Hospitals hold valuable data trapped in separate systems, from patient records to administrative databases. Considering this, implementing a ‘data fabric’ could change healthcare decision-making by giving a comprehensive view of their operations.
This unified approach could help hospitals better predict patient demands, manage resources, and improve care delivery in an increasingly complex healthcare environment, said Michael Cronin, co-founder and managing director of digital transformation specialist OpenSky Data Systems.
“The way to explain it is, it’s a set of technologies which allow seamless integration of data, regardless of whether it is on premise or in the cloud, or hybrid in both,” he said.
Although data fabrics will be of benefit to any organisation, hospitals and other healthcare providers, with their wealth of data, can benefit particularly significantly.
“A hospital is a more complex environment than most, but what you’re working with, whether it’s in a hospital or an SME, is, you have data from lots of different sources.
"People are making decisions based on information from specialist systems, so the problem that brings for leaders is, when I look at
data from my HR system, from my finance systems and service delivery systems, I am looking at data that isn’t interrelated. The result is that you don’t get the full picture,” he said.
Implementing a data fabric, which collates data from a variety of sources, changes this.
“One of the primary uses of a data fabric is it allows you to use the data that you have,” Cronin said.
Naturally, not having access to the data makes it hard to make decisions.
“If you’re looking at separate systems, let’s say admissions have gone up but you don’t know why. You could be making decisions not knowing why that was the case,” he said.
However, someone who is looking at a dashboard built on top of a data fabric will have the ability to look at multiple data sources together, such as, for instance, public data reporting that says flu cases have gone up. Meanwhile, real-time data on patient wait times, bed occupancy, and staff availability can be used, along with historical information and epidemiological information, to optimise resource allocation.
“Hospitals already have this data, it’s all sitting there in specialised systems, but it’s not being brought together to give them what they need,” said Cronin. “There is work to do, as the data isn’t going to merge itself.”
Indeed, the range of data a healthcare provider will want to look at is vast, including administrative systems such as finance, billing in the case of private hospitals, human resources, supply chains, asset management, electronic health records and specialised software in clinical areas including radiology, medical imaging and more.
Hospitals already have this data, it’s all sitting there in specialised systems, but it’s not being brought together to give them what they need
Consequently, building a data fabric is a form of digital transformation and this is where OpenSky works, starting with a consultative process that defines the goals and the scope of existing systems and data sources, then working to implement data classification as a key part of the strategy.
“The work isn’t in building the report, the visualisation; it’s about making the data integrable, and that’s done by labelling data and the other data engineering work that we do,” he said.
Machine learning (ML) can also be used to augment this data classification process.
Indeed, ML and other artificial intelligence (AI) technologies offer much to healthcare, but here again a data fabric is essential as it ensures regulatory compliance, said Cronin.
“We always say that your data fabric is your first step as it brings in governance and labelling, so you have sensitivity labelling, for instance. Then you can bring in access controls about who can see what, and then, if you are running AI on top of that, you have the controls in place such as tracking where your data is coming from, how it has changed, who changed it, and how long you keep it for,” he said.