How to create a data analytics center of excellence in healthcare
Building the framework for effective & agile decision-making
Your decisions are only as good as your data; your data is only as good as your platform.
Introduction
There’s no shortage of data in healthcare. Even before COVID-19 accelerated the digital transformation of healthcare with telehealth, mobile apps, wearable apps, remote monitoring, and more, the healthcare industry was generating several thousand exabytes of unstructured and semi-structured data annually.
How do you describe the possibilities that lie within that much data? “Wealth of information” doesn’t even begin to cover it…
That is a magnitude of potential insight into needs, trends, costs, risk drivers, and efficiencies that can barely be imagined – to power better decision-making, reduce risks, and leverage opportunities for improvement and growth.
But without the processes for making sense of it, it doesn’t really matter how many bytes you have. Your data is only as good as your analytics.
So, how do you go about transforming your processes into a system for transforming data into meaningful business insights? You create an Analytics Center of Excellence.
This article outlines best practices for doing so, no matter your size, goals, or the size of your internal analytics team. We’ll walk you through the process of setting goals, building a team, and measuring your success.
Highlights of this article
- What is an analytics center of excellence?
- 6 needs of healthcare entities that an analytics center of excellence can deliver
- Building your analytics center of excellence team
- Training for excellence
- Making use of the right tools and data sets
- Setting goals & measuring progress
- Using analytics to tell a story
- Key benefits of your analytics center of excellence
- The don’ts: the risks of getting it wrong
- How to think about budget
- Building your analytics center of excellence step by step
What is an analytics center of excellence?
An analytics center of excellence is more than a framework of data crunching and technology. It’s a framework for thinking about data and creating a culture built around defined goals and agreed-upon outcomes.
It’s driven by people and processes and collaboration as much as by data. And no matter how you organize the components, if you haven’t laid the foundation for a culture of analytics excellence, the most sophisticated technology in the world will not produce it.
It’s a framework for thinking about data and creating a culture built around defined goals and agreed-upon outcomes.
Establishing the culture ensures that you can move forward on the path to analytics maturity in collaboration and with the confidence that success and failure will be shared in a spirit of curiosity and a desire to learn and improve. You can set realistic goals and track progress. You can course correct when necessary. A culture of analytics excellence promotes agility and accountability across functions and skillsets.
In short, an analytics center of excellence is a framework for transforming data into meaningful business insights that can be leveraged across an entire organization for better decision-making. It provides the tools to tackle questions with deeper analysis that provides better answers, in less time.
What you do with it depends on the goals you set for your organization and the opportunities you want to pursue. Your center of excellence will help you put your data to practical use – to define your path, steer clear of obstacles, and get to where you need to go.
6 needs of healthcare entities that an analytics center of excellence can deliver
The shelf-life of data in the healthcare industry is rapidly shrinking. The global coronavirus pandemic – while an extreme event – made this clear. Managing the risks of a rapidly spreading virus meant knowing what hospitals were experiencing daily, across entire regions.
1. Timely insights
You can’t make good decisions using out-of-date data or data that takes weeks or even days to analyze. Timely insights are a key healthcare business need that an effective process of analytics can provide.
A survey of data experts and business users as part of MicroStrategy’s recent Global State of Enterprise Analytics found that 25% of business experts gave up on getting an answer they needed because the analysis took too long. When seeking data required to make a business decision, only 3% of employees could do this in seconds – for 60% it took hours or days. Depending on the use case, outdated data can have negative impacts on decision-making – and some public healthcare datasets lag by years.
2. Effective storytelling
Healthcare organizations must be able to communicate effectively across a broad audience. Analytics can provide a form of translation that enables your data to tell a story. Effective storytelling across the myriad of constituencies of a healthcare business means being able to break complex ideas down and highlight key ideas in multiple ways, whether as talking points or visuals.
3. Data literacy training
Nearly half of all employees report that their companies don’t offer, or don’t make it clear that they offer, any form of data training. This lack of data literacy training is a potential liability in any organization, but in healthcare, it’s also a missed opportunity.
