Virtual systems enable personalised healthcare
Healthcare is an expensive and growing cost for countries around the world. Health costs account for around 10 % of GDP in OECD countries and are rising by 7-11 % each year, driven largely by ageing populations and chronic diseases. Against this backdrop, productivity in the sector has not increased. “Healthcare has not achieved productivity gains because it remains labour-intensive and has not adopted digitalisation or automation like other sectors over the last 30-40 years,” says Balázs Benyó, full professor in the Department of Control Engineering and Information Technology at Budapest University of Technology and Economics. Personalised, precision solutions in healthcare could help to alleviate some of the rising pressures, improving patient health – particularly in intensive care. In the DCPM(opens in new window) project, which was funded by the Marie Skłodowska-Curie Actions(opens in new window) programme, Benyó and his colleagues created virtual patient, digital twin models of the human metabolic, cardiovascular and pulmonary systems specifically designed to guide ICU clinical decision-making. “The goal is to enable individualised ICU care in these core areas,” explains Benyó.
Digital twins for personalised medicine
Digital twins are widely used in industry and scientific research as a way to model system behaviour and develop accurate predictions of how to optimise performance. “Our models serve this function by accurately transforming clinical data, which is often insufficient, to provide a clear physiological picture of the patient’s status,” adds Benyó. Personalised models must provide accurate predictions of specific responses to different treatment choices with relatively limited data. The solution developed in DCPM is a minimal modelling approach, capturing key dynamics which can be identified with the available data. These capabilities can then be used within clinical settings to personalise, optimise and automate care – reducing burden and cost, and increasing productivity and outcomes.
Validating models and implementing systems
DCPM’s main achievements include introducing a personalised glycaemic control treatment into the ICU at the University Hospital of Liège, Belgium. This has been spun out into a company, Insilicare, and is used in 7 Belgian ICUs, and in Malaysia, Hungary and New Zealand. The team validated a minimal cardiovascular model and its ability to accurately predict cardiac output and haemodynamics in septic shock. This is now entering its first validation-focused clinical trials. The researchers also developed and validated a multiscale lung mechanics modelling framework to guide invasive and spontaneous breathing mechanical ventilation, which is needed by up to 80 % of all ICU patients. “This is currently being tested in multiple forms in Belgium, China, Malaysia and New Zealand – and has been extended to out-of-hospital chronic respiratory disease applications,” notes Benyó. Other achievements include the development of: a blood glucose level measurement method using electromagnetic waves; neural network models predicting future variability of patients’ insulin sensitivity; a novel passive breathing sensor system; and multiple ultra-low-cost non-invasive devices for ICU and outpatient care, such as blood oxygen sensors.
Integration into European healthcare systems
Several of DCPM’s systems are entering clinical trials, while the pulmonary mechanics digital twins are moving towards uptake and adoption into care pathways. “All three areas are well along regulatory and first steps of commercialisation pathways,” says Benyó. Long-lasting research collaborations established in the project have also broadened the research into other areas, such as emotion recognition therapy for autism spectrum disorder and personalised nutrition delivery.
