Deep learning in a clinical context the big picture from “big data"
Clinical data is a source of large volumes of data: it is not uncommon to have orders of magnitude more variables than observations. The amount of collected information is continuing to grow as technology constantly pushes the boundaries of information that can be collected.
With such overwhelming volumes of data it is not a surprise that all of this data isn’t fully utilised throughout the clinical decision making process. For example, a critical patient admitted into the intensive care unit undergoes a barrage of tests during their stay generating a large amount of data points, but only single pieces of these are actually acted upon[i]. An alternative, and potentially much more powerful, approach would be to develop a tool that allowed clinicians to visualise this big data in an integrated platform and see the “big picture”, helping to guide clinical decision making, diagnosis and even indicate prognosis or future medical events.
Such tools may come in the form of artificial intelligence / machine learning[ii], something of a trending topic. The development of (un)supervised learning tools would offer the possibility of improving diagnosis, prognosis and best course of treatment, and would enable physicians to tailor their response in real-time. For years now machine learning has been successfully implemented in complex tasks by stock traders and programmers for data analysis and statistical models. However, successful applications for this technology in health and clinical medicine are still fairly limited[iii]
Work conducted during the course of my PhD will explore how machine learning tools can be applied to improve the clinical decision making process and potentially offer deeper insights to aid diagnosis, treatment and prognosis.
This work is currently ongoing and is a key part of my thesis looking at implementation of machine learning tools for “big data” integration, analysis and visualisation to improve clinical decision making.
[i] Sittig, D. F., Wright, A., Osheroff, J. A., Middleton, B., Teich, J. M., Ash, J. S., . . . Bates, D. W. (2008). Grand challenges in clinical decision support. Journal of biomedical informatics, 41(2), 387-392.
[ii] Chen, M., Hao, Y., Hwang, K., Wang, L., & Wang, L. (2017). Disease prediction by machine learning over big data from healthcare communities. Ieee Access, 5, 8869-8879.
[iii] Ng, A. (2016). What artificial intelligence can and can’t do right now. Harvard Business Review, 9.
Nathan Russell
A 1st year PhD Student at the University of Otago investigating how the clinical decision making process can be improved through implementation of machine learning tools for “big data” integration, analysis and visualisation. With a background in biology and clinical immunology my interests lie in what information can potentially be extracted from medical data as efficiently as possible which can be useful both at the bedside but also in biomedical research.