Common Data Elements: Critical Assessment of Harmonization between Current Multi-Center Traumatic Brain Injury Studies

Abstract

Standardization and harmonization of data collection in studies on traumatic brain injury (TBI) is of paramount importance for meta-analyses across studies. Nearly 10 years ago, the first set of Common Data Elements for TBI (TBI-CDEs v1) were introduced to achieve these goals. The TBI-CDEs version 2 were developed in 2012 to broaden the approach to all ages, injury severity, and phases of recovery. We aimed to quantify the degree of harmonization of these data elements in three large, prospective multi-center studies conducted within the International Initiative for TBI Research (InTBIR). Data variables of the Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI; adult and pediatric patients in Europe and Israel), Transforming Research and Clinical Knowledge in Traumatic Brain Injury (TRACK-TBI; adult and pediatric patients in the U.S.), and Approaches and Decisions in Acute Pediatric TBI (ADAPT; international study on severe pediatric TBI) studies were indexed and matched to the second version of the TBI CDEs. We focused on the CDE sub-categories of ''Acute Hospitalized'' (AH) and ''Moderate/Severe TBI: Rehabilitation (Rehab). All ''Core'' and ''Basic'' level CDEs were considered. Closely related elements were reduced to one variable to prevent over-representation. Categorical elements and text elements for the same variable were likewise merged to one element for analysis. Following reduction and merging of related elements, 21 Core, 46 Basic AH, and 50 Basic Rehab elements were deemed harmonizable across studies. Gaps in global applicability were identified for four of the TBI CDEs and many of the outcome instruments, which are only available in the English language. Agreements of Core and Basic study CDEs for the AH domain with the TBI CDEs were respectively 81% and 91% for CENTER-TBI, 76% and 93% for TRACK-TBI, and 85% in ADAPT for both domains. For the domain Rehab, agreement with Basic TBI CDEs was 84% for CENTER-TBI, 94% for TRACK-TBI, and 71% for ADAPT. Non-harmonization was largely caused by absence of the elements in the studies. For elements present, the compatibility of coding with TBI CDEs was 90-99%. The degree of harmonization was greatest between CENTER-TBI and TRACK-TBI with 81-87% overlap within the TBI CDE sub-categories. The high degree of harmonization of study variables among these studies demonstrates the importance and utility of common data elements in TBI research. It also confirms the potential for future meta-analyses across these large studies, especially for CENTER TBI and TRACK TBI. The global applicability of the TBI CDEs needs to be improved for them to become a global standard for TBI research. CENTER-TBI, TRACK-TBI, and ADAPT, along with other studies within the InTBIR Initiative, provide a platform to inform further refinement and internationalization for the next version of the TBI CDEs.

Addressing Key Clinical Care and Clinical Research Needs in Severe Pediatric Traumatic Brain Injury: Perspectives From a Focused International Conference

“…Reliable quantification of the burden of TBI is difficult to achieve due to inadequate standardization of data definitions and severity classification. Common data elements have only recently begun to be standardized and harmonized in pediatric TBI ( 11 ). Approximately 80% of pTBI cases can be classified as mild (Glasgow Coma Scale [GCS] score ≥13) with negative imaging findings and these cases are often seen by family practitioners; thus, epidemiological estimation based on hospital admission or emergency department (ED) visits underestimates the true TBI incidence as much as 4–5-fold ( 12 ).…”

Promoting global collaboration for brain health research

“…16 On the clinical side we have seen the value of harmonisation of variables among relevant studies to promote greater comparability across collaborating research projects. 17 Machine learning and artificial intelligence techniques based on big data are increasingly being used in both understanding and diagnosis of neurological disorders and offer a new model for personalised management. Machine learning techniques could be used to delineate the categories and predict the patients' outcomes with various conditions.…”

“… 16 On the clinical side we have seen the value of harmonisation of variables among relevant studies to promote greater comparability across collaborating research projects. 17 …”

“…Thanks to the international collaborative projects, the field of functional neuroimaging has advanced substantially, showing the value of big data science 16. On the clinical side we have seen the value of harmonisation of variables among relevant studies to promote greater comparability across collaborating research projects 17…”

Evidence Limitations in Determining Sexually Dimorphic Outcomes in Pediatric Post-Traumatic Hypopituitarism and the Path Forward

“…CDEs for adult TBI have recently been defined (Version 1.0) and are being developed for pediatric TBI as well (143)(144)(145). Already, multicenter studies enrolling TBI patients worldwide (e.g., ADAPT, TRACK-TBI) (143,(146)(147)(148) are collecting core datasets which allow for more direct comparisons of study differences, analyses of patients across studies, and the assembly of larger, common study populations. This presents an opportunity for neuroendocrinerelated CDEs to be developed and included for pediatric and adult TBI.…”