Embracing Bundles: Let Data Be Your Guide

Bundling is premised on viewing healthcare as a continuum, but most of today’s healthcare systems use electronic medical records (EMRs) developed for episodic fee-for-service billing. While many in the cardiovascular community are at the beginning of this experiment, some health systems participated in the earlier Bundled Payments for Care Improvement (BPCI) Initiative and have insights to share.    

On July 1, cardiovascular practices in 98 geographic areas will begin participating in new payment models that, according to the Centers for Medicare & Medicaid Services website (CMS.gov), were designed to “support clinicians in providing care to patients who receive treatment for heart attacks, heart surgery to bypass blocked coronary arteries, or cardiac rehabilitation following a heart attack or heart surgery.” Currently, this program is not voluntary; if your health system is located in one of the CMS-selected “metropolitan statistical areas,” then you should expect to participate through 2021. And, if you are like the team here at the Wake Forest Heart and Vascular Center, you may be working alongside some colleagues who are excited to be part of this move toward population-focused healthcare and others who are worried about their ability to succeed in an environment where the goals seem familiar but the rules have changed. 

I’ll admit to feeling slightly daunted. Bundling is requiring us to rethink how we handle many, if not all, of our operational and financial activities. We have tasked new work groups with creating processes for assessing our performance and identifying opportunities for improvement. Our financial analysis team is modeling cost data for the defined populations by diagnoses while our quality team is identifying outcomes data for the same populations.

I spoke with Catherine McCarver, MBA, MHA, who participated in the BPCI initiative, first at Duke University Health System and now at Vanderbilt Heart and Vascular Institute (VHVI), where she is the associate operating officer.  

The BPCI initiative ran from 2014 to 2016, testing the bundling concept for a variety of conditions. At Duke, as director of Quality and Growth for the Heart Center, Catherine facilitated performance improvement teams focused on the cardiovascular bundles. Now, at Vanderbilt, she is responsible for VHVI operations, including the cardiac care bundle projects. She shared insights from her team’s journey and how they intend to transform challenging new requirements into opportunities to use data to continue improving their system.

Why did Duke and VHVI volunteer for BPCI?

I was at Duke when BPCI was launched, but I would say that for both health systems the decision to participate was prompted largely by an interest in learning proactively about the models, definitions and expanded analytics but also because these projects reflected our focus on innovating care. 

What surprised you in the early days? Did you need to make changes to operations, systems, staffing?

Surprises naturally came from having access to data not previously available, such as postacute care costs and how care transition decisions impact costs and utilization. Having those data, and being able to pair them with clinical evidence, enabled us to make important changes to operations. Bundle and quality measure definitions were an early focus as organizations evaluated outliers, variability and inclusion/ exclusion criteria. For example, to help reduce variability and improve patterns of care, we implemented care paths (or plans of care) and protocols for managing complications and medications. To address outlier and high-cost elements, we examined the evidence around use of high-cost medications and evaluated the data showing how we were using higher-cost equipment. We also stepped back and defined our routines for visits and for supporting patients in the postacute phase.

Which data were your team analyzing within utilization and cost?

The data from cost accounting systems were critical  to understanding costs and utilization. We were able to make the costs of services transparent to the clinical team, evaluate how often services were used and understand how much those services contributed to an admission or episode. The cost accounting data may be used in combination with productivity system data and benchmarking to evaluate labor costs in key services relevant to the bundle population—for example, the invasive lab for the PCI bundle or the operating room for the valve bundle.

How did you assess performance?

For our initial bundling efforts, we focused on cost, efficiency and quality perspectives. We used episode-of-care cost and phase costs, starting from the initial index admission.

How are you tackling the new bundling program?

At VHVI, we are translating ideas and structure from our early experiences. Right now, we are planning to use similar measures but will customize them to the populations in the new models. We may use process measures, based on population and areas of improvement, to ensure that the process changes take effect and are sustained.

[[{"fid":"22741","view_mode":"media_original","type":"media","attributes":{"height":512,"width":600,"style":"font-size: 13.008px; width: 180px; height: 154px; margin: 5px; float: left;","alt":" - c-mccarver","class":"media-element file-media-original"}}]]

How will you use data to work with teams and manage costs within the bundles?

We will bring summarized (trend) data as well as detailed cost and quality information to the bundle team to evaluate opportunities. The teams will systematically review groups of costs (for example, medications, supplies and diagnostic testing) against both clinical evidence and value to make changes to use rates or reduce use variability. They will also evaluate the cost impact of adverse outcomes, such as complications or readmissions, and determine strategies to mitigate them.  

What is the most important lesson you have learned for success in bundling?

First, understand the definitions and identify the patients in the bundle. Then, study the data to identify the outliers and the causes of variation. Those data were essential for helping us identify improvement opportunities. For example, while there are clinical indications for using higher-cost implants, such as some types of stents or higher cost anticoagulants, the teams came to consensus about the best approaches and situations to use those resources. As you go through the bundling process, each team member should contribute by thinking critically about current evidence and best practices and applying them to their practice.

What advice would you offer institutions at the starting line of the bundle marathon?

Overall, engage the team. Every team member’s perspectives and expertise need to be included. Specifically, use your cost and length-of-stay data to identify and capitalize on internal savings opportunities. Ultimately, use the project and CMS’s data to help you understand the postacute costs and improvement opportunities. 



Megan S. Berlinger, MHA, is the business administrator for the Heart and Vascular Center at Wake Forest Baptist Medical Center in Winston-Salem, N.C.