Low-value clinical practices in adult traumatic brain injury: an umbrella review protocol.

INTRODUCTION: Traumatic brain injury (TBI) leads to 50 000 deaths, 85 000 disabilities and costs $60 billion each year in the USA. Despite numerous interventions and treatment options, the outcomes of TBI have improved little over the last three decades. In a previous scoping review and expert consultation survey, we identified 13 potentially low-value clinical practices in acute TBI. The objective of this umbrella review is to synthesise the evidence on potentially low-value clinical practices in the care of acute TBI. METHODS AND ANALYSIS: Using umbrella review methodology, we will search Cochrane Central Register of Controlled Trials, Embase, Epistemonikos, International Prospective Register of Systematic Reviews (PROSPERO) and PubMed to identify systematic reviews evaluating the effect of potential intrahospital low-value practices using tailored population, intervention, comparator, outcome and study design questions based on the results of a previous scoping review. We will present data on the methodological quality of these reviews (Assessing the Methodological Quality of Systematic Reviews-2), reported effect sizes and strength of evidence (Grading of Recommendations, Assessment, Development and Evaluation). ETHICS AND DISSEMINATION: Ethics approval is not required as original data will not be collected. Knowledge users from five healthcare quality organisations and clinical associations are involved in the design and conduct of the study. Results will be disseminated in a peer-reviewed journal, at international scientific meetings and to clinical, healthcare quality and patient-partner associations. This work will support the development of metrics to measure the use of low-value practices, inform policy makers on potential targets for deimplementation and in the long term reduce the use of low-value clinical practices in acute TBI care. PROSPERO REGISTRATION NUMBER: CRD42019132428. © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

State-of-the-art synthesis of evidence on low-value clinical practices in the care of acute traumatic brain injury (TBI).

Represents a crucial step towards the deimplementation of low-value practices in acute TBI care.

Adopts an integrated knowledge translation model to ensure the results are relevant to decision makers.

For feasibility reasons, our synthesis is restricted to systematic reviews published in English since 1990.

The scope of review and the inclusion of systematic reviews precludes meta-analysis.

Introduction

Traumatic brain injury (TBI) is the main cause of mortality from injury in people under 45 years of age,1 and it leads to approximately US$60 and €33 billion in total medical costs in the USA2 and Europe3 each year, respectively. Moreover, outcomes following TBI have not improved significantly in the last four decades.4 5 Intervention and treatment options for TBI are multiple, but many lack robust evidence of their effectiveness.6 7

Low-value clinical practices, defined as a test or procedure that is not supported by evidence and/or could expose patients to unnecessary harm8–15 consume up to 30% of healthcare budgets.9 16 In the past decade, the medical community has turned towards the deimplementation of low-value practices as a promising means to reduce the strain on healthcare budgets, free-up resources and reduce harm to patients.17 Physicians report using low-value practices because of a lack of alternative treatment options, fear of legal consequences but also because of lack of guidelines on low-value care.15 18 The Brain Trauma Foundation, among others, publish guidelines on TBI care.19 However, emphasis is on practices that should be adhered to rather than practices that should be avoided. Choosing Wisely publish recommendations specifically targeting low-value practices but few pertain to TBI care and many are based uniquely on expert consensus.11 A previous scoping review and expert consultation survey identified 13 potentially low-value clinical practices in acute TBI care.20 These practices represent potential targets for guidelines, overuse metrics and deimplementation interventions. However, before recommendations can be made, we need to synthesise the evidence base for these practices.

Interventions and treatment options for acute TBI have been the subject of multiple systematic reviews.21 22 Given this large body of available evidence, evidence maps have previously been used to summarise evidence from systematic reviews on acute TBI interventions.7 23 However, these evidence maps were not designed to target low-value practices and focused on moderate to severe TBI when the mild TBI population represent great potential for reducing low-value care. In addition, previous reviews have not provided a synthesis of effect sizes or strength of evidence. The objective of the present study is to synthesise the evidence on potentially low-value intrahospital clinical practices in acute adult TBI.

