PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Brasure M, Lamberty GJ, Sayer NA, et al. Multidisciplinary Postacute Rehabilitation for Moderate to Severe Traumatic Brain Injury in Adults [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2012 Jun. (Comparative Effectiveness Reviews, No. 72.)


Topic Refinement

The initial topic of rehabilitation for TBI for this comparative effectiveness review was nominated to the Effective Healthcare Program through a public process. The topic development materials and our conversations with AHRQ and the nominator clarified the intent of the nomination as follows: to evaluate all forms or types of rehabilitation for all ages and severity levels of TBI, with an emphasis on rehabilitation services provided more than 6 months after the initial injury. Subsequent to the nomination, we recruited Key Informants, including content experts, who cautioned against a review of all ages and severity levels because these are separate bodies of evidence. Specifically, TBI in children and early adolescents is associated with additional complications caused by early stages of brain development.10 Additionally, any impairments sustained after mild TBI tend to differ from those related to moderate to severe TBI.4 Key Informants also argued against an arbitrary 6-month cutoff, emphasizing that rehabilitation timing is unique to each injury. They suggested a more meaningful clinical designation, such as postacute reflecting the time in which patients were considered medically stable and ready to participate in rehabilitation. We formulated initial Key Questions with information gleaned from Key Informant discussions and preliminary literature searching, while maintaining the intent of the original nomination. After approval from AHRQ, we posted preliminary Key Questions to the public Effective Healthcare website. These questions proposed evaluating evidence of effectiveness or comparative effectiveness for most types of postacute rehabilitation (any intervention addressing sustained cognitive, physical, or behavioral impairments) at the specific intervention level or overall program level.

The public comment period provided valuable feedback to our Key Questions, especially: (1) that our proposed scope was excessively broad and might result in conclusions with little meaning; and (2) that multidisciplinary rehabilitation is the commonly accepted approach to sustained impairments from moderate to severe TBI. Based on this feedback—with which members of our Technical Expert Panel (TEP) agreed—we significantly revised the Key Questions to avoid an overly broad scope that could add complexity to an already complicated topic. A broader scope would also have overlapped with the IOM systematic review of cognitive rehabilitation that was already underway. The topic nominator emphasized two priority areas: the effectiveness of multidisciplinary rehabilitation and of cognitive rehabilitation. Thus, our review evaluates the evidence of effectiveness for multidisciplinary postacute rehabilitation for moderate to severe TBI in adults as determined by the primary outcomes of productivity and community integration.

Search Strategy

We searched relevant bibliographic databases to identify evidence for this review. These databases included:

  • Cochrane Central Register of Controlled Trials (CENTRAL)
  • PsycINFO
  • Physiotherapy Evidence Database (PEDro)

We searched for randomized controlled trials (RCTs) and prospective cohort studies published from 1980 to the January 2012. The nature of postacute rehabilitation has transformed over the last 30 years, and studies conducted since1980 reflect programs and services most relevant to the topic today.48

Our search strategy was based on a concept analysis that identified key concepts and relevant controlled vocabulary and natural language. We combined these bibliographic database searches with backwards citations searches of relevant recent systematic reviews. The concept analysis and search strategy appear in Appendix A. We adapted the strategy to conform to controlled vocabulary and indexing used in the other bibliographic databases.

Triage and Screening

We screened bibliographic database search results to identify eligible studies in two stages: triage and screening. During triage, two independent investigators reviewed titles and abstracts of all references resulting from the bibliographic database searches to exclude ineligible studies. Studies not excluded by both investigators during triage underwent screening. Two independent investigators reviewed full text to determine if studies met inclusion criteria. Differences in screening decisions were resolved by discussion or, if necessary, with the help of a third investigator. Eligibility status and one exclusion reason were documented for all studies evaluated at the screening stage.

