What is the ROI of investing in a healthcare data analyst? A real example of saving millions
A few years ago, a health system came to us and said, “We need to build an observation wing.” In the past, there had been an observation wing on the campus, but when changes in patient demographics and demand no longer justified the wing, it was ultimately re-purposed to meet growing emergency department (ED) needs.
But then an interesting trend emerged after the health system re-purposed the wing. Observation patient volumes steadily rose, while Medicare reimbursements for the ED population continued to decline. As a result of the new trend, the health system’s leaders were looking for ways to optimize clinical resource allocation for their ED observation patients.
Using a healthcare-specific EDW to make an informed decision
The health system already had a fully functional enterprise data warehouse (EDW) in place. By digging deep into its insights with our help, their leadership expected to make data-informed, strategic decisions about the details of the observation wing.
As we sat down to discuss their requirements for the new observation wing, we also asked them what answers they expected to get from their EDW. They specifically wanted the following three questions answered: (1) How many observation beds are needed for the wing? (2) What clinical staffing will be needed for the new wing? (3) What will it cost to build the new wing?
Leadership seemed be asking good questions — the kind of questions that are typical for this type of capital improvement project. And we had answers.
How we would leverage data to answer leadership’s questions:
Question 1: How many observation beds are needed for the wing?
Answer: To answer this question, we would need to leverage clinical data from the EMR source mart to identify trends for patients classified as observation patient types.
Question 2: What clinical staffing will be needed for the new wing?
Answer: Calculating the clinical staffing needed for the new wing would require a combination of clinical data as well as human resources data to build a staff model using historical patient volumes.
Question3: What will it cost to build the new wing?
Answer: We would need to mine costing data in order to forecast the cost to build the new wing. Incidentally, costing data suggested that to convert space to an observation room would be $.5 million to $1 million per bed to repurpose an existing bed. So estimates for a 5-bed observation wing were between $2.5 million to $5 million.
Whoa — taking a step back to ask the right questions
As we began looking into the data that could provide answers to their questions, we took a step back and put the work on pause. We then asked the health system if the decision to proceed with the build had already been made. If it was a foregone conclusion, that would lead us to pursue one course of action, namely to get the answers to their questions. If however, the decision had not been made, but in fact could be influenced by what was learned by mining the data, we would suggest a different course of action.
A teaching moment
This became a teaching moment for the client partner. They told us the decision to build had not been made and they would welcome further analysis to better inform the decision-making process. With that acknowledgement, we began looking for clues in the data that might lead us to understand the root cause for increased observation patient volumes.
Clues discovered from the EDW’s data:
Clue 1: Roughly two out of three observation patients had a complaint of some chest pain. This is a very common complaint for patients presenting to the ED that end up as observation patients, so this was not a big clue. However, the fact that so many of the observation patients had the same concern was a clue.
Clue 2: Nearly every patient that presented in observation had been a previous patient. This was revealing because it meant the increased observation volumes did not represent a material increase in new patient volumes.
Clue 3: Of those patients that complained of chest pain, four out of five patients had been previously diagnosed with heart failure from the cardiology clinic. This was a BIG clue. It meant these patients were aware of cardiology care available through the clinic, yet they were presenting at the ED. Why?
Clue 4: For those observation patients with chest pain, three out of four arrived between five and 10 p.m. Why? What was significant about this time frame? Why would most of the patients cluster around this five-hour window? One quick phone call to the cardiology clinic answered the lingering question — the clinic closed at 5 p.m.
Booyah! The answer was right in front of us. The data led to the understanding that because the cardiology clinic closed at 5 p.m., patients feeling chest pain went to the next available known care location to get attention: the emergency department.
Data-driven recommendations to achieve the best ROI
The analyst leading the data mining effort recommended the following:
1) Extend the cardiology clinic hours until 10 p.m.
2) Don’t spend the $2.5 million to $5 million for an observation wing.
The health system was thrilled. So were we. Think about this: if the analyst on this project were paid $75,000 a year as a base salary, the value of this decision to avoid unnecessary significant capital spend justified her or her role for 30 to 60 years.
And that is serious ROI from the data warehouse.
Making the most of your data analyst
In today’s value-based care environment, where margins are shrinking, it is critical that health systems get the best ROI possible from their improvement projects. A data-driven approach with an EDW is a good place to start. But it is just as important to ask the right questions and empower your data analysts to use data to really make a difference. Organizations that take this approach to their improvement projects will be most successful at making the change in this new world of value-based care.
I’d love to hear from you. Have you realized the best ROI possible from your healthcare improvement projects? If not, do you have questions about how to realize these types of gains from your analysts?
Would you like to use or share these concepts? Download this quality improvement presentation highlighting the key main points.