Using analytics from Qlik and predictive modeling from DataRobot, University Hospitals of Morecambe Bay NHS Foundation Trust in England is striving to change the culture of healthcare.
Driven by data, UHMB, based in Kendal in the north of England near the Lake District, is moving from merely delivering healthcare -- treating patients who come through hospital doors in need of medical attention -- to health optimization.
Using predictive modeling, the organization is attempting to determine not only demand for medical attention, such as a probable spike in surgeries post-COVID-19 given all the delayed medical care occurring during the pandemic, but also identify risk in individual patients for particular maladies such as hypertension that could lead to stroke.
"We've got a massive opportunity to fundamentally transform healthcare," said Rob O'Neill, head of analytics at UHMB, during a breakout session during QlikWorld 2021, Qlik's virtual user conference.
Six weeks before the conference, O'Neill's father-in-law died of a stroke. Walking in the hallway between his living room and kitchen after watching the evening news, he collapsed and never regained consciousness.
Hypertension is the biggest risk factor for stroke and affects about 25% of the world's population. Hypertension, however, is difficult to diagnose, and about 40% of those who are hypertensive are undiagnosed, according to O'Neill.
Using data, however, healthcare organizations can predict it.
Urgent care, meanwhile, is a consistent source of concern for healthcare organizations.
According to O'Neill, 50% of the emergency rooms in the U.K. received a substandard rating. Demand for ER beds has been increasing in recent years -- even before the pandemic -- yet to reduce costs hospitals have been reducing their supply.
With data, however, hospitals can better match supply and demand in the ER, and improve the quality of patient care.
The biggest issue facing healthcare organizations, of course, remains COVID-19.
Over the past 15 months, hospitals have been inundated with patients suffering the often deadly effects of the coronavirus. As vaccines becomes more available, however, case volume is declining in many countries. That means hospitals will soon be able to treat many of the patients who have had non-life-threatening conditions but been unable to get treatment.
Rob O'NeillHead of analytics, University Hospitals of Morecambe Bay NHS Foundation Trust
Analytics can be the means to enable healthcare organizations to manage the backlogged patients who will be seeking care once the pandemic ends.
"We can use data, specifically autonomous machine learning models integrated into our Qlik Sense dashboards, to transform the way health systems model the flow of patients," O'Neill said. "Traditional approaches to healthcare business planning are no longer enough. In fact, they've never been enough."
Qlik and DataRobot have been key to UHMB's transformation.
Taking data from the many disparate sources that produce healthcare data, the organization uses Snowflake to warehouse its data, break down data siloes and transform it for analytic consumption. With Qlik and DataRobot both able to connect to Snowflake, UHMB is able to extract data into Qlik for its business intelligence needs and DataRobot for predictive modeling.
The healthcare provider is then able to embed the models from DataRobot into Qlik dashboards for access and visualization in a familiar environment alongside related analytics assets.
"Data are growing quickly in terms of volume, variety and velocity, and we need to be able to harness this data really quickly and integrate them directly into our analytics platform," O'Neill said. "The exciting thing is that we can create a cycle, flowing between our elastic cloud data platform, our AI models and our analytics solution."
And that cycle is now the foundation for UHMB's decision-making process, according to Kelly Heys, analytics and data science manager at UHMB.
"Being able to forecast activity [with DataRobot] and then integrate it into our existing Qlik Sense dashboards has allowed us to seamlessly introduce predictive analytics into our reporting," she said. "That's a fundamental building block for us as an organization so we can be more proactive when dealing with demand for our services rather than reactive."
Blueprint for success
While UHMB is using Qlik for its analytics needs and DataRobot to do predictive modeling, many organizations -- both in healthcare and other industries -- continue to resist data-driven decision-making.
Data-driven decision-making, however, is what has enabled many organizations to survive the pandemic and is what will enable them to continue to succeed as analytics becomes the norm rather than the exception.
To begin changing the culture of healthcare organizations and make data the driver behind all decisions, O'Neill developed a six-step blueprint:
- lead by being the one to articulate a vision that includes changing the culture and strategy and includes a delivery plan;
- excite the organization by involving leadership in the development of a strategy and deliver quick the successes that will lead to more investment in the strategy;
- innovate by understanding what the problems are and delivering successful modeling solutions;
- disrupt by being prepared to be the person that challenges existing processes and delivering new solutions with analytics;
- persist even when met with resistance; and
- repeat by finding opportunities, building successful analytic solutions, implementing them and moving on to the next one.
"Healthcare affects us all," O'Neill said. "It affects us personally, and those around us. We're at the tipping where the democratization of predictive analytics within healthcare is a reality. And as the globe begins to navigate its way out of the COVID-19 pandemic, data-driven healthcare delivery is an absolutely essential component of the recovery and restoration process."