Inclusion Criteria

We included controlled trials and prospective cohort studies assessing multidisciplinary postacute rehabilitation for moderate to severe TBI in adults age 16 and over (consistent with the definition of adult used by the TBI Model Systems programs and similar research conducted in other countries).

We aimed to include all studies of multidisciplinary interventions. We chose the term “multidisciplinary” for this topic because a clear definition of comprehensive programs does not exist. However, screening studies to determine whether interventions were multidisciplinary was challenging and could result in an inappropriate set of included studies. For example, the “multidisciplinary” screening criterion could lead to inconsistent inclusion of studies of similar interventions simply because some more clearly specified the disciplines involved. Further, clinical practice typically involves many disciplines in delivering these interventions, thus the interventions are to a degree inherently “multidisciplinary.” For these reasons, we chose not to explicitly screen by the term “multidisciplinary.” Finally, our emphasis on community integration outcomes helped assure exclusion of studies examining very specific interventions, such as those aimed at improving memory or gait. We also specifically excluded domain- or impairment-specific interventions such as specific skill building to enhance memory or social skills training even if provided by a multidisciplinary team.

We limited studies to those enrolling at least 75 percent moderate to severe TBI patients. Certain rehabilitation programs are geared to the broader brain injury populations or can include mild TBI patients. However, because our emphasis was on moderate to severe TBI, we felt that including studies addressing the broader brain injury population would not provide the relevant data to draw conclusions specific to this population.

Studies were deemed eligible if they reported one of our preselected primary or secondary outcomes. Primary outcomes included:

  • Return to school, work, or training (or other measures of productivity)
  • Community Integration as measured with (described in Table 2):
    • The Mayo-Portland Adaptability Inventory (MPAI)
    • Craig Handicap Assessment and Reporting Technique (CHART)
    • Craig Handicap Assessment and Reporting Technique Short Form (CHART-SF)
    • Community Integration Questionnaire (CIQ)
Table 2. Primary outcome scales measuring community integration.

Table 2

Primary outcome scales measuring community integration.

As the most relevant outcome, we selected participation demonstrated by productivity or community integration measures. We accepted any definitions of productivity and selected measures deemed most appropriate for measuring community integration. We selected four primary outcome measurement instruments, as follows. First, we selected the MPAI as the most appropriate outcome measurement scale for the population addressed in this review (current version, MPAI-4). The MPAI was specifically developed to evaluate rehabilitation programs in the postacute brain injury population.22 Additionally, the MPAI was recommended by the TBI Common Data Elements Outcomes Workgroup as a supplemental global outcome measure that summarizes overall impact and incorporates functioning, activities, and participation.49 This group also cited the utility of this measure in evaluating progress in rehabilitation. The second scale we selected, the Craig Handicap Assessment and Reporting Technique (CHART), is another promising measure that incorporates community integration assessment in the postacute TBI population. The CHART addresses the ICF’s participation domain and has been tested in TBI populations.50 This scale is available both in the full version and a short form (SF) version. The CHART-SF has been suggested as a core measure of social participation by the TBI Common Data Elements Outcomes Workgroup.49 Finally, we selected the Community Integration Questionnaire (CIQ), which was developed for and has been used extensively in TBI populations and within the TBI model systems programs.51

We did not prespecify all secondary outcome measurement instruments. Instead, we chose to include studies with scales that incorporated community integration or quality, satisfaction with life or other measures of global functioning applicable to community settings. Prespecified secondary outcomes scales included the Extended Glasgow Outcome Scale (GOS-E), the Disability Rating Scale (DRS), and the Satisfaction with Life Scale (SWLS). We identified other scales during the screening process. Descriptions of all secondary outcome measures appear in Table 3. Other measures considered secondary outcomes during the screening process (i.e. not selected a priori) included the EuroQOL (EQ 5D); the Perceived Quality of Life Scale (PQOL); the Brain Injury Community Rehabilitation Outcome-39 (BICRO-39); the Quality of Life Inventory (QOLI); Quality of Community Integration Questionnaire (QCIQ); and the Newcastle Independence Assessment Form (NIAF). We deemed outcomes patient-centered if they (1) directly related to life participation; (2) encompassed indicators of resumption to previous roles in the family and community or quality of life; or (3) addressed functioning in as community settings.