Methods and analysis

Given the multitude of systematic reviews available for the clinical practices identified as potentially low-value (over 60 were identified in our scoping review), we opted to conduct an umbrella review (a systematic review of systematic reviews).20 While the former aimed to fill a knowledge gap on medical overuse for acute injury care by identifying all potential low-value clinical practices, the latter will synthesise the evidence on the low-value practices pertaining to TBI. The review will be conducted according to published guidelines.24–26 In the absence of reporting guidelines for umbrella reviews, we will use the applicable Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Protocols.27

Eligibility criteria

The project steering committee comprising clinicians (two emergency physicians, seven critical care physicians, one neurosurgeon), methodologists (four) and health system managers (three) used the population, intervention, comparator, outcome and study design (PICOS) framework to develop specific research questions for each potentially low-value clinical practice (table 1).20 28 We will consider systematic reviews of original studies evaluating the effectiveness of predetermined clinical practices in acute TBI in adults (≥16 years old) without restriction on location of publication but limited to studies published in English since 1990.25 26

Table 1

PICOS for each clinical practice

#	Clinical practice
Mild traumatic brain injury
1	Population: adults with acute mild traumatic brain injuryIntervention: validated clinical decision rule (eg, CCHR, CHIP, NEXUS II, NOC)Comparator: nonePrimary outcome: false negative rate (intracranial injury, neurosurgical intervention)Secondary outcomes: sensitivity, specificityStudy design: systematic review
2	Population: adults with acute mild complicated traumatic brain injuryIntervention: routine repeat head CT in absence of neurological deteriorationComparator: none or no repeat head CT in absence of neurological deteriorationPrimary outcome: progression of intracranial injurySecondary outcomes: neurosurgical intervention, mortality, change in management, hospital length of stayStudy design: systematic review
3	Population: adults with acute mild traumatic brain injury and on anticoagulant and/or antiplatelet therapyIntervention: routine repeat head CT in absence of neurological deteriorationComparator: none or no repeat head CT in absence of neurological deteriorationPrimary outcome: progression of intracranial injurySecondary outcomes: neurosurgical intervention, mortality, change in management, hospital length of stayStudy design: systematic review
4	Population: adults with acute mild traumatic brain injury who are negative on head CTIntervention: neurosurgical consultationComparator: none or no neurosurgical consultationPrimary outcome: hospital admissionSecondary outcomes: neurosurgical intervention, mortality, ICU admission, repeat head CT, hospital length of stayStudy design: systematic review
5	Population: adults with acute mild complicated traumatic brain injury who are not on irreversible anticoagulationIntervention: intensive care unit admissionComparator: admission to regular ward or step-down unitPrimary outcome: neurological/medical decline, neurosurgical interventionSecondary outcomes: medical interventions, mortality, adverse events, hospital length of stay, discharge destinationStudy design: systematic review
Moderate and severe traumatic brain injury
6	Population: adults with acute traumatic brain injury on antiplatelet therapyIntervention: platelet transfusionComparator: no platelet transfusionPrimary outcome: GOS or GOS-ESecondary outcomes: mortality, adverse events, hospital and ICU length of stayStudy design: systematic review
7	Population: adults with basal skull fractures without evidence of cerebrospinal fluid leakageIntervention: antibiotic prophylaxisComparator: no antibiotic prophylaxisPrimary outcome: meningitis (confirmed by lumbar puncture)Secondary outcomes: GOS or GOS-E, mortality, surgical correction in patients with CSF leakage, non-CNS infection, hospital and ICU length of stayStudy design: systematic review
8	Population: adults with acute traumatic brain injury and no refractory intracranial hypertensionIntervention: therapeutic hypothermiaComparator: no therapeutic hypothermiaPrimary outcome: GOS or GOS-ESecondary outcomes: intracranial pressure, mortality, adverse events, hospital and ICU length of stayStudy design: systematic review
9	Population: adults with acute traumatic brain injuryIntervention: antibiotic prophylaxis for external ventricular drain placementComparator: no antibiotic prophylaxisPrimary outcome: ventriculostomy-related infectionSecondary outcomes: GOS, mortality, hospital and ICU length of stayStudy design: systematic review
10	Population: adults with acute traumatic brain injury and no refractory intracranial hypertensionIntervention: neuromuscular blocking agentsComparator: no neuromuscular blocking agentsPrimary outcome: GOS or GOS-ESecondary outcomes: intracranial pressure, mortality, adverse events, hospital and ICU length of stayStudy design: systematic review
11	Population: adults with acute traumatic brain injuryIntervention: plasma transfusionComparator: no plasma transfusionPrimary outcome: GOS or GOS-ESecondary outcomes: mortality, adverse events, hospital and ICU length of stayStudy design: systematic review
Severe traumatic brain injury
12	Population: adults with acute severe traumatic brain injuryIntervention: albuminComparator: any other colloid-containing fluids (dextrans, modified gelatins, hydroxyethyl starches) or isotonic crystalloid fluids (saline 0.9% and balanced salt solutions such as compound sodium lactate, Plasma-Lyte)Primary outcome: GOS or GOS-ESecondary outcomes: mortality, adverse events, hospital and ICU length of stayStudy design: systematic review
13	Population: adults with acute severe traumatic brain injuryIntervention: antiseizure prophylaxis (levetiracetam or phenytoin)>1 weekComparator: antiseizure prophylaxis <1 week or no antiseizure prophylaxisPrimary outcome: late post-traumatic seizureSecondary outcomes: GOS or GOS-E, mortality, adverse events, hospital and ICU length of stayStudy design: systematic review

