An Overview of Published Research about the Acute Care and Rehabilitation of Traumatic Brain Injured and Spinal Cord Injured Patients

Introduction

T raumatic brain injury (TBI) and spinal cord injury (SCI) have substantial personal, financial, and economic costs. Globally, at least 10 million people each year suffer a TBI resulting in death or hospitalization (Murray and Lopez, 1996). TBI can result in a wide range of ongoing impairments such as cognitive problems, physical disabilities, and changes in personality and behavior (Khan et al., 2003). A recent Australian study estimated lifetime costs of $2.5 million and $4.8 million per case for moderate and severe TBI, respectively, and an overall annual cost to Australia of $8.6 billion (Access Economics, 2009). The worldwide incidence of SCI is reported to be 46–81 per million population (Ackery et al., 2004). Aside from the effects of paralysis on movement, SCI can cause a range of ongoing physical and medical problems, including spasticity, pressure ulcers, and bowel and bladder dysfunction (McDonald and Sadowsky, 2002), that impact upon the independence and psychological well-being of people with SCI (Ackery et al., 2004). The estimated lifetime cost of SCI in Australia is $5 million for paraplegia and $9.5 million for quadriplegia, with a total annual cost to Australia of $2 billion (Access Economics, 2009), representing a substantial financial burden to the community.

In prioritizing the limited resources available for TBI and SCI research, it is important that resources are directed to areas of greatest need, and that unnecessary duplication of existing research is avoided. Grant applications often build on previous research, resulting in a tendency to fund already strong areas of research rather than less productive areas that may be more important. Sometimes research in a particular area continues well beyond the time when equipoise no longer exists because previous research was not known or fully understood. For example, a cumulative meta-analysis of 64 randomized controlled trials (RCTs) comparing aprotinin to placebo for reducing perioperative bleeding during cardiac surgery demonstrated that a large number of RCTs were conducted after the point where a systematic review (SR) of existing literature would have established the efficacy of this intervention (Fergusson et al., 2005). This resulted in patients needlessly being denied a proven treatment through randomization to a placebo group in redundant RCTs. Fergusson's review underlines the importance of research prioritization; where research evidence is strong, guidelines and best practices can be promoted with a greater focus on optimizing the translation of evidence into clinical practice; where research evidence is weak, new knowledge needs to be generated through the conduct of primary research.

Comprehensive knowledge of the extent and quality of existing neurotrauma research evidence, and where research gaps exist, is useful in planning research and strategically directing research resources to areas of need. Since 2005, three large projects have been undertaken to create user-friendly, online neurotrauma research evidence resources for this purpose: the Global Evidence Mapping (GEM) Initiative, based in Australia, and the Acquired Brain Injury Evidence-Based Review (ERABI) and the Spinal Cord Injury Rehabilitation Evidence (SCIRE) Project, both based in Canada.

The GEM Initiative (www.evidencemap.org) has developed evidence maps describing the volume, nature, and characteristics of TBI and SCI research in pre-hospital, acute, and rehabilitation/long-term care settings. Topics for the GEM evidence maps were developed and prioritized by in-depth consultation with neurotrauma stakeholders (i.e., patients, carergivers, clinicians, researchers, and policymakers). Systematic searching and selection of evidence relevant to prioritized topics was undertaken. Study and patient characteristics were then extracted from relevant articles to form individual evidence mapping reports for a range of TBI and SCI topics across pre-hospital, acute, and rehabilitation care settings. Since the inception of the project in 2007, a total of 58,263 abstracts have been screened, and 1644 relevant neurotrauma publications mapped. Fifty-three evidence maps for high-priority questions have been completed (Bragge et al., 2011).

The ERABI project (www.abiebr.com) is designed to be a comprehensive, accessible, up-to-date review of the current evidence in acquired brain injury (ABI) that provides levels of evidence for individual topics. The primary goal is to provide access to updated research syntheses for front-line clinicians and decision makers. The ERABI team has reviewed over 27,000 titles containing the keyword “brain injury,” which led to 6000 abstracts being reviewed to determine eligibility. Ultimately 732 articles were selected for data abstraction and quality assessment using the PEDro and Downs and Black tools, of which 205 were RCTs (134 rehabilitation and 71 acute ABI). Information was summarized by topic in a table format and levels of evidence were assigned for each topic according to a modified Sackett criterion (Teasell et al., 2007). The ERABI is organized into modules or chapters that were divided by the investigators into topics familiar to clinicians reflecting brain injury epidemiology and efficacy of rehabilitation services, and then according to functional domain, for example motor, cognitive, sensory, and emotional functions.

The SCIRE project (http://www.scireproject.com/) is similar to the ERABI in methodology and intent with some overlap of team membership. The SCIRE team has reviewed 1650 SCI published studies which cover a comprehensive array of topics across the continuum, such as impairments (e.g., spasticity), activities (e.g., mobility), participation (e.g., employment), and environment (e.g., housing) (Eng et al., 2007).

