Prime Healthcare’s BPCI Success: Leveraging Data for Continuous Performance Improvement

At a time in U.S. healthcare history when leveraging data for continuous clinical and financial performance improvement has never been more important, the leaders of some organizations are already demonstrating concrete results from having invested early in data analytics for performance improvement.

One of those moving quickly along in the journey is the Ontario, California-based Prime Healthcare, a healthcare management company that owns and operates 45 acute care hospitals in Alabama, California, Florida, Georgia, Indiana, Kansas, Michigan, Missouri, Nevada, New Jersey, Ohio, Pennsylvania, Rhode Island, and Texas. As explained on its website, “Founded in 2001, Prime Healthcare has emerged as one of the nation’s leading hospital systems. Prime Healthcare was founded by Dr. Prem Reddy, a physician with a mission to save hospitals, improve community healthcare and provide the very best care to patients. Our philosophy is that all healthcare is local. Each of our hospitals,” the website explains, “serve the unique needs of their communities while providing exceptional care with the strength and support of Prime Healthcare. Through a physician-directed and patient-centered model, passionate dedication to quality, evidence-based best practices, operational efficiency, technology and capital investment, our hospitals are recognized among the best in the nation.”

The leaders of Prime Healthcare have been partnering with the Marlborough, Massachusetts-based Persivia, which on March 17 announced in a press release posted to the company’s website “the launch of a new COVID-19 surveillance module within its SolitonTM AI engine,” with that functionality to be made available to Persivia’s install base of “20 million covered lives across the country.” As an established customer organization, Prime Healthcare will be able to take advantage of that technology, even as its long-term partnership with that solutions provider has already resulted in ongoing advances in clinical performance management. With regard to Prime Healthcare’s data-driven aspirations, Ahmad Imran, M.D., vice president of value-based care at Prime Healthcare, spoke last week with Healthcare Innovation Editor-in-Chief Mark Hagland, regarding the organization’s current initiatives. Below are excerpts from that interview.

Tell me about your organization’s use of data right now, in the context of the COVID-19 pandemic, and more broadly?

We are a data-driven company, and are very heavily focused on performance and outcomes. We’ve been recognized by IBM Watson Health three times as a top health system in the country, and every year, we have hospitals recognized as among the Top 100 Hospitals in the Country; six of our hospitals were recognized this last year. We want to deliver good, compassionate care, and also put into place a good process.

And AI relates to our core. We leverage anything technology-based to capture information more in real time rather than retroactively. So, for example, we partnered with CMS [the federal Centers for Medicare & Medicaid Services] in 2018, around bundled payments. They wanted to make sure we delivered good care while keeping utilization management within CMS goals. So we did a lot of analysis of post-discharge utilization. So for that, with the help of CMS, a real-time patient-monitoring tool that would help us share information with our post-acute providers—physicians, SNFs, and home health, to track patient utilization. So in this way, it helps us a lot, AI, it helps us to create rules to help our team to focus on areas of opportunity.

Can you drill down a bit on the leveraging of AI for post-acute utilization management?

For example, CMS will say, this is your heart failure patient. And I’ll make up numbers. So CMS will say, if anyone’s taking care of this patient, you should be able to manage that patient, so we’ll allocate $30,000 for a 90-day period, including hospital stay. And hospitals tend to be the major cost driver. So we built a background engine that would help us look at comorbid conditions such as diabetes, or the patient belongs to a certain socioeconomic class, or their age—what it’s like for this patient to be readmitted or for the patient to utilize post-acute resources.

We always get feedback when we reach out to the post-acute network. They’ll say, you’re sending us sicker patients. So what this tool does is that it factors in all these factors and gives you more of a competitive analysis. It helps us analyze optimal utilization. We’re not just looking at pure demographics, such as age; it makes adjustments so that we have more meaningful discussion.

Do you have any metrics to share around the progress you’ve made using the tool?

We do. So what we did is that before we launched this program with CMS, we looked at the historic spend of all our patients in every geographic area, and of each clinical episode. How much would a CHF [congestive heart failure] patient historically spend over the past three years, for example? And in our post-acute network, we can leverage our presence. And as claims start coming from CMS, we start analyzing. We track the utilization using our AI tool, and also marry the information from claims as well. We look at what we predicted. We were very close to our predicted impact around utilization from a cost perspective, and we were nearly at 98-percent accuracy. It also gave us an opportunity to say where we could further refine this process in years to come. And we were able to see that there were cases were our utilization was relatively high, figuring out how we could leverage our presence, through care management, etc. So we can manage patients, keeping quality at the forefront as well.

How many years have you been enrolled in the bundled payment initiative?

The BPCI [Bundled Payments for Care Improvement] Program was started by CMS in October 2018, and that’s when we started. We started with 99 clinical episodes, and comparing those with claims data, we’ve now grown the program out to 125 clinical episodes. That speaks to our confidence in the tool and our process, and making sure that we can deliver good care. But also, I’d like to point out that in our model, we bear risk as a system. We’re not passing that risk along to any physicians or post-acute networks. We have confidence that we can manage the risk. We should be able to partner with anyone.

What have the biggest challenges been so far, and the biggest lessons learned?

The biggest challenge has been bringing everyone onto the same page—physicians, clinicians, everyone. And everyone involved in the 90-day episode of care has a different business model. For example, in the hospital, it’s called geometric-mean length of stay, GMLOS. What that means is that hospitals get paid according to a certain DRG [diagnosis-related group]. When it comes to a heart failure patient without any comorbidities, normally, that patient should have a stay of only three days. If there is no justification for additional length, we won’t be paid extra. Now, when the patient transitions into post-acute care, [the post-acute care organizations] gets paid per diem, as long as the patient meets medical necessity. CMS has calculated that a CHF patient should normally stay about 14 days, and should be sent home after that; sometimes that can be stretched. In short, everyone has a different source of reimbursement. Physicians are paid per visit, for example. So making sure everyone walks towards the main core principle means of taking the best care of the patient within the same goals, that’s been our biggest challenge. We want to make sure we treat the patient right but that we don’t break the bank; that’s been our biggest challenge.

What have been the biggest lessons learned overall, in that regard?

That the better partnership we can have with our post-acute network, the better we can do. Anyone who’s performing better according to CMS’s standards, will also help us in terms of performance—so it’s to establish a partnership with the post-acute-care providers that can demonstrate better outcomes.

In other words, to be selective in your network development and management?

Yes, but at the same time, per CMS’s mandate, we always give patients a selection, recommending post-acute networks to them, but give them their choice. We have to respect patient choice.

What would your advice be to colleagues who would want to move forward in BPCI?

There’s still a lot of confusion about cost and length of stay. Many people think that attempting to minimizing cost will have an impact on outcomes; that’s not true. We have to best utilize resources, for better outcomes.