Big data analytics has provided insight into the fields of cancer research and public health, but still faces computing and data-sharing barriers
This article forms part of a series on big data in healthcare.
Research activities have generated more scientific data in the last few years than at any other time in history. This data, in combination with information contained in electronic health records, has the potential to generate new knowledge to advance health research. Automated analysis of structured and unstructured data is currently being used by many researchers to create new hypotheses, discover unknown scientific relationships, and gain a fresh perspective into diseases.
Genomics research, an area of research which sequences and analyzes whole-organism genomes, has benefitted greatly from big data analytics (BDA). In the field of cancer, the need for faster, less generalized, and more cost-effective methods to translate research into targeted solutions for patients has triggered researchers to examine the disease at the genomic level. Researchers have used BDA to understand the genomic diversity of breast cancer by sequencing and analyzing cancer genomes into 10 breast cancer classification groups with distinct clinical outcomes and possible underlying biology. Genomic predictive analysis has also helped facilitate the discovery of new candidate disease biomarkers, or indicators for pathology, for translational research across many areas of research.
In addition to genomics, BDA also has applications in public health research. BDA has been used to compare air quality data from polluted areas to health care data on respiratory disease. Epidemiologists have used BDA to gather information on the spread of disease by examining social and sexual networks, and government and clinical databases are being analyzed using BDA to find cost-effective treatments for a variety of conditions to support health policy decision-making.
However, there are numerous barriers to using big data analytics for health research. Insufficient computing power and BDA adoption pose challenges, but a more compelling barrier is the lack of willingness to share data. Many researchers, institutions, and organizations protect their data fiercely, a very traditional stance which stems from complex factors like competition, funding, and intellectual property. This unfortunately poses a problem, as any type of research that uses BDA benefits from access to data, its integration, and collaboration between healthcare sectors. In an effort to facilitate the sharing of data, the Montreal Neurological Institute has recently implemented an ‘open science’ initiative to make all research data and analyses available at the time of publication to other researchers and organizations. This in turn will hopefully accelerate the translation of research into new treatments for patients and serve as a model for others to follow suit.
Dawson SJ, Rueda OM, Aparicio S, and Caldas C. A new genome-driven integrated classification of breast cancer and its implications. The EMBO Journal, 32, 617–628, 2013.
Murdoch, TB and Detsky, AS. The inevitable application of big data to healthcare. Journal of the American Medical Association, 309(13): 1351-1352, 2013.
Noor AM, Holmberg L, Gillett C, and Grigoriadis A. Big Data: the challenge for small research groups in the era of cancer genomics. British Journal of Cancer, 113(10):1405-12, 2015.
Montreal Neurological Institute Open-Science Initiative announcement:
Big Data Analytics in Health White Paper by Canada Health Infoway:
Written by Fiona Wong, PhD