4. Trust in healthcare data
Analytics can help improve data quality, including
- Simple intuitive processes for data collection
- Automated data audits
- Dropdowns to reduce manual errors
- Integration to update linked systems
These measures, combined with a well-honed analytics strategy, can be used to build trust in health data, which reduces resistance to using the data and leads to informed decisions that optimize healthcare delivery and better patient outcomes.
5. Disciplined data infrastructure & scalable data modeling
Analytics can establish a disciplined data infrastructure so that your information is streamlined across multiple sources, with a source data inventory and consistent naming conventions, data formats, definitions, logic, and rules.
A scalable and strategically designed model of data repository, capable of evolving with new requirements, is a key benefit for keeping up as data grows in volume and complexity.
6. Simple tools
Perhaps most importantly, analytics can provide simple tools for frontline workers, management teams, executives and other stakeholders who need timely answers to important questions and better-informed decisions. Research shows only 44% of frontline employees have access to data – compared to 81% of management and executive teams.
A well-honed analytics strategy can be used to build trust in health data that leads to optimized healthcare delivery and better patient outcomes.
Building your analytics center of excellence team
There’s no one-size-fits-all model for building your analytics center, but there are some key roles for ensuring that you get the most out of the data at your disposal.
Here is an overview of some important players in the process and the skills and focus they contribute.
- Data Engineers: manage the collection, curating, cleansing, and management of raw data in preparation for its use by analysts and data scientists; their focus is on infrastructure, scale, data quality, and data functionality – these people are key to getting your analytics center of excellence started
- Data Scientists: use data to draw insights and develop predictive models utilizing artificial intelligence (AI) and machine learning (ML)
- Analysts / Analytic Consultants: take on a more business-facing role, determining use cases, gathering specifications, and developing reports, dashboards, trend analyses, and data outputs for clients
- Data Translators / Storytellers: translate data into simple and practical business solutions, communicated by story themes for each audience – this is a newer role
- Knowledge Consultants: integrate advanced analytics with perspectives from outside disciplines to generate answers to business questions; few organizations have team members in this role, but external resources are available
- Business Users: the business domain experts and decision-makers
- Leaders: senior executives and department leaders who lead the charge, motivate the team, hold others accountable, and demonstrate the data-driven decision-making approach
Ensuring the right mix of talent
There are more than “roles” to consider in putting together your team of excellence. Focusing on the talents needed instead of just assigning roles will ensure that you have the necessary skills to convert data analytics into insights and answers accessible across your organization.
An unrealistic approach to roles, in fact, is the reason some organizations fail to capitalize on the insights generated by data analysis, according to a 2019 Harvard Business Review article.
Author Scott Berinato writes that it’s a problem that can often be traced back to unrealistic expectations about the role of data scientists, who may lack the communication skills to “sell” the results of their analytics to decision makers.
“Data teams know they’re sitting on valuable insights but can’t sell them,” Berinato writes. “They say decision makers misunderstand or oversimplify their analysis and expect them to do magic… Executives meanwhile, complain about how much money they invest in data science operations that don’t provide the guidance they hoped for.”
By focusing on the talents needed, from project management to data wrangling, to communicating the results to stakeholders, Berinato argues, analytics teams are better positioned to leverage their insights.
Training for excellence
Getting the most from your center of excellence team should include ongoing and accessible training and opportunities to improve. Certifications and continuing education in both technical proficiency and healthcare analytics will ensure that your team has the skills and confidence to dig deeper, ask the right questions, and challenge results when necessary.
Education and certifications should seek to establish competence and understanding around a wide variety of industry and analytics topics including healthcare finance and economics, health cost guidelines, benchmarks, hierarchical groups, risk scores, predictive analytics, and population analytics, among other subjects.
Finding a partner who can provide easy-to-use and high-quality educational tools, with options that include integration into your own learning management system, is critical to the success of your analytics center of excellence.
Making use of the right tools & data sets
Reliable tools and data sets are key to ensuring that your data analytics center provides the answers and insights you are looking for. Whether you “buy” or “build” the analytics you need, you should begin with an awareness of the tools and models available for managing, processing, and using your data.
The right third-party tools can be an effective way to ensure the quality of your data so you can drill down to the answers you need, with confidence in the results.
