Authenticated user menu

Search
Resource
23 November 2015
Total views

Big Data in healthcare

By Dr Lauren Megahey, ADEPT Clinical Leadership Fellow (ST4 in Psychiatry of Old Age)

What exactly is ‘Big Data’ and how is this relevant to health professionals?

These are two of the questions I hoped to get answers to at the ‘Big Data and healthcare’ workshop at the FMLM Northern Ireland Regional Conference in Belfast last week, delivered by Dr Michaela Black, Senior Lecturer in Computer Science at the University of Ulster.

What is big data?

We were advised that what Big Data is can depend on who you ask! This is not a single technology, technique or initiative. It refers to technologies and initiatives that involve data too diverse, fast-changing or massive for conventional technologies, skills and infrastructure to address efficiently. In short, this is data that is too vast for humans to comprehend without the assistance of computer models. It can also be viewed as the science of handling data to get actionable learning.

Dr Black explained this very well using examples of large companies such as search engines, shopping websites and supermarkets. For example, when online, we notice that advertisements can be tailored to items we have recently been looking at on other websites, and when we receive supermarket vouchers these are usually for items that we have recently purchased. In fact, data about all of us is continually being generated and stored, enabling companies to provide personalised services and targeted advertising.

How can big data be used in healthcare?

Big Data can be very useful in healthcare, and this opportunity is not currently being fully utilised. It could be very beneficial to be able to translate the idea of personalised customer service into healthcare. The fact that much healthcare data is longitudinal gives added wealth. 

Ninety percent of the work in Big Data involves the actual collection of data. There can be many difficulties around this, eg who currently has access to data and are they willing to share it? We need to know which sources to get data from and how to verify their reliability. 

Once data has been collected, it needs to be put into a format that people are comfortable with. The data needs to be managed – for instance ‘cleaned’, or joined-up if it’s fractured data. Analytics can then be used in a number of ways, eg for visualisation; to provide descriptive models of what is happening such as clustering or showing patterns; to help develop decision models; for predicting disease progression; or to develop change models, using knowledge about disease over time or across demographics.

The use of Big Data in healthcare has huge potential. Data is being collected at an incredible rate, with ninety percent of the world’s data being created in the past two years! Healthcare data is one of the fastest growing segments of the digital world, increasing at a rate of about 50 per cent per year. Dr Black illustrated with some examples already being used in healthcare, such as a tool called CordMatch, which uses Big Data techniques and a unique matching algorithm to quickly find cord blood matches for cancer patients in need of a stem cell transplant.

The opportunities and limitations

Participants of the session were also asked to think about and discuss questions such as: what data do you currently have easy access to; what do you think you could learn from this; what additional data could you get access to; how would this change from descriptive to predictive or prescriptive?

Dr Black rotated around each group helping them to think about these questions. One group suggested using data available from maternity systems to aim to reduce the rate of caesarean sections, by grouping patients who have caesarean sections into clusters, looking at which groups have the highest rates and then specifically aiming interventions to those groups. 

It was acknowledged that there may be some ethical issues around the use of data. There was discussion around this, with a suggestion to ask patients if they agree to anonymised data being used and shared.

Overall this was a very useful workshop. With so much data available, it makes sense to utilise it to improve the planning and delivery of healthcare. With the rapid expansion of the amount of data available this will become an increasingly important area.

Array ( [0] => sitewide [1] => advert_external_leaderboard [2] => not_front_desktop [3] => advert_external_wideskyscraper [4] => attachments [5] => node-social-accelerators [6] => node_resource [7] => related_content [8] => member_attachments_for_non_members [9] => advert_internal_desktop )