CCHR, Canadian Computed Tomography Head Rule; CHIP, CT in head injury patients; CNS, central nervous system; CSF, cerebrospinal fluid; GOS, Glasgow Outcome Scale; GOS-E, Glasgow Outcome Scale-Extended; ICU, intensive care unit; NEXUS, National Emergency X-Radiography Utilisation Study; NOC, New Orleans Criteria; PICOS, population, intervention, comparator, outcome and study design.

We will use the Cochrane definition to identify systematic reviews. We will consider a review to be systematic if it clearly stated a set of objectives and reported explicit eligibility criteria, an extensive search strategy (a refined search strategy ran on Medline or Cochrane Library and at least one other database)29 30 and reproducible methods to identify, select and critically appraise the findings of the included studies.24

Outcomes

Primary and secondary outcomes were identified for each of the evaluated clinical practices by the project steering committee and are described in a PICO format in table 1.

Search strategy

In consultation with an information specialist, we will develop comprehensive literature search strategies separately for each clinical practice to be studied (see table 2 for a preliminary search strategy in PubMed). We will search systematic reviews using the Cochrane Library, Excerpta Medica Database (Embase), Epistemonikos,31 PubMed and the International Prospective Register of Systematic Reviews32 from 1990 to up to 6 months prior to submission for publication. Using a snowball approach, we will screen the references of included studies in addition to previous reviews on this subject.7 21–23 33

Table 2

Search strategy for hypothermia in PubMed

Concepts	PubMed search strategy	Research
Injury	“Craniocerebral Trauma”[Majr] OR [diffus* AND axonal injur*[Title/Abstract]] OR “head trauma”[Title/Abstract]OR “head injury”[Title/Abstract]OR “head injuries”[Title/Abstract]OR “brain trauma”[Title/Abstract]OR “brain injury”[Title/Abstract]OR “brain injuries”[Title/Abstract]OR “cerebral trauma”[Title/Abstract]OR “cerebral injury”[Title/Abstract]OR “cerebral injuries”[Title/Abstract]OR “craniocerebral trauma”[Title/Abstract]OR “craniocerebral injury”[Title/Abstract]OR “craniocerebral injuries”[Title/Abstract]OR “TBI”[Title/Abstract]OR “traumatic brain injury”[Title/Abstract]OR “traumatic brain injuries”[Title/Abstract]OR “brainstem trauma”[Title/Abstract]OR “brainstem injury”[Title/Abstract]OR “brainstem injuries”[Title/Abstract]OR “Head Injuries, Closed”[MeSH:NoExp]OR “Brain Injuries”[MeSH:NoExp]OR “Craniocerebral Trauma”[MeSH:NoExp]OR “Brain Hemorrhage, Traumatic”[MeSH] OR “Diffuse Axonal Injury”[MeSH:NoExp]OR “Coma, Post-Head Injury”[MeSH:NoExp]OR “Head Injuries, Penetrating”[MeSH:NoExp]OR “Intracranial Hemorrhage, Traumatic”[MeSH] OR “Skull Fractures”[MeSH]	#1
Clinical practice	“Hypothermia”[Mesh] OR “Cryotherapy”[Mesh] OR “Body Temperature”[Mesh] OR “artificial hibernation”[Title/Abstract]OR “body cooling”[Title/Abstract]OR cold*[Title/Abstract]OR cool*[Title/Abstract]OR “cooling therapy”[Title/Abstract]OR “cooling therapies”[Title/Abstract]OR cryogen*[Title/Abstract]OR cryother*[Title/Abstract]OR cryotreat*[Title/Abstract]OR hypotherm*[Title/Abstract]OR normotherm*[Title/Abstract]OR refrigeration*[Title/Abstract]OR temperature*[Title/Abstract]	#2
Filter for systematic reviews	(((systematic review[ti] OR systematic literature review[ti] OR systematic scoping review[ti] OR systematic narrative review[ti] OR systematic qualitative review[ti] OR systematic evidence review[ti] OR systematic quantitative review[ti] OR systematic meta-review[ti] OR systematic critical review[ti] OR systematic mixed studies review[ti] OR systematic mapping review[ti] OR systematic cochrane review[ti] OR systematic search and review[ti] OR systematic integrative review[ti]] NOT comment[pt] NOT [protocol[ti] OR protocols[ti]]) NOT MEDLINE [subset]) OR (Cochrane Database Syst Rev[ta] AND review[pt)] OR systematic review[pt]	#3
Total	#1 AND #2 AND #3	#4

