NYU Langone Health builds a data-driven culture featuring Tableau visual analytics

For the past 15 years, NYU Langone has followed a vision to transform itself “from a distributed and ad-hoc management and decision-making organization” to one that is “centralized, planned, and data-driven.” Among the tenets of this vision are achieving quality improvements throughout the organization and democratizing access for all users. In these areas and many others, visualizing data plays an important role.

“I learned from my years as a researcher about the importance of using data to tell stories,” said Cho. “In research, you often use storytelling to make your findings more engaging with people who aren't familiar. This skill has translated well to the hospital, where we look at data to tell us stories about how each patient is doing, and what trends we're facing on the macro level.”

The continuous development and refinement of Tableau visualizations at NYU Langone has yielded powerful tools for analyzing highly relevant clinical data, such as the High Risk Care Coordination dashboard shown in Figure 1, which reports on cases for patients with at-risk conditions. More than 100 dashboards use over 800 key performance indicators to track a variety of measures, from operational analytics and hospital resource capacity, to patient census and lengths of stay, to employee safety metrics and patient experience factors, to COVID-19 leadership scorecards and detailed financial performance. These dashboards represent a myriad of data sources, which the NYU Langone team has expertly wrangled over the past decade to yield meaningful, actionable results.

Data analysis also plays a key role in managing financial activities at NYU Langone, including tracking ROI, providing executive visibility into financial performance, and predicting caseloads and other economic drivers. The fiscal health of the organization depends on the well-rounded collection and analysis of financial data, including payor receivables, revenue cycles, purchasing, payroll, contractor relations, and clinical activity.

Tools such as the One Finance: Enterprise to Department dashboard, provide summary and detailed information to all organizational levels, from executives down to individual care team members, enabling insights into operating margins, revenue, expenses, and service volume. NYU Langone also uses Tableau to predict future readings of this data so that leadership and other decision makers can plan and adjust their budgets accordingly.

“Our organization has a real thirst for using data to drive decisions, and at the same time, we're hyper-focused on performance,” said Steve Chatfield, Assistant Vice President in Finance, reporting directly to the Chief Financial Officer. “So it's a cultural thing that you're expected to know the data in order to stay on top of expenses and trends, and to be accountable for knowing about issues before the people higher up ask questions. Jeff's team does a great job in providing the data sources so our Finance team can develop the dashboards.”

Some of the dashboards help to further promote this accountability by ranking physician performance. Among other benefits, this tactic helps identify opportunities for documenting patient conditions, which in turn ensures accurate reimbursement and quality rankings. Visual analysis in Tableau enables this to be a much more dynamic, proactive process than it was in years past.

“Previously, we'd produce all these ad hoc reports,” said Chatfield. “At meetings, we'd identify a blip or a bubble, and everyone would sit around and wonder: What's behind that? We'd have to take it away, investigate it, and then meet again one week later to discuss. Now, we simply click on it, and there you go—the discussion can continue in real time.”

Population health management and payor network integration represents yet another area where finances, operations, clinical activity, and patient outcomes intersect over the thoughtful management and presentation of relevant data at NYU Langone. NYU Langone’s clinically integrated network, NYUPN, actively manages more than 500,000 lives through shared savings contracts with seven managed care plans including commercial, Medicare Advantage, and Medicaid product lines.

The task of managing the health of a population and associated financials is complex, given that the data takes many forms: claims data, patient rosters, contract and eligibility details for all patients in all plans, pharmacy files, and quality files that reference potential and actual gaps in care. Differing file formats, data schemas, and naming conventions for each payor add further complexity, as does the reality of different payors providing information on different cadences at different times of each month.

In the past, manual juggling of all this data made for chaotic, often unproductive, payor operations. To address the need for clarity, and to put the health system in a proactive stance for addressing communications and performance, NYU Langone embarked on a multi-year, in-depth strategic initiative focused on using payor data in better ways to make more informed decisions.

“The old way was not sustainable, especially considering we're always growing, we're engaging in new contracts with payors, plus taking on more patients and more providers—and so the complexity continued to grow,” said Carrie Rooke, Senior Director of Operations for Network Integration. “We needed a better way of linking things on the back end, and getting accurate, workable data in front of us every day. That's where the IT and business teams really came to the table and said: How do we get this done?”

The solution was to devise better ways of aggregating and normalizing the multitude of data types and formats, and then build dashboards in Tableau that enable visual analytics to drive action. The project involved not only cleaning up the payor data, but also overlaying it with clinical data in Epic in order to drive insights on the level of individual patients. For example, if a patient recently had a mammogram, this fact would show up in the data immediately, and the system wouldn't trigger an unnecessary reminder to reach out about the test being due.

This level of data integration at both the software interface and human interface levels generates new kinds of efficiency for Rooke's team when conducting outreach activities across the payor and clinical networks. Tableau dashboards produce accurate summary views and action lists based on timely, paid services patients might need. Precision in the data roll-ups that enable drill-down capability into details help the staff hit strategic targets and track both patient and in-house communications such that no cases get missed. Having data and insights readily at hand also confers greater authority on NYU Langone as an organization, positioning them to hold payors accountable for errors or omissions in the data they provide, as well as to pursue more relevant conversations with the patients themselves.

Rooke cited a few recent examples. “In 2020, we identified a massive data gap with one of our payors, and because we were organized on our side and could visually present our data, we were able to make them clearly accountable for correcting the situation,” she said. “On the flip side, we recently attracted the attention of executives at one of our health plans based on how we use Tableau, and that has generated a lot of excitement about new use cases that we can all apply to achieve even greater levels of collaboration and measurable improvement.”