Computational modelling allows researchers to simulate and study complex systems – including disease – at multiple levels, powered by significant achievements in computing power and software. Dr Jacob Barhak is an independent Computational Disease Modeller. He draws on his multidisciplinary expertise to help machines comprehend healthcare. Read More
The extensively published (and patented) Reference Model for Disease Progression was developed by Dr Barhak in 2012 initially for application to diabetes.
Dr Barhak observed that while there were many models of this chronic disease, not all were good representations (or ‘fit’) of the real-world clinical data – that is, what actually happens with patients. His approach to this problem was to create a league of disease models and validate these using publicly available results from clinical trials, allowing a test of how well the models and assumptions fit the existing data.
Importantly, the Reference Model is an ensemble model, meaning that it can assemble lots of different models and compare them, assessing their credibility. As such, it is also a ‘fitness engine’ in which the models (and combinations of these) are ranked according to how they fit the real-world data. As further information is added to the Reference Model, it accumulates this additional knowledge meaning that it is constantly evolving.
…
At the start of the COVID-19 pandemic in 2020, Dr Barhak saw the potential of the Reference Model to improve our understanding of this highly infectious disease and adapted his previous work to create the first multi-scale ensemble model for COVID-19. As the name suggests, it uses multiple models describing cells/organs, individuals, and populations.
The COVID-19 model incorporates models of infectiousness – the level of infectiousness of each individual from the time of infection. The models of transmission reflect the probability of contracting the disease through interactions with infected individuals. The models of response reflect the behavioural choices of each individual affecting the number of their interactions with others in light of the pandemic. Several mortality models incorporate the probability of dying from COVID-19 by age, and time of death since infection (in days). The recovery model defines the condition of recovery. Finally, observation models are implemented to correct for the fact that the reported numbers are not always accurate.
Dr Barhak is using publicly available data and testing how these can be explained using the Reference Model. To date, the disease is not yet fully explained and his ongoing work is dedicated to refining and improving the approach to better explain the existing COVID-19 data. If successful, the Reference Model would provide a valuable tool to better comprehend new pandemics and more effectively inform public health interventions.
