Traumatic brain injury (TBI) occurs when a sudden trauma causes damage to the brain. It is a major health problem and the most common cause of permanent disability in people under the age of 40 years. Yearly cost from TBI in Europe exceeds 100 billion Euros. Recent statistics show a steep increase in the incidence of TBIs, with an increase of 21 % over the last five years, threefold greater than the rate of increase in population. Despite this, TBI has been seriously underrepresented in medical R&D efforts compared to many other, less significant health problems.

The TBIcare project provides an objective and evidence-based solution for management of TBI by improving diagnostics and treatment decisions for an individual patient. A strictly evidence-based approach, it realises the objectives of developing:

• a methodology for finding efficient combinations of multi-modal biomarkers used in statistical models to objectively diagnose and assess an individual TBI patient, and
• a simulation model for objectively predicting the outcome of the planned treatment of an individual TBI patient.

These objectives are supplemented by the realisation of: a software solution to be used in daily practice to diagnose and plan treatments; new approaches for extracting information from multi-source and multi-scale physiological databases for management of an extremely heterogeneous disease; and innovative data quantification methods for the clinical TBI environment. Thus, TBIcare transfers the scientific Virtual Physiological Human (VPH) concepts to clinical practice.

Policy Context

TBICare is an EU co-funded in FP7 under the VPH framework project and is co-ordinated by VTT Technical Research Centre of Finland. The consortium includes GE Healthcare Ltd. (UK) and GE Healthcare Finland Oy with the participation of 5 more organisations from 4 European countries: Finland, United Kingdom, Lithuania and France.

Description of target users and groups

Healthcare professionals treating patients with TBI.

Description of the way to implement the initiative

The project has two scientific objectives that are supplemented by three technical objectives. Together they define how the work in the project is carried out.

• Scientific Objective 1: Development of a methodology for finding efficient combinations of multi-modal biomarkers in statistical models to objectively diagnose and assess an individual TBI patient

The first objective is addressed by using an approach in which a high number of vital signs or biomarkers, relevant to TBIs, are explored from sets of heterogeneous data. These include, for example, structural and functional changes visible in imaging data (computerised tomography, CT; magnetic resonance imaging, MRI; positron emission imaging, PET), changes in electrophysiology (electroencephalography, EEG); changes in bedside multimodality monitoring parameters including systemic cardiac and respiratory physiology, intracranial pressure (ICP), and brain chemistry (monitored by oxygen sensors and microdialysis); and changes in metabolomics visible in the blood. We define sets of biomarkers from several thousand brain injury cases retrospectively, and from several hundred TBI cases and healthy controls prospectively. The goal is to build statistical models allowing standardised and objective interpretation of data from a single patient. The diagnostic rules are derived by comparing the patient data to the most similar cases in a database using statistical inference.

• Scientific Objective 2: Development of a simulation model based for objectively predicting outcome of the planned treatment of an individual TBI patient.

Work towards this objective uses the aforementioned statistical models as basis for the construction of a simulation model. Due to the unique responses to treatments, the simulation model must be individualised. The model is personalised for each patient separately using data only from similar cases. Various approaches can be used for the simulations, such as, concepts from system dynamics or Bayesian networks. In the TBIcare concept individual physiological measures and various treatments form the building blocks of the system dynamics model which is used to predict the outcome.

The simulation model provides important information both for scientists and clinical practitioners. It helps a scientist to better understand a human as a system – a viewpoint central to the Virtual Physiological Human. A clinician is able to test the influence of various treatments by first simulating them. As the variability of the individuals and traumas is huge, we do not expect that a simulation model built from hundreds of cases is enough for reliable prediction of the outcome. However, our aim is to develop a strictly evidence based simulation model for objectively predicting the outcome of treatment and rehabilitation of an individual TBI patient. The model provides objective evidence based information about the most probable outcome and will be a step towards scientifically valid approach for the treatment planning. This model will be a basis for future development, where an increasing amount of validated clinical data will continuously improve the reliability and usability of the model. In addition, this kind of model may be used to optimise the diagnostic procedure in TBIs, e.g. it may advise the clinician to take some further tests in order to improve the reliability of the model for a certain individual.

These scientific objectives are supplemented by realisation of three technical objectives:

• Technical Objective 1: a software solution to be used in daily practice to diagnose and plan treatments;
• Technical Objective 2: new approaches for extracting information from multi-source and multi-scale physiological databases for management of an extremely heterogeneous disease;
• Technical Objective 3: development of innovative data quantification methods for the clinical TBI environment.

By reaching these objectives, TBIcare transfers the scientific Virtual Physiological Human (VPH) concepts to clinical practice.

Technology solution

The first scientific objective is addressed by using an approach in which a high number of vital signs or biomarkers, relevant to TBIs, are explored from sets of heterogeneous data.

These include, for example:

• structural and functional changes visible in imaging data (computerised tomography, CT; magnetic resonance imaging, MRI; positron emission imaging, PET),
• changes in electrophysiology (electroencephalography, EEG and electro-impedance spectroscopy, EIS);
• changes in bedside multimodality monitoring parameters including systemic cardiac and respiratory physiology, intracranial pressure (ICP), and brain chemistry (monitored by oxygen sensors); and
• changes in metabolomics visible in the blood, as well as
• genomic and proteomic information.

Biomarkers are available from several thousand brain injury cases retrospectively, and from several hundred TBI cases and healthy controls prospectively. Statistical models using these data allow standardised and objective interpretation of data from a single patient. The diagnostic rules are derived by comparing the patient data to the most similar cases in a database using statistical inference.

