The Three Systems Critical to Healthcare Analytics Success

Healthcare analytics problemsThroughout my career, I’ve been lucky to work with a number of enlightened health systems at various stages of the journey to become data-driven healthcare organizations. They all have something in common: a strong desire to use analytics to improve the lives of patients, rather than just churn out metrics used to punish staff who underperform.

The leaders in these organizations talk to me a lot about the difficulties they encounter with analytics-driven care improvement. I’ve found that most of their challenges and concerns could be grouped into three core problems:

  1. Analysts spend too much time gathering data and producing reports and not enough time analyzing data.
  2. Despite reports and dashboards, data still doesn’t seem relevant. The organization’s reports and dashboards are not actionable.
  3. Improvements can be made, but the organization can’t sustain them. As soon as the organization moves on to the next project, prior wins disappear.

Addressing all three of these challenges is essential to long-term success with analytics.

Analytics-Driven Healthcare Improvement Starts with the Analytics System

The first challenge—giving data analysts the tools and information needed to do good analysis—is addressed by making data easy to find and easy to access. An analytics system ensures that data is aggregated, easy-to-access, and distributed efficiently. I’m a technologist, and my job is to help organizations create a solid analytics infrastructure that incorporates an enterprise data warehouse (EDW) and powerful analytics applications. But I am also keenly aware of how much beyond technology is required for success. (My colleagues have written extensively on the analytics system and why it is essential, so I won’t expound on it here.)

Many of us technologists like to think we hold the keys to success with technology, but the truth is, we do not control it. For technology to drive real clinical and financial transformation, we have to closely partner with other areas of the business in order to address the second two challenges—actionable dashboards and sustainable improvement.

The Challenge of Building an Effective Healthcare Best Practice System: Actionable Data

Have you ever heard the following?

“We have many reports and dashboards, but our data does not seem relevant and is not actionable.”

You can have a well-built data warehouse and flashy dashboards, but if the combination is displaying data that isn’t relevant or actionable for improving patients’ lives, your efforts to improve care with analytics will fail.

An Example: Actionable Heart Failure Readmissions Data

As an example, let’s say you’re looking at a dashboard that shows your heart failure readmission rate as a flat number with no further data as context. There’s nothing actionable about that information. Knowing that your heart failure readmissions are high is insufficient to drive change even if the numbers are accurate. You can react to the numbers and create a fire drill, but you can’t drill down into that data to discover what is really affecting readmissions. If you’re just measuring rates and not measuring the processes that affect those rates, you’ll have a hard time improving.

But how do you know which processes you need to focus on to really move the needle on heart failure readmissions? You have to look at the evidence base to identify the best practices proven to reduce readmissions. Then, you must measure how well your organization performs on these best practices. If you have limited resources—and most health systems do—you must ensure that your time and energy is spent on interventions that have proven value.

An organization’s analytics initiative must focus on what is actionable, relevant, and important. Ask the question, “What practices are really going to affect the outcome that we want to change?” Then look at scientific literature and widely accepted clinical consensus to identify the relevant best practices. Incorporating this evidence-based knowledge into your analytics is what we at Health Catalyst call a best practice system.

What Is a Healthcare Best Practice System?

A best practice system involves applying evidence-based best practices to care delivery. By systematically incorporating best practices into your improvement program, you can be sure that your interventions focus on what is actionable and relevant.

I like to illustrate the need for a good best practice system by talking about the weather. In the past, predicting the weather relied on techniques like consulting the Farmer’s Almanac. When I was a kid, my mom told me, “No one knows why, but the almanac is accurate. It has been used since the 1790s.” It sounded like magic devised by some kind of mystical farming society with powers beyond my capacity to understand.

The truth is, if you do a little bit of research you learn that the accuracy of the Farmer’s Almanac is folklore. Most of us are much more inclined to trust modern technologies that crunch a lot of data to predict the weather — we no longer rely on the groundhog and folklore to tell us what the weather is going to be. Rather, we rely on meteorologists who use a sophisticated best practice system built on science, evidence, and technology.

The same should apply to healthcare. The decision as to what we should measure to drive clinical and financial improvement shouldn’t be based on thumb-to-the-wind guesswork (or sunspots!). Rather, we build effective analytics initiatives around proven best practices and research. We need to look to modern technologies, science, and evidence to analyze patient care and determine the best way to improve it. Even when deciding what care to provide, we need to rely on evidence and modern best practices, not just the gut feelings of physicians.

Evidence-based Medicine: Don’t Do All the Groundwork Yourself

Building a robust best practice system takes effort, but the best analytics solutions save you from having to start from scratch. Instead, they come with a best practice system already in place. You can then tweak or fingerprint that baseline best practice system to meet your specific needs.

For example, if you’re setting out to improve heart failure readmissions, a good analytics partner will already have these pieces of the best practice system in place:

  • Evidence-based definitions of the target population
  • Best practices for reducing heart failure readmissions
  • The process metrics for these practices that must be tracked

These definitions, best practices, and process metrics are what is reflected and measured by an analytics dashboard focused on heart failure.

