Ambulance Victoria uses analytics and modeling to serve the expanding needs of a growing population.
by Aaron Lazenby, November 2013
Since 1883, the ambulance services that operate in Victoria, Australia, have addressed the emergency medical needs of the country’s most southeastern mainland state. Over the years, those services have consolidated into one entity—Ambulance Victoria—answering calls for help from the metropolis of Melbourne and the surrounding rural territory, serving 227,000 square kilometers and 5.5 million people. How has an organization that started with a single horse-drawn ambulance evolved to earn a global reputation for superior patient outcomes?
“We’re an organization with a proud history of using data very widely to improve the welfare and survivability of our patients,” says John Dousset, manager of enterprise architecture at Ambulance Victoria. For example, paramedics are armed with special notebooks to capture empirical data about each patient at the time of treatment. And Dousset’s team at Ambulance Victoria has embarked on an ambitious program to use data collection and analytics to ensure more patients get the care they need.
Here Dousset talks to Profit about how Ambulance Victoria has evolved, the innovative ways his coworkers are using data to visualize solutions, and what technology he is watching now.
Profit: How has Ambulance Victoria changed in recent years?
Dousset: Ambulance Victoria has a long history of serving the community, and has done so with a lot of support from the community. In 2008, Ambulance Victoria was created as a single entity by merging the state’s three remaining ambulance services: Metropolitan Ambulance Service, Rural Ambulance Victoria, and the Alexandra District Ambulance Service. We now serve the entire state with 250 branches, 3,000 paramedics, and 1,000 volunteers, with four fixed-wing and five helicopter aircraft for fast connections between rural communities and major specialist facilities in the metropolitan region.
On average, we’ve seen about 4 percent of annual growth in demand for our services since 2000. This is due to factors such as population growth, an aging population, increasing numbers of people living alone, increased outpatient services and early discharges, and limitations to other health services in rural settings. Ambulance Victoria received approximately 830,000 requests for assistance in 2011–2012, which is fairly substantial. But we can’t simply increase resources to service that growing demand, so we have had to figure out how to do more with what we have.
Profit: How does technology help you do that?
Dousset: To be more efficient with our resources, we knew we needed to do two things. First, we needed to improve our operational decision support in real time: Where are our resources? Which resource should go where? Second, we needed to work on the business side: What is our profit, what is our loss, what is our margin? How can we improve our transport costs?
Last year Ambulance Victoria had a running deficit of about AU$3.3 million, but the organization still undertook a major investment in Oracle analytics and reporting. Part of the reason upper management embraced this plan was that I could demonstrate how we could reuse the same technology stack to enable multiple business processes. We didn’t need to buy a unique technology stack just to support different business processes.
We invested in Oracle technology and were able to use one platform to launch a major analytics and reporting program, centered around three business objectives. The first was to introduce advanced resource modeling: What would happen if we increased the number of paramedics? What would happen if we increased the number of cars? The second was centered on workforce planning, because an aging population also means an aging workforce for us. The third was centered purely on financial modeling, including margin models and models about product cost. This is critical because with the creation of Ambulance Victoria we picked up multiple systems with multiple business processes and variable definitions of our data.
Profit: What is your strategy for modeling planning scenarios?
Dousset: Modeling is one of those things people get very passionate about. I know people who do modeling in Excel, and yes, that’s a form of modeling. But I would call it fairly labor intensive and a very flat model.
Instead, I use “What if?” modeling. So, for example, if I were working on a resource model, I would ask, “What’s the impact on my resource demand for the next five years, if I introduce new scenarios or variables?” And I would take raw data out of our finance system, find those spreadsheets used in our finance department, pull them into a wire framework in Oracle Essbase, and allow people to develop many and varied scenario-based “what-ifs.” Once the wire framework is developed in Oracle Essbase, fresh data and multiple scenarios can be reloaded and altered in a very timely manner—just minutes after they get the data. Then we can give those results to senior executives to assist with planning and budgeting via standard or custom Oracle business intelligence reports.
Profit: How can analytics help managers or paramedics in the field do their jobs better?
Dousset: There are probably half a dozen key events that occur in the treatment of a patient, starting from the time of an emergency call. For example, when our paramedics accept a call, they push a button on an in-vehicle mobile data terminal [MDT] device. That is an event that arrives in our real-time data feed along with a geospatial coordinate that gives us the ambulance’s location. Every five seconds we receive an update on the vehicle location’s coordinate, which tells us that an ambulance is en route from point A (say, the scene of an accident) to point B (presumably a hospital).
Right now, we have a proof of concept going that would allow hospital workers to see how many ambulances are scheduled to arrive—and when they should expect them. We are also starting to figure out how to publish data, such as what resources are stationed at the hospital and which hospitals are full or becoming full. When we get that information to our hospital emergency departments and our own team leaders, then they can start to make some real-time decisions to improve the overall resource management.
Analytics and reporting should be simple enough for everyone to use to assist in their daily
Business intelligence should not be something that’s away in the corner of the office and difficult to consume. Really, it’s about having the correct information starting point and the right level of visualization to fulfill a business need—such as real-time operational decision support, reporting for our CEO and board, and financial modeling for the CFO. Analytics and reporting should be simple enough for everyone to use to assist in their daily tasks. Where possible, users should be enabled to drill up and down the data presented.
Profit: How does data support the management of the organization?
Dousset: In my department, we have two big monitors that are out for everyone to see. One shows real-life, real-time data about where our resources are. My CEO can bring around members of the board and show them, “Here’s this hospital. Five ambulances are currently en route, and five ambulances have been at the hospital for the last hour.”
The other monitor allows users to drill down on information from a summary financial view. They can see a simple bar graph at the executive level and drill all the way down to a specific transaction. One of the changes I have made is to move to a model of exception-based reporting. If the report is green, that means we are meeting all our goals, so don’t do anything. If the report is red, drill down and find out what needs to be fixed.
Profit: What kind of scenarios would make a report go red?
Dousset: It could be anything ranging from ambulances at a hospital for longer than 40 minutes to a patient in a nonthreatening situation who has been waiting for attention for more than an hour. It’s really quite limitless as to what the exception can be. That’s going to be one of our challenges, I think, and one I really want to tackle: What are our KPIs [key performance indicators], and what are the alerts we put into reports? Right now a lot of what we have is driven on our staff’s experience and the problems they’ve identified.
Profit: What else are you excited about, looking forward?
Dousset: Looking downstream, one of the things that really interests me is figuring out how to leverage Twitter and other social media tools. Are there things we can access and blend into our analytics and reporting? Social media could also be useful in providing context for major events and helping us plan an appropriate response.
Also there’s the innovation side of it: Can I share data with other emergency service organizations this way? I’m keen for B2B integration. What if I can send members of other organizations an XML feed in real time, guaranteed messaging, with a common message model? Then they send me something back that I can surface through a tablet or a device in the field. Lo and behold, our paramedics can see a fire line, they can see where other resources are—like police officers and other paramedics—and they are provided more information to assist in their in-field decisions.
In the end, Ambulance Victoria is a vital member of the overall healthcare system that saves lives. It is about resource management and providing the best care possible to the community in the most efficient and effective manner. We’ll always be looking for innovation and technology to help keep Ambulance Victoria as one of the leading ambulance services in the world.
Aaron Lazenby is editor in chief of Profit.