Using data to gain new insights to improve healthcare delivery for over 750,000 citizens.
As a public service accountable to the 750,000 citizens of New Brunswick as both patients and taxpayers, the Department of Health is always looking for ways to provide better service at a lower cost.
Think about the last time you visited a hospital or a doctor’s office.
From checking in to heading home, you interacted with a number of people, systems, services, and maybe even pharmaceuticals. While thanks to our incredible healthcare system in Canada you never saw a bill for any of it, there are associated costs behind the scenes.
In New Brunswick, the Department of Health keeps very precise records of all costs incurred while delivering healthcare to New Brunswickers both in and out of province. This helps forecast and budget at a macro level. But the Department of Health wanted to get even more value out of this data.
They wanted to be able to more completely understand what health care services individuals were receiving and build profiles of their top patients based on age, gender and diagnoses to inform health care delivery decisions.
The Department of Health selected NP Digital to apply our proven processes to turn this complex data challenge into an opportunity to better understand healthcare delivery.
To empower the Department of Health’s analysts with the ability to really understand where health care dollars are consumed and the profiles of those consumers, many types of data would need to be securely aggregated in the new system. It goes without saying that any solution needed to safeguard patient privacy.
With this in mind, the data science team set about preparing the Department of Health’s data for analysis. This meant following the team’s proven process for gaining insights and creating what’s known as a “data lake,” a single data repository that would securely bring together an aggregate view of the costs and services provided by the province and its hospitals and physicians, as well as hospitals and physicians outside of New Brunswick.
Once the data lake was created, a new schema was designed to allow analysts to develop custom filters to create a “single view of the patient” that is both comprehensive and accurate.
Because of the variety of the data, privacy requirements, and the expectation that analysts need to be able to ask and answer complex queries very quickly, the system was built using Hortonworks Hadoop, an open source, industry leading data science platform.
With the proof of concept now deployed and in use by the Department of Health’s analysts, results have been exceptional.
Thanks to the data lake, patient analyses schema, and a friendly yet powerful user interface, the Department of Health’s analysts are now able to ask difficult questions of their data and quickly discover answers.
Specifically, cost and patient profile analyses are performed much faster with much less duplication of effort. The new system is powerful and flexible, allowing for deeper investigation and understanding of what’s driving costs.
The new system is also infinitely scalable. This initial project was a proof of concept linking a limited number of siloed data sets. Being able to integrate any new data sets, the Hortonworks powered data lake is primed to become even more useful as other types of data are added for analysis and inspection.
Finally, all of this data has primed the Department of Health to get started with predictive analytics. By comparing past and real-time data, this will allow for even better interventions to create and deliver world-class healthcare.