These three neurotrauma evidence databases form an excellent platform for neurotrauma research planning, synthesis, and implementation. However, their potential use as a combined resource has not been explored. The aim of this study was to combine the outputs of GEM, ERABI, and SCIRE, to provide an overview of existing TBI and SCI research articles, and to inform the identification of knowledge translation (KT), systematic review (SR), and primary research opportunities.

Methods

A team of four neurotrauma researchers including a rehabilitation physician (M.T.B.) and a physiotherapist (P.B.) undertook the process of synthesizing topics and relevant articles from the GEM, ERABI, and SCIRE projects as follows.

All ERABI TBI topics (i.e., chapter headings) were tabulated alongside all GEM TBI topics (i.e., titles of evidence mapping reports), and all SCIRE SCI topics were listed alongside all GEM SCI topics. Topics within and between ERABI, SCIRE, and GEM were merged where there was conceptual overlap between them. For example, the ERABI chapters “Psychiatric interventions” and “Neuropharmacology” were merged to create the single topic “Pharmacological therapies for neurological recovery;” the topic “Pharmacological management for raised intracranial pressure” was created by merging the GEM topic of the same name with the ERABI chapters called “Acute interventions” and “Neuropharmacology.” Agreement was reached among the four researchers regarding the face validity of combining topics.

Evidence yields for each topic were calculated by summing the number of relevant SRs, RCTs, and all other study designs identified. For topics that overlapped between GEM/ERABI or GEM/SCIRE, duplicate references were identified and removed.

Topics where no relevant research had been identified were removed. Although these topics reflect areas of research/stakeholder interest in TBI and SCI, this project was designed to describe the volume and nature of extant literature rather than to identify research questions.

“Lumping” or “splitting” of the remaining topics was undertaken based upon consideration of topic complexity and yield. For example, the topics “TBI seizure prophylaxis” and “TBI seizure management” were lumped into “TBI seizure prophylaxis and management;” the topic “management of hand, limb and motor function” was split into “non-surgical” and “surgical” topics. Decisions on lumping and splitting were made through discussion and agreement between two clinicians (M.T.B. and P.B.).

To optimize the utility of the resource, short topic names (approximately 10 words) were developed encompassing the key information from the chapter headings and/or evidence mapping titles that created each topic.

Each topic and its associated yield (number of SRs, RCTs, and other studies) was categorized according to:

• Type of neurotrauma: TBI or SCI

This created a total of 8 tables outlining the neurotrauma research evidence gathered by the three projects.

Results

Supplementary Table 1 summarizes the topic synthesis process (see online supplementary material at http://www.liebertonline.com). A total of 145 topics (66 TBI and 79 SCI) were identified across the three resources. The 145 topics yielded 3466 articles: 1256 covering TBI topics (1143 studies of interventions and 113 non-interventional studies), and 2210 covering SCI topics (1797 interventional and 413 non-interventional studies). The numbers of SRs and RCTs were comparable for TBI (31 SRs and 274 RCTs) and SCI (50 SRs and 267 RCTs). However, almost double the number of studies other than SRs or RCTs was identified for SCI (1893) compared to TBI (951). Tables 1 –4 list the TBI acute, acute non-intervention, rehabilitation, and rehabilitation non-intervention topics and yields; Tables 5 –8 contain the equivalent SCI topics and yields.

Discussion

This comprehensive overview of published studies in TBI and SCI is the product of years of systematic searching, study selection, and appraisal by several research teams spanning two countries, using recognized and robust methodologies. Resources that catalogue research evidence across an entire research field, known as scoping or evidence-mapping studies, enable an examination of the extent, range, and nature of research activity, and identification of research gaps (Arksey and O'Malley, 2005; Katz et al., 2003). This information can be used by funding agencies to evaluate research needs and identify priorities; by researchers to aid research planning by gaining an overview of what has been done in a range of topic areas; and by clinicians to rapidly access research evidence in a given topic area (Bragge et al., 2011). We have demonstrated the utility of this dataset by using it to identify topics with relatively high KT potential for an international neurotrauma KT consensus workshop held in Denver, Colorado, in October 2010.