Selection process

We will manage all citations with EndNote software V.X8.2 (Clarivate Analytics, 2014). We will identify and remove duplicates using electronic and manual screening.34 To ensure reliability when selecting studies for a given practice, two sets of 100 citations will independently be evaluated and then discussed by the reviewers. Pairs of reviewers (PAT, LM, IF, KMB) will then independently screen all identified records using titles, abstracts and full texts, consecutively. Any disagreement will be resolved through discussion between reviewers and, if necessary, consultation with a senior author (AFT). Potentially eligible studies excluded using full texts will be described in a PRISMA flow chart.

Data items and abstraction process

Using a standardised data abstraction form piloted on a representative sample of five studies, pairs of experienced reviewers (PAT, LM, IF, KMB) will independently extract the following data: first author, title, year of publication, databases used and date of the last search; population(s), intervention(s), comparator(s), outcome(s) and study designs included; measures of association and their respective measure of heterogeneity; tools used to assess the quality (risk of bias) of original studies and overall rating from the authors. Any disagreement will be resolved through discussion between reviewers and, if necessary, consultation with a senior author (AFT). When information is available in figures only, we will abstract graphical data using computer-assisted software.35 36 Furthermore, we will contact study authors (up to three email attempts) when information is unclear or unavailable.

Methodological quality assessment

Two reviewers (PAT, LM) will independently critically appraise the quality of systematic reviews using the Assessing the Methodological Quality of Systematic Reviews (AMSTAR-2) tool.37 Methodological quality will be categorised as low (0–3), medium (4–7) and high (8–11).

Synthesis

Results will be presented according to current recommendations for umbrella reviews.38 For each low-value practice, we will present the number of studies, study designs and patients included, the quality of the reviews (AMSTAR-2), effect sizes for primary and secondary outcomes (forest plots) and strength of recommendations (Grading of Recommendations, Assessment, Development and Evaluation).

Potential limitations

To ensure the feasibility of the review, we will restrict our search to low-value practices identified in the scoping review and expert consultation study, which may lead us to miss some low-value practices. However, given the robust search strategy used in our scoping review and the fact that experts were asked to add any other practices they considered low value, it is unlikely that important low-value practices have been missed. By targeting systematic reviews rather than original studies, we may miss some evidence. However, given the availability of high-quality, up-to-date reviews in TBI care suggested by our scoping review, we think it unlikely that we will miss a large body of evidence. For certain clinical practices, we may not identify any high-quality, up-to-date reviews. These practices will be the subject of systematic reviews in subsequent phases of the research programme. Finally, for feasibility reasons, we limited this umbrella review to reviews published in English since 1990 as per recommendations for umbrella reviews.25 26 These limitations should have negligible impact on results since few systematic reviews were published prior to 1990 and most published reviews are likely to be written in English.25 26

Potential impact

This review is part of the Canadian Program on Monitoring Low-Value Clinical Practices in Injury Care (Canadian Institutes of Health Research #113664), aiming to evaluate the effectiveness of an audit-feedback module targeting low-value clinical practices in acute injury care. The results of this review will be used to inform the development of quality indicators to be integrated in the audit-feedback module.

We will use state-of-the-art methods to optimise the sensitivity of our search strategy and the robustness of results. Results will be synthesised graphically. Ultimately, this research will inform the development of metrics, guidelines and deimplementation interventions, all targeting low-value injury care. The reduction of low-value clinical practices in acute TBI care has the potential to reduce pressure on strained healthcare budgets, free up resources, reduce adverse events and improve patient outcomes.

Ethics and dissemination

Ethics approval is not required as original data will not be collected. This study will be disseminated in a peer-reviewed journal, international scientific meetings, to knowledge users through clinical and healthcare quality associations (Choosing Wisely Canada, Trauma Association of Canada, American College of Surgeons—Committee on Trauma, International Federation of Emergency Medicine, Institut national d’excellence en santé et en services sociaux, Brain Trauma Foundation) and to patient partners associations (Brain Injury Canada).

Patient and public involvement

No patient or public representatives will be involved in this study.