Modern Healthcare: Big backers buy upstart Health Catalyst’s targeted data approach

Big backers buy upstart Health Catalyst’s targeted data approach

Chris Keller

Two decades ago, at a time when most hospitals were still documenting their patients’ health records by hand, Steven Barlow and Thomas Burton developed an analytics program to augment Intermountain Healthcare’s home-grown electronic health records.

It wasn’t an easy task.

Databases with analytical platforms in the retail and banking industries provided no model for the rapidly changing complexities of healthcare.


What they needed was a data warehouse and analytics system that was more flexible than anything on the market – one that could let relationships between key variables be determined after the system was built, not beforehand as was usually the case.

And what they built allowed Intermountain begin to measure quality improvements in real time. It also gave the Utah-based system a better way to track its overall cost of care.

Soon other health systems became interested in replicating the design, said Dan Burton, Health Catalyst’s CEO. But when Intermountain passed on commercializing the system, Barlow and Burton founded the Salt Lake City-based company, which since its founding in 2008 has grown to 500 employees with 50 major clients, including some of the largest healthcare systems in the country.

Health Catalyst is one of the more successful start-ups in a field that has drawn major interest from the giants of the information technology world. McKesson, SAS, IBM, Microsoft and Epic have all made major investments in the space.

“The market is very segmented and likely to grow as more health systems take on risk based payments,” said Dr. Maulik D. Majmudar, a cardiologist and associate director, Healthcare Transformation Lab at Massachusetts General Hospital.

Maintaining positive margins under new payment models like accountable care or bundled payments requires health systems to closely monitor patients and their own performance by pulling, aggregating and analyzing data contained in their electronic medical records as well as administrative claims and pharmaceutical records. It’s created a business opportunity, but also a possible stumbling block.

Many health systems, especially those that are not connected to well-capitalized large systems, lack the capacity to conduct this type of analysis. If they turn to an outside vendor, it still requires in-house expertise to make use of the insights.

Analytic software and data warehouse operators like Health Catalyst can identify operational inefficiencies, especially when there is a agreed upon set of metrics that should be followed. For instance, it can identify if most clinicians in a large healthcare system are giving patients a vaccine required by Medicaid.

But sometimes it is hard to make sense of the data, especially if it’s not focused on a specific variable like vaccine administration. Also, information can be difficult to find within electronic medical records. Professionals with a background in medical research and informatics are often needed to connect the dots.

To fill the gap Health Catalyst has in-house experts who can aid in implementing its data warehouse products. Health systems can also hire these professionals under a separate service agreement to do more advanced analysis. The company is also trying to develop automated tools they hope can do some of the analytical heavy lifting.

How it works

Health Catalyst’s data warehouse culls information from multiple sources, including payment systems, pharmaceutical forms and electronic health records. Unlike traditional databases that start with organizing reams of data into a repository, Health Catalyst’s system focuses on culling data that’s relevant to the process or disease the hospital wants to focus on.

Was a set of safety standards met? Were certain tests administered? Health Catalyst makes it possible to track these metrics in real time so variations from best clinical practices are surfaced right away.

“There is so much data, healthcare organizations often need a starting point,” said Christopher Hutchins, an executive in the analytics department of Great Neck, New York-based Northwell Health, who used Health Catalyst products in his previous position at Partners Healthcare in Boston. “Otherwise, it’s like trying to boil an ocean.”

Children’s Hospital of Wisconsin, which consists of two hospitals and a surgery center, is using Health Catalyst’s data warehouse to ensure its clinicians follow safety recommendations created by Solutions for Patient Safety, an organization seeking to reduce hospital acquired conditions at children’s hospitals across the country.

For example, the organization designed a set of interventions that used together can lower the number of central line-associated blood stream infections (CLABSI). The recommended practices cover the best ways to insert and maintain central lines. Seemingly innocuous procedures – such as using alcohol-impregnated caps on the end of each central line’s connection port to keep them as clean as possible – ultimately made a big difference.

By following these guidelines, the hospital reduced CLABSI by 54% in the past two years. It also reduced the rate of overall hospital inquired infections by about 44% over the past three years.

Other systems have used Health Catalyst to rethink what medical supplies are ordered, limit unnecessary diagnostic screenings, shorten the discharge process once a patients is cleared to leave the hospital, and ensure follow up appointments are scheduled within five days of discharge.

The model has won admirers from some of the biggest players in both the health and venture capital communities, who have already invested $265 million in the firm. Health Catalyst backers include Cedars-Sinai of Los Angeles, CHV Capital (an Indiana University Health Company), Epic Ventures, HB Ventures, Kaiser Permanente Ventures, Leavitt Partners, MemorialCare, Norwest Venture Partners, Partners HealthCare, Sands Capital, Sequoia Capital, Sorenson Capital, Tenaya Capital and UPMC Enterprises.

In a blog post on the company’s web site, Dale Sanders, vice president of product development, boils down data driven decision making to a few principals. Assess the situation, make a diagnosis, avoid false positives, identify best practices, execute and communicate the decision, and monitor and adjust after execution.

It isn’t as simple as it sounds, especially if the information is difficult to locate or there is no predetermined set of variables that can be monitored. In those cases professionals with a background in medical research and informatics are needed to track down, validate and analyze the data. Hospitals systems that have used Health Catalyst extensively usually have a deep talent bench in data analytics.

Take Texas Children’s Hospital as an example. The team size for a data warehouse and analytics project ranges from 10 to 15 people. However, its electronic medical records team has over 100 people, so they can pull experts in database design or statistical modeling into a specific project.

Even seemingly straightforward tasks can require a tremendous amount of expertise and time. It took Texas Children’s six months to build the right filters to define type one diabetes because physicians throughout its system were coding it differently or placing in different portions of the EMR. Only by cross-referencing patient records with pharmaceutical data was it possible to get a complete picture.

Hospital information needs often fall into three broad categories, said Darren Dworkin, Chief Information Officer at Cedars-Sinai in Los Angeles. They are looking for reliable answers to predefined problems like meeting safety standards; they are looking to discover the root cause of a problem; or they are looking for patterns and insights from multiple data points.

Health Catalyst’s technology automates the first category. The company’s developers are now trying to create software that will make it easier for hospitals to analyze problem areas or identify patterns in data that will generate opportunities for process improvements.

Two years ago the company sold an equity stake in exchange for the right to use and sell analytical models developed by Boston-based Partners Healthcare. In 2013, after Partners became responsible for about $100,000 million in risk-based payments, the system hired Health Catalyst to track key cost measures like which providers had unusually high referrals to post-acute care facilities.

“We didn’t really know specific referral patterns until we saw the analytics,” said Dr. Timothy Ferris, senior vice president of population health management at Partners Healthcare. After uncovering the patterns, the organization created an education program for its clinicians designed to reduce unnecessary referrals.

In 2013 Partners Healthcare invested $1 million in Health Catalyst and began working on clinical applications. These apps, based in part on Partners experience as a Pioneer ACO, could, for example, identify patients with congestive heart failure and offer a series of interventions designed to keep them healthy and out of the hospital after discharge.

In 2015 the relationship took a step farther when the two organizations entered into a $30 million agreement to refine the program’s population health platform. Partners also increased its ownership stake in Health Catalyst and agreed to let the firm license and commercialize analytical models it developed in-house.

“The idea is to pull out generalizable aspects of care and make analytics widely available” to health systems or community hospitals without a deep bench of data specialists, said Dr. Ferris. “They don’t need to reinvent the wheel.”

Related News