Brain Injury, February 2007; 21(2): 107–112
A systematic review of the rehabilitation of moderate to severe acquired brain injuries
ROBERT TEASELL, NESTOR BAYONA, SHAWN MARSHALL, NORA CULLEN, MARK BAYLEY, JOSIE CHUNDAMALA, JIMMY VILLAMERE, DAVID MACKIE, LAURA REES, CHERYL HARTRIDGE, CORBIN LIPPERT, MAUREEN HILDITCH, PENNY WELCH-WEST, MARGARET WEISER, CONNIE FERRI, PAT MCCABE, ANNA MCCORMICK, JO-ANNE AUBUT, PAUL COMPER, KATHERINE SALTER, ROBERT VAN REEKUM, DAVID COLLINS, NORINE FOLEY, JOZEF NOWAK, JEFF JUTAI, MARK SPEECHLEY, CHELSEA HELLINGS, & LINH TU St. Joseph’s Health Care London, Ottawa Rehabilitation Centre, and the Toronto Rehabilitation Institute, and affiliated with Physical Medicine and Rehabilitation at the Schulich School of Medicine and Dentistry, University of Western Ontario, Faculty of Medicine, University of Ottawa, and Faculty of Medicine, University of Toronto, Ontario, Canada (Received 16 June 2006; accepted 4 December 2006)
Abstract Objective: To conduct a systematic review of the rehabilitation literature of moderate to severe acquired brain injuries (ABI) from traumatic and non-traumatic causes. Methods: A review of the literature was conducted for studies looking at interventions in ABI rehabilitation. The methodological quality of each study was determined using the Downs and Black scale for randomized controlled trials (RCTs) and non-RCTs as well as the Physiotherapy Evidence Database (PEDro) scale for RCTs only. Results: Almost 14 000 references were screened from which 1312 abstracts were selected. A total of 303 articles were chosen for careful review of which 275 were found to be interventional studies but only 76 of these interventional studies were RCTs. From this, 5 levels of evidence were determined with 177 conclusions drawn; however of the 177 conclusions only 7 were supported by two or more RCTs and 41 were supported by one RCT. Conclusion: Only 28% of the interventional studies were RCTs. Over half of the 275 interventional studies were single group interventions, pointing to the need for studies of improved methodological quality into ABI rehabilitation. Keywords: Brain, brain injuries, craniocerebral trauma, evidence-based medicine, rehabilitation, review literature
Introduction Acquired brain injury (ABI), particularly of traumatic etiology, is one of the leading causes of death and lifelong disability in North America, particularly in children and adolescents [1, 2]. In western developed countries, the incidence of traumatic
brain injury (TBI) is estimated to be approximately 250–300 per 100 000 population [3]. It is estimated that the annual incidence of TBI in the United States is 1.5 million [2]. In Ontario, a province with a population of over 12 million people, the Ontario Brain Injury Association estimates that the total
Correspondence: Robert Teasell, MD, FRCPC, Chair, Department of Physical Medicine and Rehabilitation, St. Joseph’s Health Care London, London, Ontario, Canada. Tel: þ1 519 685 4292, ext. 44559. Fax: þ1 519 685 4023. E-mail:
[email protected] ISSN 0269–9052 print/ISSN 1362–301X online ß 2007 Informa UK Ltd. DOI: 10.1080/02699050701201524
108
R. Teasell et al.
annual number of brain injuries is over 18 000 with nearly 4000 annual injuries occurring in the pediatric population alone (0–14 years) [4]. Such figures demand that the treatment approaches used in ABI rehabilitation be optimized. Evidence-based practice, ‘attempts to implement the results of research trials (evidence) and to translate those results into clinical practice, with the presumed goal of improving the effectiveness of clinical care’ [5]. Thus, due to its potential to improve patient care, evidence-based reviews such as this one are important to assess the efficacy of clinical practices. The Evidence-Based Review of Rehabilitation of Moderate to Severe Acquired Brain Injuries (ERABI) was designed to comprehensively review current practices in acquired brain injury (ABI) rehabilitation and aims to achieve the following: (i) identify effective treatment interventions; (ii) identify gaps in the literature deserving further research and; (iii) serve as an accessible tool for clinicians in an effort to encourage improved evidence-based practice.
include focal brain lesions, anoxia, tumours, aneurysm, vascular malformations, and infections of the brain [6]. Given that ABI is a loosely defined term, studies referencing ABI are often similarly vague in regards to their subject populations, and might include any combination of persons with traumatic brain injuries, diffuse cerebrovascular events (such as a subarachnoid hemorrhage) or diffuse infectious disorders (such as encephalitis or meningitis). For example, some studies might include persons with all types of brain lesions apart from stroke (e.g., brain tumours) as ABI subjects. Most ABI patients have a traumatic etiology; in fact, many of the studies involve TBI patients alone. For the purposes of this evidence-based review, we used the definition of ABI employed by the Toronto Acquired Brain Injury Network [6] in which ABI is defined as ‘damage to the brain that occurs after birth and which is not related to congenital disorders, developmental disabilities, or processes that progressively damage the brain’.