For example, to answer questions around the cost of readmissions, you might start with high-quality, third-party benchmarks and analysis to validate your data before going deeper to flag avoidable readmissions. The right data and analytics approach will enable you to break down the specific underlying factors you can use to drive improvements, including:
- Sources of readmissions
- Diagnosis codes
- Doctors or attending physicians associated with readmissions
- Facilities with highest readmissions
- Care and services received between admissions
The most useful analytics products and software enable a nuanced and precise examination of the subject matter, to reveal prescriptive insights you can use to take action and achieve results.
Setting goals & measuring progress
The goal of your analytics center of excellence is a broad one: to create the infrastructure that transitions your business to a fully data-driven decision-making process. But the success of the bigger goal depends on your ability to communicate the benefits, to identify and root out barriers, and to establish and meet smaller, more easily achieved goals with defined metrics for success.
Barriers can include things like mistrust of the data, a reluctance to move away from legacy tools and processes, and a sense of unfamiliarity with new approaches. A data literacy plan can be a good place to start, beginning with establishing a common language for talking about data and data insights across the organization.
A phased plan can also help you communicate the larger narrative of how a better system of analytics will drive process improvement. Ideally, your plan will distinguish between the different parts of your audience to tailor messaging.
The training of frontline workers, managers and leaders will help them achieve the fluency needed to incorporate data-driven decision-making into their roles.
Goals should incorporate insights from a broad cross section of participants, with communication, feedback and learning actively encouraged.
Using analytics to tell a story
Analytics is more than a way of organizing data into agile formats for crunching numbers and spitting out stats.
Analytics is a system for transforming data into meaningful insights. It’s about making practical use of available information to reach better decisions, grow, reduce costs, risk, and effort, find opportunities, and gain a competitive advantage.
Analytics is a system of thinking about your data as a source of inspiration and ideas. The right system of analytics provides more than answers – it guides you into asking the right questions. To make your data work for you, you should think of analytics as a way of making your data tell a story.
The right analytics can deliver powerful stories, with layers of information and insights.
The four categories of analytics illustrate the way that analytics can be used to establish a variety of story lines.
Descriptive analytics: tells the story of what happened using an analysis of historical data and trends.
Diagnostic analytics: helps explain why it happened, by examining cost drivers, risk assessment and correlations.
Predictive analytics: tells the story of what is likely to happen next, by modeling algorithms based on historical data. At the patient level, predictive analytics can be used to enhance treatment and intervention decisions, especially for patients with complex medical histories or multiple medical conditions and most at risk for readmission or disease progression.
At the organizational level, predictive analytics can help anticipate and achieve effective resource allocations to:
- Match patient utilization patterns, seasonal needs, and demographic changes in patient populations
- Identify missing reimbursement codes to improve financial performance
- Optimize supply chain management
Prescriptive analytics: tells the story of what could be done to influence an outcome by identifying at-risk populations based on current data and/or real time monitoring, defining timely treatments, testing potential outcomes, and recommending a course of action.
Whether you are using analytics to identify and adjust patient utilization patterns, to create treatment plans, or to optimize supply chain management, you are essentially telling a story. A story organizes a flood of data into an accessible narrative that can move your decision-making from conceptual to concrete. A story also brings everyone along in a spirit of collaboration toward a shared goal.
The right analytics can deliver powerful stories, with layers of information and insights. The most compelling stories, of course, are the ones that you can show. When you combine analytics with visual storytelling that uses compelling, succinct, and easy-to-understand dashboards and charts, you multiply your storytelling power.
Key benefits of your analytics center of excellence
When you get it right, your analytics center of excellence will yield a wide array of benefits, including:
- Greater confidence in business decisions
- Measurable, actionable insights into opportunities and emerging trends
- Reduced costs, improved profitability, and value
- Increased quality
- Greater trust in your data
- Improved accessibility to data for easy-to-use results
- More effective collaboration with fewer silos between departments and roles
- Operational efficiency and better allocation of resources
- Greater accountability
- Accelerated learning
- A more competitive organization
The Don’ts: The risks of getting it wrong
An effective process for turning analytics into meaningful insights doesn’t just happen.
If you set up your analytics center without a shared understanding of what you want to achieve, you risk setting unworkable goals and undermining confidence in your data.
