As an organization participating in the federal Pioneer accountable care organization (ACO) program, Allina Health, based in Minneapolis, has a natural incentive to improve clinical performance. Even so, the decision on the part of senior leaders at the 13-hospital, 90-clinic, 26,000-employee Allina to commit to achieving exceptional outcomes—senior executives strove to be in the top decile of performance on every meaningful, measurable quality metric—was ambitious enough to give the project a semifinalist award in this year’s Healthcare Informatics Innovator Awards program.
In population health, where no one single metric exists, the organization would demonstrate that it could measurably improve a health index measure. Further, in terms of affordability, the organization would be at or close to mid-market cost for the patients it serves, according to Allina Health officials. Indeed, Allina Health set out to make this commitment by 2016. To move in that direction, Allina adopted a multi-pronged strategy that encompassed new payment models and contracts, care redesign, and technology and analytics.
Specifically, Allina team members aggregated clinical, financial, operational, patient satisfaction, and other data into its enterprise-wide data warehouse (EDW), to create consistent views of the data. They then decided to use analytics to identify the clinical programs in greatest need of optimization and waste reduction, focusing initially on congestive heart failure (CHF) readmissions, spine care, length of stay, and venous thromboembolism (VTE) care. They chose to drill down particularly intensively on reducing CHF-related 30-day readmissions.
The leaders who created that EDW are the founding members of Salt Lake City, Utah-based data warehousing and analytics vendor Health Catalyst, a company that “aggregated a lot of information from a lot of different sources, allowing for data to be pulled out of Epic and put it in place so our leaders could use it clinically and drive improvement,” says Tim Sielaff, M.D., Ph.D., chief medical officer of Allina Health. “We have been doing this for a long time. Our organization is organized around clinical service lines such as oncology, cardiology, and orthopedics, so care processes and improvements of care processes driven by data is baked into our genes,” Sielaff says.
Tim Sielaff, M.D., Ph.D.
When Health Catalyst the company was formed and the data warehouse that Allina was running independently became bigger and more complicated, the cultural fit to connect the organizations was obvious, Sielaff notes. “It was a way of helping to exponentially accurate the function of our data resources within the organization,” he says. Simply put, Sielaff says the goal is for Allina to serve as a resource for the rest of the community and country. “As we learn how to do things well, such as develop dashboards and improvement projects, we have a vehicle to share those with society now,” he says. “The ability for us to use data to do improvement was something we have been doing, and the collaboration with Health Catalyst has made it better by that much, made it that much more rapid cycle, and that much more rigorous based on what we might have been able to do on our own.”
Sielaff notes that data by itself is not very useful; the goal, he says, is to turn data into information and information into knowledge, and from there enact improvements. “Our organization is designed around care process improvement, and that is always driven by data that is in the hands of people close to the work—clinicians in an OR, catheterization lab, or clinic, they have that information and knowledge available to them to help create improvement,” he says.
The way Allina does improvement originates oftentimes out of identifying variation in care, Sielaff continues. “Where there is variation that is not warranted by the needs and preferences of the people we serve—our patients—that’s where we start working on improvement. Any organization that starts looking at variation within any part of an organization, will find it. The ability of putting that data into the hands of clinicians so they can determine if it’s warranted, relative to our patients or not, and then take that from there, that’s the sauce. Trust and transparency are the two key tactics that allow us to enact improvement.”
Sielaff offers examples of finding variations in care and then driving improvement. It can be anything from the timeliness between having an abnormal diagnostic mammogram and a biopsy to bleeding rates post-percutaneous coronary intervention to the cost and value of delivering a spinal fusion operation, he explains. “There are literally hundreds of examples of using data to drive improvement, which is almost always less expensive. Better care is linked to less expensive care, and that is value. The value equation is outcomes divided by cost, and I say [multiplied by] appropriateness. And value doesn’t mean cheap,” he says.
Taking it to the Next Level
In January, Allina Health announced that it signed a 10-year agreement valued at $100 million to combine analytics technology, clinical content and personnel with Health Catalyst. As Healthcare Informatics Senior Contributing Editor David Raths reported at the time of the announcement, “The Allina employees currently working in data warehousing, analytics and performance improvement technology will become onsite Health Catalyst team members in phases beginning [January]. (Health Catalyst has committed to retain all of the former Allina Health staff.) Also, Allina will gain access to Health Catalyst's full technology, content and deployment expertise to accelerate outcomes improvement at Allina. The partnership's governing committee will annually identify a prioritized list of improvement projects, each designed to provide measurable care improvement and financial value to Allina, and as success is realized, both partners will share in the economic benefits.”
Further, Allina said it would become a "living laboratory" and national showcase for outcomes improvement, featuring the latest developments in analytics-enabled improvement from Health Catalyst's portfolio. Each year, the committee governing the partnership will agree to a prioritized list of data-driven improvement projects with specific, measurable outcomes goals for each project. Economic rewards will be based on the attainment of these goals, according to Raths’ report.
Sielaff adds that the partnership allows Allina to “take its internal resources that it grew locally and connect them with a whole other world of really smart people who are connected with multiple other health systems. It was felt that we would really be able to accelerate our improvement opportunities, the rigor and complexity of the analysis we can do, and benchmarking across multiple sites.” Health Catalyst has a similar long-term, shared risk relationship with Partners Healthcare in Boston, and just recently announced one with Tacoma, Wash.-based MultiCare Health System.
As such, Allina has now linked all of its cost information to clinical outcomes to administrative data and lab data, and those multiple aggregation sources allows its leaders to make the analyses and conclusions more rigorous and robust, and hopefully more accurate and correct, Sielaff says. “We have the ability to link cost profiles to our individual surgeons across homogenous groups of patients. This allows us to help understand what best practices are, and where there might be opportunities for cost savings and improved care processes. The variation that we see in cost and quality and outcomes of delivering care are much more transparent and accurate, by virtue of the aggregation of that data via Health Catalyst. The accuracy, rigor and actionability of that information has gone up significantly,” he says.
The results for the health system have been remarkable. In the readmissions arena, Allina clinicians and care managers have achieved a 17-percent 30-day CHF readmissions rate reduction in 10 of the 11 hospitals doing cardiac care, along with a 30-percent increase in patient participation in the system’s heart failure program, a 36-percent reduction in CHF patient ED visits, and a 20-percent reduction in length of stay for CHF patients. In the spine care arena, the health system has been involved in a full initiative to deliver care that supports the Triple Aim in the treatment of spine disorders and lower back pain. Among other achievements, Allina has implemented a collaborative spine care coordination program; achieved a 16-pecent reduction in length of stay for spine care procedures; achieved a 36-percent reduction in post-operative complications; and achieved a projected $2.7 million in savings through supply standardization. Allina leaders continue to work forward across the CHF, spine care, and VTE clinical areas, and continue to show ongoing significant clinical and financial outcomes progress, its officials say.
Sielaff additionally notes that readmissions work is not one single thing, but rather a complicated problem that represents the continuum of care from the patient being in the hospital to getting back in the community. “So if you’re transitioning from your hospital bed down to radiology, there is a risk of something happening during that transition—the same as leaving the hospital to going home. It’s about making sure you have the resources to help that individual when he or she is back out in the community,” he explains.
Allina’s data analytics allows its leaders to identify those people who are high risk of readmission. An air traffic control dashboard tells them who is at the highest risk of readmission, and allows the clinical team to deploy special resources to them, Sielaff says. “It gives us the ability to see that when someone is in the hospital, or even better, before he or she comes to the hospital. That’s driven by our ability to identify the folks who need that additional attention and provide that. It’s a combination of analytics and clinical care,” he says.