investigation. We relied on the original authors description of the alternative treatment to make these classifications, although on several occasions we had to rely on our judgment regarding the nature of the comparison. We used only studies that reported direct statistical comparisons between treatments in this analysis. For studies that compared cognitive rehabilitation with more than 1 alternative treatment, cognitive rehabilitation was compared separately with each of the alternatives. For the purpose of this analysis, we collapsed studies using psychosocial or other cognitive interventions as the alternative treatment. This resulted in the comparison of 47 treatment conditions, representing 1801 patients (table 1). It is clear that cognitive rehabilitation provides clinical benefits, compared with not receiving any treatment, and substantial benefits compared with conventional rehabilitation. Cognitive rehabilitation produced greater improvement than pseudotreatment, psychosocial treatment, or an alternative cognitive intervention in about two thirds (64%) of the study comparisons. Overall, cognitive remediation resulted in a significant benefit compared with the alternative condition in about 79% of all treatment comparisons (see table 1). There were 22 treatment comparisons involving 616 patients primarily with TBI. In those studies, 18 of the 22 (81.8%) comparisons demonstrated a differential benefit in favor of cognitive remediation. Among the 25 treatment comparisons involving 1185 patients primarily with stroke, 19 (76%) demonstrated a differential benefit favoring cognitive remediation. In no case was there evidence that the alternative condition was superior to cognitive rehabilitation. Table 2: Evidence-Based Recommendations for Cognitive Rehabilitation: Practice Standards Intervention Visuospatial rehabilitation Cognitive-linguistic therapies Specific interventions for functional communication deficits, including pragmatic conversational skills Specific gestural or strategy training for apraxia Memory strategy training Strategy training for attention deficits Recommendations Recommended for persons with visuoperceptual deficits associated with visual neglect after right hemisphere stroke Recommended during acute and postacute rehabilitation for persons with language deficits secondary to left hemisphere stroke Recommended for persons with TBI Recommended for persons with apraxia after left hemisphere stroke during acute rehabilitation Recommended for persons with mild memory impairments from TBI, including the use of internalized strategies (eg, visual imagery) and external memory compensations (eg, notebooks) Recommended during postacute rehabilitation for persons with TBI. Insufficient evidence exists to distinguish the effects of specific attention training during acute recovery and rehabilitation from spontaneous recovery or from more general cognitive interventions.