Gathering studies across an entire research field also enables identification of and reflection on research trends and patterns of activity. Both TBI and SCI are characterized by one topic area with a high volume of identified studies: cognitive therapies for TBI, and prevention and management of urinary tract problems post-SCI. Concentrations of research efforts in particular topic areas may reflect the identified priorities of patients, clinicians, and other stakeholders, which are subsequently areas for which funders direct resources; for example, problems in the urinary tract have a substantial impact on quality of life post-SCI. Alternatively, it may be that the inherent complexity of the topic requires a number of studies examining different facets of the problem. Conversely, there were areas with clinical importance and impact on quality of life, such as substance abuse following TBI, for which very little published research was identified; and other areas with a relatively high volume of primary studies but no published synthesis of existing knowledge. These findings demonstrate the variability in the volume and type of research efforts directed at different TBI and SCI research topics. Identification of and consensus about important research topics are complex issues that are subject to a variety of potential influences. Recent events such as the 2011 International Conference on Spinal Cord Medicine and Rehabilitation, which focused on developing an agenda for SCI rehabilitation research over the next decade (International Spinal Cord Society/American Spinal Injury Association, 2011), indicate that the neurotrauma community is actively engaged in dialogue regarding research priorities. This overview of SCI and TBI research evidence contributes to this dialogue.

There was a marked disparity in the number of non-interventional studies of SCI (413) compared to TBI (113). This may reflect the relative difficulties in coding and classifying TBI compared to SCI. Most brain injuries (80%) are mild; however, data on mild TBI are difficult to capture because these injuries are often either not treated, or treated outside of hospitals. Furthermore, TBI often occurs in multi-trauma and military settings where data is underreported or incorporated into general death and injury statistics (Hyder et al., 2007; Langlois et al., 2006). In relative terms, epidemiological data on SCI may be easier to capture and aggregate, as SCI injury classification is usually based on the internationally accepted ASIA Impairment Scale (American Spinal Injury Association, 2002).

The primary limitation of this current exercise, which is common to similarly broad research overviews, is that we did not include quality appraisals or levels of evidence for each of the topic areas. However, it was not the intention of this project to provide definitive statements regarding study quality or effectiveness of interventions. Rather, these tables provide the starting point for further exploration and discussion. The reader may access information for each topic regarding study characteristics and quality appraisal by accessing the original research syntheses from the source projects (i.e., GEM, ERABI, and SCIRE). It is acknowledged that tables outlining numbers of SRs, RCTs, and other studies have the potential to be misleading. Various caveats should be considered when evaluating KT, SR, and primary research opportunities from these tables. In topic areas with a number of SRs, RCTs, and other studies, the quality and consistency of the evidence needs to be considered when determining KT potential. For example, if all the RCTs in a topic area were negative, the KT priority would be to warn clinicians against use of these approaches. For areas where there are multiple RCTs and other studies, consideration of the breadth of the topic and the heterogeneity of these studies will determine the feasibility of conducting a systematic review. Finally, for non-intervention questions (e.g., diagnosis and epidemiology), an RCT is not feasible, and furthermore, even for interventional topics, there are various limitations to the conduct of RCTs in neurotrauma populations; for example in SCI due to heterogeneous and relatively small populations at single sites (Tator, 2006). Another limitation to RCTs is the nature of the interventions, which, in the case of therapy delivered by a therapist, may preclude blinding.

In developing this resource, decisions on the creation, lumping, and splitting of topics were somewhat arbitrary, although it should be emphasized that the topics were drawn from established evidence synthesis projects that have undertaken considerable work on TBI and SCI topic development and classification. As a result, the tables may not match the needs and preferences of all users. However, this is inevitable in the context of an overview of this breadth, and it is unlikely there will ever be universal agreement on such issues. There is also some variation in operational definitions between projects (for example, GEM focused on traumatic brain injury and ERABI on acquired brain injury). Because the ERABI and SCIRE projects are themselves review projects, it is possible that published SRs have been under-identified in creating this resource. However, SRs were part of the GEM Initiative's search strategies, and furthermore, SRs known to the other two projects have been considered in the preparation of this article. For example, since these tables were created for our Denver workshop, SRs have been published or identified by the SCIRE team in the areas of autonomic dysreflexia, venous thromboembolism, orthostatic hypotension, non-pharmacological respiratory management, bone density/management of osteoporosis, pressure ulcers, and sexual and reproductive health. A final limitation to consider is the relative emphasis in the ERABI and SCIRE projects on rehabilitation literature, although this was offset by the GEM resource, which focused on equal coverage of acute and rehabilitation topics.

Like all evidence resources, keeping this information up to date and maximizing its utility is an ongoing challenge. This unique resource allows new studies to be contextualized within the existing dataset of TBI and SCI research from GEM, ERABI, and SCIRE, to facilitate informed consideration of research opportunities and priorities. Further work to enhance this resource could include refinement of the topic list, developing reference lists that link to each topic, incorporating a live update search for each topic to enable new evidence to be identified, and online versions of the resource.

In conclusion, this de-duplicated dataset of SRs, RCTs, and other studies synthesizes 145 TBI and SCI topics from three large neurotrauma evidence synthesis resources. This has enabled the identification of KT, SR, and primary research opportunities and has highlighted research trends in TBI and SCI. This resource has demonstrated value in neurotrauma research planning, and can also facilitate rapid access to topic-specific evidence for clinicians, and inform strategic planning for funders.