QuesGen Systems is pleased to continue working with CENTER-TBI on their large European Traumatic Brain Injury (TBI) Study aimed at improving the care for patients with Traumatic Brain Injury (TB)I injuries. The partnership began in 2013 and will continue through study completion. This study is led by Andrew Maas, MD, Ph.D. Emeritus Professor of Neurosurgery at Antwerp University Hospital and David Menon, MD, Ph.D, Professor of Anaesthesia at the University of Cambridge whose goal is to bring together many of the world’s leading TBI experts and the latest technologies in an effort to tackle the silent epidemic of TBI. This research initiative includes over 100 hospitals across Europe, China, India, and Australia.
The CENTER-TBI research study is developing objective, multi-modal classification schemes and outcome measures for traumatic brain injury (TBI). The objective of the study is to improve the characterization and classification of TBI with the inclusion of emerging technologies. The study is also hoping to identify the most effective clinical care using Comparative Effectiveness Research (CER) approaches. The project’s goal is to provide high-quality evidence in support of treatment recommendations and guidelines.
CENTER-TBI is an incredibly complex project that requires advanced data programming and data wrangling. It is a large collaborative project supported by the European Union Framework 7 program, that includes a multi-center, longitudinal, prospective, observational cohort study (Core study) with highly granular data collection across all severities of TBI. It also includes a Registry which collects basic administrative data.
CENTER-TBI faced many challenges due to the complexity of this study. The Core study was extremely challenging in terms of developing a user-friendly, intuitive e-CRF and in developing a dedicated data extraction tool (INCF-Neurobot) to facilitate analyses. The study was unique in its design to address TBI across all levels of severity. On enrollment, patients were differentiated by care pathway: ER stratum (discharged from the emergency room), Admission stratum (hospital ward), and ICU stratum (primary admission to the intensive care unit). This stratification posed substantial challenges to the integration of all variables relevant to the three strata into one e-CRF. The CENTER-TBI clinical dataset alone comprised a combination of over 2,500 distinct discrete and longitudinal measurement concepts, with the latter involving both regularly and irregularly sampled time points. Furthermore, the dataset skip-logic design was stratum-specific.
“While many data-collection options were available, we opted for QuesGen because of their in-depth experience with TBI studies over the last 10 years including the TRACK-TBI Pilot. QuesGen had implemented the system in the definitive TRACK-TBI study.” – Dr. Andrew Maas, Emeritus Professor of Neurosurgery at Antwerp University Hospital
Working with QuesGen
QuesGen designed and built a system where basic data validation/range checking was implemented in the eCRF. Completion status was also logged allowing sites to easily identify subjects for which various data elements were incomplete or required attention. Data element descriptions were included in the eCRF so that it acted as a standard operating procedure for the sites. Raw data from the eCRFs was stored on a centralized, secure relational database located within the European Union for data governance considerations. This raw data was transferred to a second relational database in order to store the ‘clean’ data. Re-mapping of some of the database tables as well as the transformation of time-series clinical data from eCRF ‘wide’ format to a ‘long’ format more suitable for analysis was conducted using documented pipelines. Data in the clean database was de-identified: All dates were zeroed to the Unix epoch and site-identifiers anonymized. Where free text entry had been permitted, it was checked both by hand and by automated algorithms to ensure that identifiable information was not present.
“This complex study model required partnering with a company that could provide a data collection system with a dedicated study specific e-CRF with stratum-specific workflows which QuesGen Systems was able to provide.” – David Menon, MD, Ph.D, Professor of Anaesthesia at the University of Cambridge
CENTER-TBI has provided valuable lessons for undertaking multinational and trans-Atlantic collaborative studies – both in general and for TBI in particular – in the following areas:
Data curation: A great deal of effort – much greater than anticipated – went into the curation (cleaning) of data. This work was performed by a multidisciplinary data curation team (DCTF) with expertise in clinical data science, eCRF design, statistics, and outcomes as well as study management domains. The data curation team examined data not only for missingness and plausibility but also for multivariate consistency by cross-checking variables with other related concepts in the database. The data curation team was sizeable and international, making communication challenging. A formal ticketing system was used to log issues and their resolution.
“QuesGen personnel contributed substantially to the curation process. The QuesGen team provided experience and expertise that was invaluable to the CENTER-TBI team and the overall study.” – Dr. Andrew Maas, Emeritus Professor of Neurosurgery at Antwerp University Hospital
Data quality issues: Throughout the study data quality problems were identified, then investigated to identify whether these were structural issues (e.g. variances in data type that was unanticipated at design time), site-specific issues (e.g. unanticipated variances in data element interpretation due to local or language-related misinterpretation) or simply isolated random errors. Data quality problems were addressed in three broad ways. Firstly, for a small number of systematic data entry inconsistencies, it was possible to transform data or unify concepts across time-points and documented pipelines were created to do this. Range and consistency checks were updated in the eCRF if the original choices were found to be too permissive. Secondly, where systematic issues were identified, there was a robust process involving a dedicated team to go back to sites to understand and identify problems and implement solutions (including, for example, process-validation at source or ongoing training needs analysis). Finally, where common but non-systematic errors were identified, eCRF rules were updated and the subjects re-flagged as being incomplete so that sites could go back and make corrections. The entire process of data curation required nearly one year of work by the 18 members of the DCTF. The result is a high-quality dataset of immense size, that is further linked to imaging, biomarker and genetic repositories.
Harmonization across InTBIR studies: Joint analysis teams from CENTER-TBI and TRACK-TBI are now embarking on the first cross-study analyses, and as with data curation, it seems clear that harmonization may be more resource intensive than had been hoped, analysis across studies will still be possible, though federated approaches to analysis may be needed. Some of the challenges are inherent to differences in the databases between the two studies, but differences in image storage and processing and regulatory barriers also contribute.
Results have not yet been published as the total duration of CENTER-TBI including full analysis of all results is 6.5 years with an expected end date of April 2020. Ultimately, the result will be a high-quality dataset of immense size, that is further linked to imaging, biomarker, and genetic repositories. Analysis of CENTER-TBI has just started and the first results are anticipated over the next few months. You can expect and look forward to many findings and publications around this exciting project.
“QuesGen offered a shorter start-up period and better opportunities for harmonization between US and EU partners.” – David Menon, MD, Ph.D, Professor of Anaesthesia at the University of Cambridge
Assessments included in the CENTER-TBI study: GOAT, GOSE, TMT, CRS-R, RAVLT, RPQ, SF-12V2, QoLIBRIOS, PHQ9, SF-36, GAD-7, TUG, QoLIBRI, 10m Walk