An industry vertical analysis: Healthcare and big data
Big data is such a broad enterprise business resource that we can only speculate the impacts on industrial verticals over the next years. We do know that industrial sectors that are building the “industrial internet” of networked devices, grids, transportation network and electronic records are the ones to be systemically transformed. Through this series of blogs we will uncover each impacted industrial sector with associated big data facets, opportunities and limitations. This post will focus on big data and healthcare.
Traditionally, healthcare has been analytically impaired for one reason or the other. The sector is historically data disadvantaged with very little ready access to electronic records. One of the key strategies of the World Health Organization is to modernize healthcare across developed and under-developed geographies to adopt “electronic medical records” for transactions. Even these efforts will leave us with 50 percent-unstructured texts in electronic medical record data. This also adds to the low analytic adoption.
Another crucial element to examine is supply chain moderation. The sector in general has specific point of care requirements to be handled at each local level. The similarity is often very minimal across local units, let alone regional and national levels. Analytic models have to operate at the point of requirement, based on healthcare symptoms and gaps at the specific localised requirements.
For example, sub-Saharan Africa hosts approximately 70 percent of the HIV affected population globally. For governments to provision healthcare services, the supply chain has to be modelled for the local area to level healthcare gaps and accurately predict growth. This really is a faceted big data analysis. IBM Big Data Bluemix Cloud provisioning can span a wide technological solution, bringing disparate regions to the centralized platform.
Any clinical setting also brings in voluminous amounts of textual data in the form of notes from physicians and nurses. Natural Language Processing and IBM BigInsights Text Analytics play a critical role here as well.
Ecosystem integrations around health care provider data with insurance medical claims data is equally vital. IBM Big Data strategy around avoiding moving data from source and rather federating for analysis plays a vital role here. Storage and analysis of CAT scans, MRI’s and Human Genome Data spanning terabytes per human are other massive sources of big data information.
The ongoing liberalization of health care services by governments also brings in telemedicine, remote monitoring and more networked sensor devices capturing healthcare information at frequencies pacing from seconds to minutes. IBM Infosphere Streams can be the technological backbone for monitoring sensor data in critical care units.
The challenge of big data in the healthcare sector is now really around making use of data. Correlation and What-If analysis of big data will transform the sector by providing extraordinary insights into treatment and saving lives.
The next part of this series will cover the telecommunications sector, now being defined by the sheer number of communication devices, which are approaching big data levels. Ever wondered why this sector treats you the same in spite of being a small or a big spender? Follow me on twitter and stay tuned to the Hub to find out why.