Work towards reaching the second objective uses these statistical models as basis for the construction of simulation models. Various approaches can be used for the simulations, such as concepts from system dynamics or Bayesian networks. In the TBIcare concept, individual physiological measures and various treatments form the building blocks of the system dynamics model that is used to predict the outcome. The simulation model provides important information both for scientists and clinical practitioners. It helps a scientist to better understand a human as a system – a viewpoint central to the Virtual Physiological Human, and a clinician is able to test the influence of various treatments by first simulating them. As the variability of the individuals and traumas is huge, we do not expect that a simulation model built from hundreds of cases is enough for reliable prediction of the outcome, and fully implementing prediction for an individual patient will not be feasible within the relatively short  time of this project, but rather a goal for continuing research. However, the model provides objective evidence based information about the most probable outcome and will be a step towards scientifically valid approach for the treatment planning. This model will be a basis for future development, where an increasing amount of validated clinical data will continuously improve the reliability and usability of the model.

A second modelling application can be used to optimise the diagnostic procedures in TBIs in the context of cost-effectiveness and socio-economic analyses, taking into account treatment, rehabilitation and prevention. The model evaluates the impacts of potential TBI prevention policies leading to reduction of TBI’s incidence rate caused by traffic accidents or alcohol to name but a few. Moreover, the model can be used to evaluate the significance of different policies concerning TBIs treatment and diagnosis. Policy improvements may result from improved competence of healthcare personnel, or computer aided decision support and diagnostic tools, for example. Integrating the acute treatment and rehabilitation of TBI is also a key focus of the modelling process. The hypothesis is that the earlier the rehabilitation starts after TBI the better the results, which in turn, leads to improved socio economic impacts.

For a final implementation as decision support to be taken in use in the hospital setting, obvious interoperability with the existing Hospital Information System and structure is needed. While we aim to realise that within the project where reasonably possible – it is not the main aim per se of this research project, but rather something that will be the subject of further exploitation efforts.

Technology choice: Proprietary technology, Standards-based technology

Main results, benefits and impacts

The main targets of TBIcare – improved diagnostics and treatment selection – are especially in the interests of those health professionals who treat patients at emergency departments and in-hospital wards, including intensive care units. Depending on the local settings, these may include at least traumatologists, neurosurgeons, (neuro)anaesthesiologists, and neurologists. As diagnostics and treatment decisions are the responsibility of physicians, the specialists in these fields are those with the greatest potential to adopt our results in everyday work.

While eventual implementation of outputs from TBIcare in these different settings will require individual validation, it should be noted that in the course of the TBIcare project we can only achieve a preliminary analysis. A clear prerequisite for successful marketing of these applications is to prove excellent reliability and achieve recognition by medical opinion leaders, as well as to show proven benefit for the final user to use such a system. In this sense, our post-project plans are subject to future changes, in case the current vision changes due to the outcome of later clinical trials. For all the project’s outcomes that are to be exploited further, a period of at least two years following the end of the project is expected to be necessary for additional development, certification and clinical evaluation at a large scale. The marketing phase will vary depending on the scenario in which exploitation is being pursued.  Thus, use in academic research will realistically be a much earlier target than use in clinical care. 

Research in TBIcare is broad and hence there are several potential areas for exploitation of project results. In summary, the objective is to bring new information technology approaches to the bedside, and through a digitisation of human physiology, deliver tools that of real value in clinical care and research. The research spans several disciplines and there are many potential areas for exploitation of TBIcare project results. These include: biomarker profiles involving metabolomics, proteomic and genomic analyses; algorithms for EEG processing as well as tools for EIS data analyses; fully automatic non-invasive ICP measurements; and image processing. The software decision support tool is, obviously, another major target for exploitation. System dynamics modelling in healthcare is a novel and less well-established concept and will require more work to allow proper validation of its performance in the long-run.

The future of health care is to have an integrated and personalised approach to “personalised diagnosis -> personalised treatment -> personalised follow-up and maintenance”. This requires a multidisciplinary approach where competencies from biology, biophysics, medicine, image processing, chemistry etc. are needed, and where different actors (different health professionals, patients and their (formal and informal) caregivers, service providers, and policy makers) interact closely. TBIcare will deliver the tools and the setting to further implement and expand such an integrated approach upon.


Return on investment

During the project duration itself there will not be returns on investment. Once the tool is in use, several years after the project, savings in healthcare costs may be foreseen in the form of savings on direct healthcare costs (resource usage in hospitals) and increase in quality of life of patients. Potentially this runs in the Millions of Euros, but is difficult to estimate.

Track record of sharing

We are in continuous contact with collaborative research efforts, and have close co-operation with them (Tampere Hospital, TRACK-TBI consortium, IMPACT consortium, INCF). Additionally new large research efforts have started in which members of the consortium play a main role (CENTER-TBI). We are very open to exchange of data. Furthermore we are organising a public symposium in January 2014 (http://congress.utu.fi/TTBIS2014/), meant for interaction and setting up of further co-operation.

Lessons learnt

• Collecting highly complex multi-modal physiological databases from different centres is a complex task that requires good co-ordination but also a lot of flexibility and ability to deal with many surprises. In the end the result is extremely valuable though.
• From a data-analysis point-of-view, the task of helping to improve diagnosis in TBI is an extremely complex problem, with a huge amount of different variables and extremely diverse personalised patient profiles.
• Even the smallest progress in the abovementioned points is very worthwhile though, as the issue being worked upon is a major source of costs to the healthcare system.

Scope: International