Again, I’m a technologist. I’m not an expert on best-practice content. But I can tell you from experience that successful analytics—the analytics that create meaningful improvement—combine robust, relevant best practices with impactful analytics technology.

The Healthcare Adoption System: Engaging People to Drive Sustainable Improvement

Every technologist wants to see his/her technology succeed. Even with a solid analytics system and best practice system, efforts to improve care with analytics can still flounder. I often observe organizations that implement a best practice-driven analytics system and make initial improvements but can’t sustain the gain. Either they lose the gains as soon as they move on to the next project, or they are unable to scale the success of a limited initiative across the entire enterprise.

This challenge arises from the lack of an effective adoption system. The adoption system is the people portion of any analytics endeavor. It is how best practice-driven analytics are translated into care improvement and how that improvement scales across an organization. It involves working with stakeholders throughout the enterprise to establish organizational structures that drive sustainable change.

Pillars of an Effective Adoption System

I won’t pretend that building an adoption system is easy. But it is an effective and incredibly gratifying process. The following is a brief overview of the three pillars that make up a good adoption system.

Pillar One: Permanent, Dedicated Teams

A good adoption system requires permanent organizational structures dedicated to and accountable for using data to drive improvement. At Health Catalyst, we call these permanent, dedicated teams workgroups. Workgroups tackle a specific clinical area, such as C-sections or heart failure. Each workgroup typically consists of a physician, a nurse, various operational roles, and technical staff. The exact makeup of a workgroup team varies from team to team, but the key point is that these teams are cross-functional—clinical and operational personnel working side-by-side with technical staff.

The workgroup determines how to implement improvement efforts. This determination is based on data analysis by the technical staff and applies the clinical and scientific knowledge of the rest of the team. For example, the group may note a large variation in length of stay following a C-section. The data may provide evidence that this variation is due to lack of standardization in care for these patients and decide that the first aim statement will be to increase the use of standardized order sets for these patients. The group is also responsible for vetting ideas and plans with the larger governance organization to get feedback and buy in.

An important feature of successful workgroups is that they are permanent and integrated. If you throw together a team to perform an analytics-based process improvement, and then disband the team once you start seeing improvement, you won’t sustain your gains. If the team doesn’t remain accountable and doesn’t keep monitoring performance, performance will slip.   Likewise, clinical people without access to good data to drive their decisions, or worse – technical people without clinical guidance – will not achieve sustainable improvement.

Pillar Two: Governance to Prioritize Resources

I was recently in a meeting that perfectly outlined the need for resource governance. The setting of the meeting was interesting—the walls were plastered with metrics. There were printouts of charts and graphs, printed spreadsheets, and sticky notes with the so-called “next important metrics” scrawled out. The sheer number of metrics was suffocating! The client let me know that they had previously made efforts to reduce the number of metrics tracked from 90 to four. The unfortunate outcome of several meetings was that the client decided they needed 94 metrics—and this was only for one clinical area. It is impossible to tackle all 94 metrics as the highest-priority. Leadership is required to determine what outcomes are most important for the organization and to focus on metrics that drive those chosen outcomes.

An effective adoption system has a governance structure in place to determine which metrics or improvement projects will be prioritized over others. Process improvement initiatives require permanent teams who rely on a limited number of data architects and analysts to help them. It is essential use an institution’s readiness and need to prioritize projects. This governance is best handled by a guidance team that oversees workgroups and takes into account resources and readiness and considers areas most likely to benefit from improvement.

Pillar Three: Agile Development Methodology

Effective adoption systems apply Agile principles to care improvement. Agile development is a proven method for developing software that meets the needs of the user. It involves constant testing and discussion with the user during the adoption process. These Agile development techniques have been around the software world for a few years now, but they are often new to non-technical people. Applied to care improvement, these techniques are very powerful.

In a workgroup using Agile methodology, the physician (who understands the clinical aspect of care), the nurse (who understands the day-to-day workflow), and the technologist (who understands how to expose data through the analytics system) work together to determine the best way to use data to drive improvement. They iterate together, test data, review results, incorporate feedback, and test again to develop the most efficient, data-driven care delivery system. They jointly own the outcome improvements, and they all strive to understand and resolve challenges along the way.

Three Systems for Analytics Success in a Healthcare Organization

In the field of healthcare analytics, there is more to think about than software. Don’t get me wrong, good technology and a strong analytics system are crucial. But you must also develop a robust best practice system and an adoption system that drives sustainable clinical and operational improvement. Fortunately, you can begin building your adoption system with just a single clinical workgroup. From there, the sky is the limit.

How can using these three systems transform your organization? Do you have any questions about to tackle the problems common on the journey to becoming a data-driven system?

Powerpoint Slides

Would you like to use or share these concepts? Download this presentation highlighting the key main points.

Click Here to Download the Slides

Loading next article...