Dr Laura Moss, R&D Healthcare Computer Scientist at NHS Greater Glasgow & Clyde, and Pamela Brankin, Head of Marketing and Communications at Aridhia Informatics, tell Kathryn Reilly the exciting story of their collaboration and its potential for more effective treatment for patients with Traumatic Brain Injury (TBI)
Brain injury is devastating – not only for those affected but also for their family, carers and everyone who supports patients during their often long and arduous recovery. Unlike many other diseases and injuries, Traumatic Brain Injury (TBI) has no proven effective medical therapy. However, recent advances in Big Data modelling, computation and analytics offer the hope of a new approach, one where collaboration between technology and life science industries can determine better treatment pathways and deliver cost-effective, sustainable healthcare.
To make treatment more effective, high-frequency neurointensive care data produced by patients suffering from TBI can be linked to lower-frequency clinical data to enable the detection and prediction of clinically relevant events. However, due to the data’s volume, size and often a lack of suitable computing resources, these analyses have been difficult to perform and complete in clinically meaningful timescales. In this critical care environment, providing the results of analyses available at a patient’s bedside presents technological challenges for healthcare providers in data interpretation, analysis and integration. Now a unique and innovative collaboration is working towards solving these challenges.
CHART-ADAPT is an Innovate UK-funded collaboration between Aridhia Informatics, NHS Greater Glasgow & Clyde, Philips Healthcare and the University of Glasgow and was formed to develop a platform that enables important physiological models and analyses to be rapidly implemented into clinical practice. The platform extracts high-frequency physiological TBI patient data in order to analyse and return the results directly to the bedside, informing clinical decision-making and, ultimately, improving practice.
Healthcare has only recently taken its first steps towards unlocking the potential of Big Data in critical care. Intensive care is a very technology-led area of practice. Consequently, large volumes of patient data can be generated, for example, a typical Neuro Intensive Care Unit (neuro ICU) may record more that 72 million time points of data per average patient stay.
To ensure the patient receives the best possible care, continuous monitoring of patient physiology is fundamental. Most modern TBI centres now have complex bedside monitoring devices that are capable of collecting and storing large amounts of data. These include vital signs such as: arterial blood pressure and brain pressure, medications (including the type, dose and rate of delivery), ventilator data and laboratory results.
While huge volumes of such precise high-frequency data are available, which could be used to predict future changes in physiology and guide care, it is often only summarised and saved in single ‘snapshots’, resulting in important information being lost to future analysis when no longer clinically required.
Further challenges arise when attempting to process and use this data in a clinical setting due to a lack of data-science skills and inadequate computing infrastructures to perform high-resolution data collection and analysis. This process is also time-consuming: it can often take days or even weeks before results can be produced. This is detrimental to the patient’s chances of recovery and prolongs their expensive neuro ICU stay.
With a number of challenges to overcome in turning this data into clinically useful information, a new approach is clearly required. In Europe, TBI is the most common cause of permanent disability in people under the age of 40, and the cost of managing these injuries exceeds €100 billion annually. The need is great.
TBI can produce a multitude of pathophysiologies requiring a range of different treatments – brain injuries have a number of causes including car accidents, falls, assaults, strokes etc – therefore clinical staff need as much detailed information as possible about both the patient’s documented past medical history and their current clinical status to provide the personalised care they require.
This can prove difficult. In Glasgow, for example, there are usually around 50 cases of severe TBI presented each year. This means that only a few cases of each type are seen annually and, in such small numbers, it can be difficult to characterise the full underlying pathology and thus to propose appropriate therapy. To fully understand the pathology, a larger databank is required to allow care providers to see which therapies work well for individual injury types, and to take a more targeted patient-specific approach.
The Connecting Healthcare and Research Through a Data Analysis Provisioning Technology (CHART-ADAPT) project addresses these challenges through a technology that rapidly delivers clinical decision support at the bedside. The collaboration draws on the expertise of a multidisciplinary team of scientists and clinicians led by Dr Laura Moss, an R&D Healthcare Computer Scientist with NHS Greater Glasgow & Clyde’s Department of Clinical Physics and Bioengineering. The University of Glasgow’s School of Medicine, Dentistry & Nursing offers clinical input to the project, while Philips Healthcare’s expertise in providing bedside monitoring and a robust data integration infrastructure provides the patient data for analysis.
The CHART-ADAPT framework receives patient data from this monitoring equipment, de-identifies it and then transfers it to Cloud computing services for configurable Big Data analytics. From there, the results from the analyses are re-identified and presented back into the clinical environment.
Scottish-based clinical informatics company, Aridhia, is an essential part of this collaboration. Its data analytics platform, AnalytiXagility, allows the project to analyse live streams of vast amounts of physiological and clinical data collected in a real-time clinical setting, translating it into clinically relevant information that is fed back to the patient monitoring equipment at the bedside to enable more rapid and better informed treatment decisions to be made by clinical staff. Aridhia, which has a base in the University of Glasgow’s Innovation Space at the Queen Elizabeth University Hospital, has also created an app to enable clinical teams to select different algorithms to run on a particular patient’s data to inform the best course of individual care, and supporting a rapid, personalised approach to treatment.
An earlier pilot project saw Aridhia and NHS Greater Glasgow & Clyde reduce the time taken to process sample data from patients with TBI from 16 hours to 48 minutes, clearly demonstrating the advantages of using advanced analytics in a clinical setting. The next phase of CHART-ADAPT aims to improve on this already significant achievement.
Preliminary testing of the infrastructure has begun at the Institute of Neurological Sciences, Glasgow, and live data collection has commenced. Once validated, the framework can be customised for use by intensive care units elsewhere.
The only way to analyse such complex data is to take a multidisciplinary approach, bringing clinicians, the medical device industry, programmers and data scientists together. It’s impossible for any one group to do this alone. With CHART-ADAPT we have – collaboratively – made major strides in developing software and hardware solutions that can extract, standardise and de-identify data from the intensive care environment, and created a platform to process the data in clinically meaningful timescales. The real value of this platform will be demonstrated upon seeing patient prognosis improving and treatment patterns evolve as this data-driven innovation has an impact in the real clinical environment.
You're the expert! Write for The Engine or share your articles, papers and researchAdd your content
Add your content
Sign up for Ignition, our regular, ideas-packed newsletter