You need to have your data in a structured environment that people can access, define, and trust. If you don’t build your data in a way that people can trust, you diminish the value of collecting it.
Here are some common pitfalls to avoid:
Don’t forget to reconcile the data to key financials. If there are multiple data sources, they need to agree on any overlapping data elements (e.g. Allowed PMPM, bed days per 1,000).
Don’t limit analytics to only a few data scientists. This impedes collaboration among functions and departments.
Don’t shortchange communication and translation, which are critical to the process of understanding and addressing business needs.
Don’t punish negative metrics or discourage team members from responding, pivoting, and adapting when an effort fails. This is how you learn.
Don’t boil the ocean: high-level, poorly defined goals can leave you sputtering around in high-level failures. Start with a concrete business objective to avoid the practice of delivering analytics in search of a problem.
Don’t pursue a multi-year “waterfall” approach that doesn’t deliver value until the end. Things change and priorities may shift, along with personnel. Agile methodologies and small, well-defined goals are a better option for harnessing and accelerating analytics excellence.
If you don’t build your data in a way that people can trust, you diminish the value of collecting it.
How to think about budget
When done effectively, establishing your analytics center of excellence can have a dramatic impact on your business growth and financial performance. It’s important to think about your investment in terms of the specific opportunities it will bring.
That means estimating the costs of growth opportunities, and the costs of losing such opportunities by doing nothing. Key questions to ask include:
- What is the potential impact on business growth and financial performance?
- What is the estimated value in competitive advantage?
- What is the estimated value in added growth, new products, new revenue streams or new business models?
- What is the cost of doing nothing in terms of data quality issues, duplication of effort and lack of timely insights?
“Agile sprints” can deliver repeat value within a short period of time to build and sustain momentum.
You can also think about budget by planning efficient implementation with a phased approach utilizing well-defined short segments or use cases. Such “agile sprints” can deliver repeat value within a short period of time to build and sustain momentum. Establish a definition of “complete” for each sprint to enable communication, accountability, and value measurement.
Agile sprints also reduce risk by spreading your investment out over time while generating value with each progressive sprint in the context of your larger strategy. The key skillsets and personnel needed for each sprint should be identified and include both new hires and existing experts. Some sprints may require a greater array of positions but not everyone may be needed for every sprint.
You should also think about partnering to augment existing capabilities – you don’t have to do it all internally to reap the rewards of a well-functioning center of excellence.
Building your analytics center of excellence step by step
No two organizations are exactly alike, and your needs and goals will drive the decisions you make in establishing a center of excellence in analytics. As you progress, however, you will want to keep in mind some best practices for ensuring success. Here is a checklist of key steps along the way:
- Build a shared understanding of goals and agreed-upon outcomes. Identify barriers that will have to be overcome, including potential reluctance to move away from legacy tools and processes.
- Plan your budget, being sure to estimate both the business opportunities of moving forward as well as the costs of doing nothing.
- Incorporate tools that allow you to translate data with compelling visuals for better storytelling.
- Incorporate tools that ensure you can quickly and easily access the data you need when you need it.
- Ensure that you have the analytics expertise and the data you need and the right automation for integrating data sets and testing results.
- Explore your options for buying ready-made data analytics software as a foundation builder or partnering with an analytics provider to augment your existing capabilities.
- Ensure that leadership is actively supporting the effort, with incentives for managers and frontline professionals to learn new skills, maintain open lines of communication and encourage the sharing of successes and failures to promote learning.
- Establish a data literacy plan, including transparency of terms and definitions so everyone is speaking the same language and your goals and metrics are clear and accessible to everyone.
- Train frontline workers, managers, and leaders.
- Create a series of short, concrete, and achievable goals with well-defined success metrics to deliver early value and build momentum.
- Identify the key skillsets and personnel needed for each of these goals.
- Utilize collaboration and ongoing communication across functions and departments to establish a culture of transparency, agility, and learning; analytics excellence is built around a culture more than any particular technology.
- Establish a process of agile response to early indicators and progress metrics with shared insights and suggestions for improvements.
- Put your organization on the path to excellence by starting with the right tools and expertise.