Wisconsin’s Aurora Uses Teradata To Improve Patient Outcomes

Like other industries, health care initially used technology to automate time-consuming processes such as billing, registration and scheduling. 

“What organizations discovered is that, as an indirect result of operational systems, they had collected huge amounts of data on the business and how it performs” said Phil Loftus, chief information officer at the health care group. “We are shifting our focus from putting in operational systems to figuring out how to tap this wealth of data to make better business decisions and develop data driven improvements to the business.” 



The health care group is completing Smart Chart which Loftus described as probably one of the last giant $200 million operational systems that Aurora will deploy across the enterprise.  

About 10 years ago Aurora began building a data warehouse based on an Oracle database. 

“We started taking key pieces of data from our operational systems and putting it into a data warehouse,” said Loftus. “The original reason was to get nationally recognized measures of our clinical performance by scoring them against national standards and benchmarking the results through a third, which would tell us how we were performing and where we stood nationally.” 

That drove some major changes across Aurora. Once a hospital starts measuring the performance of its departments on a national scale, and sharing the results, physicians began to pay attention. 

“You don’t want to be in the fourth quartile; you want to be in the first, especially for physicians who tend to be competitive. This allowed them to see that  they could improve their performance by adopting clinical best practice.”   

Collecting and using its clinical data helped Aurora establish a national reputation for quality. The Center for Medicare and Medicaid Services (CMS) did a six-year demonstration project with  a number of major healthcare providers providing additional reimbursement, for high quality outcomes. At the end of the project, Aurora was named the best system in the U.S. and its Green Bay hospital was recognized as one of the top performers. Aurora works with the Premier Alliance, which benchmarks clinical quality data from hundreds of hospitals  allowing them to see their performance nationally. 

“Using these results, we have been able to build up hospital clinical quality performance in most major departments into the top quartile,” said Loftus. 

The route to higher quality has certainly been data intensive. Four years ago Aurora had a terabyte of data in its data warehouse; now is has over seven terabytes. 

“The amount of data has gone up by almost an order of magnitude. We are collecting 10 times as much data as we were 10 years ago,” he added. Aurora is using the data to look at effects of length of stay, complications and readmissions. 

“Not only is there an order of magnitude more data, but we want to ask more complex questions of the data. Using conventional technology, the cost would be prohibitive. We needed a different type of computer which was specially designed to process very large data sets quickly and cost effectively.  Aster (now part of Teradata) is one of the vendors that came up with computers which were specifically designed to analyze very large datasets and do it much more cost-effectively than regular computers.”  

The system from Aster uses massively parallel processing (MPP) — taking multiple inexpensive microprocessors and dividing the computing across them for fast results which Loftus estimated at 20 to 30 times the computing power of a traditional data warehouse. Aster was acquired by Teradata last spring. 

“Teradata positions us to accommodate the fact that our data will continue to grow and our questions become more and more complicated. We can look at our data and see if we are getting the optimal outcome for each of the diseases we treat.” 

As access to large data set analysis becomes possible, the nature of the analytics has changed from looking at individual patients to looking at populations of patients with problems such as congestive heart failure or diabetes, for instance.

“This allows us to increase clinical outcomes across a whole population of patients while lowering the overall cost – the holy grail of healthcare. We can look at the treatment of whole populations of diabetic patients, making sure that they meet hemoglobin Ha1c and cholesterol LDL targets that help control the progression of the disease and reduce the incidence of major complications such as retinopathy and impaired circulation.  Using these kind of approaches with our 1.2 million Aurora patient population has enabled us to achieve savings of 42 percent versus the Segal [a national health care cost survey] trend over the past 10 years. 

Health care providers are facing a revolution in the way they get paid — from payment for office visits, services, tests and hospitalizations to payment to meet the healthcare needs of a defined patient population at a fixed cost. 

“The move is from volume to value,” is how Loftus described it. An employer is going to pay a provider to look after a patient population at a fixed cost. About 42,000 Aurora employees are in a version of that model now.

“If we can find better, more effective and lower cost ways of looking after that population, we will have the opportunity to keep some of the savings.” 

Whatever the outcome of the current Supreme Court case, or the next Presidential election, the need to provide improved patient outcomes at lower overall cost will persist.  

One of the major goals in hospitals is to reduce the average length of stay for patients.  Aurora’s system wide length of stay is 3.8 days, compared to an expected average of 4.6 days, for a similar patient population. 

“We are also working to minimize the number of 30-day re-admissions,” he said. Pending federal reimbursement guidelines will require a hospital to take care of re-admissions within 30 days, for certain specific conditions, with no additional reimbursement. 

The Aurora computer system provides a detailed clinical history for patients. It is accessible to caregivers across the system to ensure that diagnoses and treatments are based on the best available patient information.  The system also includes 550 order sets based on clinical best practice that can be used as a starting point by physicians and modified to accommodate specific patient needs. This also gives patients the benefit from Aurora’s system-wide expertise at every facility. 

“One of the major changes that we are likely to see in healthcare in coming years is the growing ability to successfully treat patients in their own home,” added Loftus. The way the U.S. health care system evolved over the last 20 or 30 years, the system was built to funnel the patient into hospitals, because they were the higher margin part of the system. Now the shift is toward treating patients in less expensive and much more convenient settings such as clinics and the individual’s home. This is an area where monitoring and medicine delivery technology are already playing a growing role. 

Aurora has around 2,300 patients who are being treated at home by laptop-equipped nurses. They use their wireless data cards to tap into the patient records, giving them the same access to patient information as they would have in an Aurora hospital or clinic. For patients with chronic conditions, like congestive heart failure and diabetes, regular monitoring can be done at home and reported by nurses or remote monitors. Lofuts expects , the use of mobile technology, such as tablets and smart phones will grow rapidly. 

Aurora is preparing for outcome-based payments from the government, state and commercial payers, using a model known as accountable care organizations. Aurora is in talks with a number of employers to provide healthcare coverage for their employees at a fixed price. Already about 42,000 Aurora employees are in a similar outcomes-based model.

Even if the Supreme Court overturns the Obama administrations health care program, care providers will have to o move to similar models.  

Inevitably, said Loftus, that places healthcare providers in the risk business.

“In the past providers were reimbursed on the basis of payment for service didn’t take on financial risk.. In the future they will increasingly take on risk in the delivery of agreed upon outcomes for defined patient populations.  Risk is the element that you are now adding that wasn’t there in the past.” 

And as risk managers in finance know, managing risk requires lots of data plus sophisticated analytics.






About Tom Groenfeldt

I write - mostly about finance and technology, sometimes about art, occasionally about politics and the intersection of politics and economics. My work appears on Forbes.com and and occasionally in The American Banker and Banking Technology in London.
This entry was posted in Uncategorized and tagged , , , , , . Bookmark the permalink.

1 Response to Wisconsin’s Aurora Uses Teradata To Improve Patient Outcomes

  1. Pingback: Wisconsin's Aurora Uses Teradata To Improve Patient Outcomes … » BlinkMoth Software Industries | BlinkMoth Software Industries

Comments are closed.