Defining severity of injury Defining acquired brain injury ABI is an umbrella term, encompassing a wide spectrum of brain injuries that generally includes traumatic and non-traumatic etiologies such as cerebral concussion, brain contusions, subarachnoid hemorrhages or other ‘acquired’ problems. ABI typically produces a potentially wide range of impairments affecting physical, neurocognitive and/ or psychological functioning. Therefore, a person with an ABI may refer to any of the following: An individual with a mild, moderate, or severe traumatic brain injury (TBI); an individual with Herpes encephalitis; an individual with viral meningitis; or an individual with acute hypertensive encephalopathy. As opposed to an insidious developmental process, an ABI infers that a person previously intact from a neurological perspective subsequently ‘acquires’ some form of brain pathology at some point during his or her lifespan. ABI can therefore result from traumatic and non-traumatic causes. Common traumatic causes include motor vehicle accidents, falls, assaults, gunshot wounds, and sport injuries [1]. Non-traumatic causes of ABI
ABI severity is usually classified according to the level of altered consciousness experienced by a patient following injury. Consciousness levels following ABI can range from transient disorientation to deep coma. Patients are classified as having a ‘mild’, ‘moderate’ or ‘severe’ ABI according to their level of consciousness at the time of initial assessment. Various measures of altered consciousness are used in practise to determine injury severity following ABI. Some of the most common measures include the Glasgow Coma Scale (GCS), the duration of loss of consciousness (LOC), and the duration of post-traumatic amnesia (PTA). Table I outlines the measures of severity used for the selection of studies in this review.
Methods Literature search strategy An extensive literature search using multiple databases (CINAHL, EMBASE, MEDLINE, and PsycINFO) covering the years 1980–2005 was
Table I. Definitions of ABI injury severity. Mild PTA < 1 hour GCS 13–15 LOC < 15 minutes
Moderate
Severe
Very severe
PTA 1–24 hours GCS 9–12 LOC < 6 hours
PTA 1–7 days GCS between 3 and 8 LOC 6–48 hours
PTA > 7 days LOC > 48 hours
Rehabilitation of moderate to severe ABI conducted to identify all published investigations which evaluated the effectiveness of any treatment or intervention related to ABI. We limited our search to studies dealing with rehabilitation and therapy. Both prospective and retrospective studies were considered, as were studies that used either experimental (randomized trials) or non-experimental designs (prospective and retrospective controlled trials, single group interventions, retrospective studies and cases studies). Studies cited in review articles, meta-analyses, systematic reviews, or in selected study articles but not identified through the original literature search strategy were also included. Unpublished data or studies were not included. Reference Manager 10.0Õ was used for database management.
109
of the brain injury took on special significance; therefore, articles which did not meet our definition of ABI or which did not fall under the moderate to severe range of injury severity were excluded. Data extraction Once an article was selected for full review, the following data were extracted: Authors, place and date of publication, inclusion and exclusion criteria, sample sizes, details regarding the population studied (i.e., type of injury, severity, source, sex, age, time since injury), treatment, outcome measures, and results. These data were summarized using large tables. Articles evaluating similar treatments were grouped together. Methodological quality assessment
Study inclusion criteria Every effort was made to identify all relevant articles that evaluated any rehabilitation intervention, which assessed any outcome measure during any time frame within the recovery period following brain injury. Two independent reviewers evaluated all reference titles obtained from the various databases or from other sources. Any reference title that appeared to involve ABI and a treatment was selected. Abstracts, from the selected reference titles were then reviewed by two independent reviewers to determine if the studies met the criteria for full review. Studies where at least 50% of the subject population met our definition for ABI and which involved the evaluation of a treatment with measurable outcomes were selected. Studies underwent full review only when two independent reviewers agreed that a study met the two inclusion criteria stated above based on the information available in the article abstract. A third independent reviewer settled any discrepancies. As indicated above, the purpose of this evidence-based review was to investigate the efficacy of rehabilitation interventions or treatments for ‘moderate’ to ‘severe’ acquired brain injury. Thus, for the purposes of this review, any studies dealing with ‘mild’ forms of ABI (Table I) were excluded and defining severity
Two blinded, independent reviewers rated the methodological quality of all selected articles. For randomized controlled trials (RCTs) two rating scales were used: The Physiotherapy Evidence Database (PEDro) rating scale developed by the Centre for Evidence-Based Physiotherapy (CEBP) in Australia [7] and the Downs and Black rating scale [8] developed in England. Studies that used a non-experimental or uncontrolled design (prospective and retrospective controlled trials, single group interventions, retrospective studies and cases studies) were evaluated using the Downs and Black rating scale only. This tool was identified in a review by the Health Technology Group [9] as one of the most appropriate for the evaluation of non-RCTs in systematic reviews. Whenever rating discrepancies occurred, the two reviewers reached consensus or a third independent reviewer evaluated the study to settle any disagreements (Table II). Determining levels of evidence Interpreting the results of individual studies. For RCTs, studies scoring 9–10 on the PEDro scale were considered to be of ‘excellent’ methodological quality. Studies with PEDro scores ranging from 6–8 were considered to be of ‘good’ quality, while studies scoring between 4–5 were of ‘fair’ quality.
Table II. Methodological quality assessment tools used in the ERABI. Scale Downs and black scale [5]
Physiotherapy evidence database (pedro) scale [4]
Features
Type of studies assessed
Suitable for assessing both RCTs and non-RCTs. Consists of 27 questions, grouped into four sections: Reporting, external validity, bias and confounding. Scores range from 0–34 10-point ordinal scale suitable for evaluating RCTs. Scores range from 0–10
All studies in this review
RCTs only
110
R. Teasell et al. Using this system, conclusions were easily arrived at when the results of multiple studies were in agreement. However, interpretation became difficult when the study results conflicted. In cases where RCTs also differed in terms of methodological quality, the results of the study (or studies) with the higher PEDro score(s) were more heavily weighted to arrive at the final conclusions. However, there were still some instances where interpretation remained problematic. For instance, the authors needed to make a judgment when the results of a single study of higher quality conflicted with those of several studies of inferior quality. In these cases we attempted to provide a rationale for our decision and to make the process as transparent as possible. It was noted that some areas of study did not include any interventional studies. Rather authors relied on information from studies that resulted from mail out surveys or interviews. This information was in fact included but levels of evidence were withheld.
Studies that scored below four (