Table 3. Descriptions of secondary outcomes scales.

Table 3

Descriptions of secondary outcomes scales.

We also included prospective cohort studies because of the ethical and operational challenges inherent in conducting rehabilitation RCTs. We considered only studies with comparators of no or alternative interventions, because the extent and timing of spontaneous recovery is not clear (e.g. studies with controls at later stages postinjury were not considered adequate). Additionally, given the number of known and unknown confounding variables affecting rehabilitation outcomes, we paid special consideration to risk of bias in grading of evidence.

Limiting included studies to those published in English is not ideal; however, studies conducted in English are more likely to be applicable to U.S. multidisciplinary postacute rehabilitation programs. We describe specific exclusion criteria used in triage and screening in Table 4. Studies meeting these inclusion criteria were used to address all Key Questions.

Table 4. Exclusion criteria.

Table 4

Exclusion criteria.

Data Extraction

We determined fields to be extracted for each Key Question and extracted data from eligible studies into tables for evidence and relevant outcomes. We believed that the complexity and heterogeneity of this condition and of multidisciplinary postacute rehabilitation required extensive data extraction. We extracted basic study information such as author; year of publication; subject inclusion and exclusion criteria; intervention and control characteristics (program or service components, timing, frequency, duration); followup duration; participant baseline demographics and other relevant preinjury and postinjury characteristics; comorbidities; injury etiology and severity; and descriptions and results of primary outcomes and adverse effects. One investigator extracted select data elements into evidence and outcomes tables, and a second investigator confirmed data extractions for accuracy.

Risk of Bias

Several tools are available to evaluate risk of bias among primary studies. Recommended practice when selecting instruments to evaluate risk of bias when conducting systematic reviews is to use instruments designed specifically for this purpose and to avoid instruments that calculate composite scores.63 We developed risk of bias assessment forms specifically for this project. For RCTs, we modified the Cochrane Risk of Bias tool64 to address specific items that may lead to risk of bias on this topic. Due to the complex nature of the interventions, we incorporated items from the RTI Observational Studies Risk of Bias and Precision Item Bank65 to evaluate intervention and comparison definitions, implementation, and outcomes issues (consistent measurement, validity and reliability of scales, objective vs. subjective measures, providers versus self-report). Building on the work of other researchers,66 we assessed whether the intervention definitions provided adequate detail, including identification of the theory or model driving the specific studied intervention, thorough details about intervention components, and documentation of the intervention in manuals or other publications. We also reviewed studies for validation that the interventions were effectively implemented via staff training and/or fidelity checks. Because many of the outcomes were measured using scales, we added an item assessing the quality and validity of the scale to our risk of bias assessment forms. We also modified the Cochrane questions to simplify the evaluation of each component by directly answering questions instead of assessing the degree of risk of bias for individual elements. We dropped the element related to blinding of participants and personnel because such blinding is unlikely with these interventions. The resulting items on our RCT risk of bias assessment forms included sequence generation; allocation concealment; blinding of outcome assessment; intervention and control description; intervention implementation; outcome measurement; incomplete outcome data; selective outcome reporting; and other issues.

We created a risk of bias assessment form for cohort studies from the RTI Observational Studies Risk of Bias and Precision Item Bank.65 We selected items for consistency with items on the RCT form, and additional items relevant to selection bias and statistical analysis. Final versions of these forms (Appendix B) contained individual items with guidance and space for responses and comments. The last item on each of the forms assigned an overall risk of bias to the study.

Two investigators independently assessed each item using the appropriate form, and then assigned an overall risk of bias assessment of low, moderate, or high to each study. Risk of bias assessments were performed only for primary outcomes. An ‘uncertain’ response was available for particular items on the forms when the determination could not be made based upon what was reported in the study (e.g. no report of blinding of outcomes assessors). We did not contact study authors for additional information. Overall assessments were subjective based upon the assessment of individual items, the magnitude of individual items and the collective risk of bias created by the individual items. Investigators reconciled discrepancies for overall risk of bias by consulting with each other and, when necessary, with a third investigator. RCTs and cohort studies with an overall assessment of high risk of bias were not used to draw conclusions about effectiveness.

Data Synthesis

The diversity of the setting, populations, interventions, controls, outcomes, and outcome measures studied precluded any quantitative synthesis of results. All eligible studies were used to address KQ1. Only studies rated low or moderate risk of bias were used to answer KQ2 – KQ5. Study results are not reported for studies rated high risk of bias. Qualitative syntheses grouped studies by population, intervention setting or type, and outcomes. We evaluated outcomes within groups when more than one study could be appropriately grouped. Results from studies evaluating program effectiveness utilizing measures we selected as secondary outcomes were used to determine consistency of effect with the participation measures selected as primary outcomes.

Grading the Evidence

We evaluated the overall strength of evidence for each primary outcome or comparison using methods developed by the Agency for Healthcare Research and Quality and the Effective Health Care Program.67 We evaluated strength of the evidence on four required domains:

  1. Risk of bias (do the studies for a given outcome or comparison have good internal validity). The risk of bias, based on study design and conduct, is rated low, moderate, or high. Because studies were assessed for risk of bias at the study level and assessments were based on the given study design, evidence level risk of bias assessments are downgraded one level for observational studies.
  2. Consistency (the degree of similarity in the effect sizes—i.e., same direction of effect—of the included studies). Consistency is rated consistent or inconsistent if possible. When evidence on comparisons was based upon a single study, we recorded “single study” for this domain and did not downgrade SOE.
  3. Directness (reflecting a single, direct link between the intervention of interest and the outcome). Directness can either be direct or indirect. Because we assessed SOE only for primary outcomes, we considered all evidence to be direct.
  4. Precision (degree of certainty surrounding an effect estimate of a given outcome). Precision is either precise or imprecise. A precise estimate is one that would yield a clinically meaningful conclusion. Relative risk estimates for dichotomous outcomes were determined imprecise if relative risk increases or reductions exceeded 25 percent; continuous outcomes were considered imprecise if the upper or lower confidence interval crossed an effect size of 0.5 in either direction.68

Two investigators worked independently to qualitatively rate each component and overall strength of evidence. Disagreements were reconciled through discussion among project team members. We rated the overall evidence for each outcome and comparison as:

  1. High: High confidence that the evidence reflects the true effect; further research is very unlikely to change the confidence in the estimate of effect.
  2. Moderate: Moderate confidence that the evidence reflects the true effect; further research may change our confidence in the estimate of effect and may change the estimate.
  3. Low: Low confidence that the evidence reflects the true effect; further research is likely to change the confidence in the estimate of effect and is likely to change the estimate.
  4. Insufficient: Evidence either is unavailable or does not permit a conclusion.

Assessing Applicability

We determined applicability of the studies according to the PICOTS format at the evidence level. Study characteristics that affected applicability include (but are not limited to): narrow eligibility criteria; patient or injury characteristics different than that described by population studies of postacute TBI; and postacute rehabilitation programs or services not typically used in current practice.69

PubReader format: click here to try
Cover of Multidisciplinary Postacute Rehabilitation for Moderate to Severe Traumatic Brain Injury in Adults
Multidisciplinary Postacute Rehabilitation for Moderate to Severe Traumatic Brain Injury in Adults [Internet].
Comparative Effectiveness Reviews, No. 72.
Brasure M, Lamberty GJ, Sayer NA, et al.

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...