Clinical Data Management: 3 Improvement Strategies

Add

They say a sign of an organized mind is a cluttered desk. Either you’re that cluttered genius or you know one in the office. Stacks of files and folders cover the desk, floor, and shelves. Reminder notes stick to any remaining available surface. How you or anyone else ever finds that one needed document remains one of life’s great mysteries.

Another great mystery is how anyone can work with the equivalent clutter in our industry: healthcare data. Like an office with stacks of paper, healthcare must contend with petabytes of data, spread across numerous disparate systems, with more accumulating every second of every day (Figure 1). It’s been estimated that the volume of healthcare data will grow from 153 exabytes, where it was in 2013, to 2,314 exabytes by 2020. Though we are developing the architecture to handle this volume, clinicians would be wise to remember that more data doesn’t always mean more insight. Sometimes more data is just more clutter.

complex hc analytics

Figure 1: Typical complex healthcare analytics environment

For an industry evolving to value-based care, clinical data management is crucial. Today, to assist an organization in answering its most pressing questions, a healthcare data analyst must navigate a complicated information infrastructure.

The first step in that navigation is to identify the proper data sources. The analyst may work with an IT resource who can help determine which data would be most beneficial to start the analytic process. Collecting raw data from disparate systems is only the beginning. Distilling the information so it focuses on a specific question or strategy enables the analyst to discover more meaningful insights. Leadership, working together with one source of truth, can identify areas for improvement and take quick action to address them.

Problems with Traditional Clinical Data Management

The standard approach to clinical data management presents two formidable challenges.

Spreadsheet Fever

Many hospital departments have developed DIY tools to confront the profusion of data. They use multiple spreadsheets, often containing information that may be inaccurate or conflicts with corporate data, which leads teams to focus on the wrong priorities or miss improvement opportunities altogether. Source data, once identified and captured, often become shadow systems with one spreadsheet linking to another and then another. This spreadsheet mania further complicates the process and introduces an unnecessary risk for error.

Changing Analytics Roles and Tools

In today’s value-based care environment, data analysts must be equipped with the right tools to identify performance gaps in the organization and to create actionable recommendations that drive improved outcomes. Providing the proper infrastructure mitigates the risk of compromising data during capture or transfer. Accurate and timely data is crucial to engaging clinicians and gaining their trust.

As Figure 2 shows, analysts spend roughly 80 percent of their time hunting and gathering data as opposed to performing high-value work, such as interpreting data and making recommendations about how to improve the outcomes.

reduced waste time

Figure 2: The current, wasteful state of clinical data management versus spending most of time on value-added activities

 

Most analysts would rather conduct strategic analyses and contribute to organizational decision making. Today, the approach to clinical data management focuses most analytic resources on the task of hunting and gathering. Creating a process that is efficient, effective, and reliable will enable analysts to spend more time on gleaning insight from the data.

Most leaders also dislike the current approach. The multiple pressures of improving care, reducing cost, and planning for the future require near-real-time data. Submitting a request, being placed in a queue, and waiting weeks for a response is no longer an option.

Three Strategies to Improve Clinical Data Management

So, how do we change the way data is managed within a healthcare organization to ensure the right data is found quickly and that all the analysts are using that data? Here are three strategies to improve clinical data management.

1. Identify the Analysts in the Organization

To align the analysts, a good first step is to simply identify the current analyst pool sprinkled throughout the organization. Finding all the analysts can sometimes be a challenge, but one way is by working with HR to get a list of positions with names similar to “analyst,” “specialist,” or “informaticist.”

2. Assess Analytic Improvement Opportunities in the Organization

Once the analyst pool is determined, elect a core analyst team responsible for assessing the risk within the organization. Some of the team’s new duties, along with the risks they are certain to uncover, include the tasks in the table below:

 

Task

Risks and Challenges

Create a report inventory and find the logical “owners” of each report. It is most practical to start with recent reports, pulled during the last year. Reports that haven’t been run in over a year are candidates for archiving. Work with the owners to prioritize the work of combing through each report to document its purpose, rules, tools used, frequency, data sources, formats used, and steps taken to produce it. This process will lead to better documentation AND fewer reports that need to be touched upon system upgrade.
Most organizations are surprised at how many reports have been created (and maintained) over the years. Any given system may have thousands of reports run out of the EMR. Folder security can sometimes force duplication (and probably modification) of the same reports with no audit trail for what is different. Additionally, the report filters and logic are often hidden from the report consumer, making it next to impossible to determine the rules and ultimately, the “source of truth.”
Gather the analysts to develop a list of core competencies and a program to provide ongoing training and mentoring.
Analytic excellence is tough to measure, but without certain core competencies, there is no way to guarantee high quality, consistent results.
Assess the degree of silos and political will to improve alignment.
Analytic silos breed duplication and potential waste. Although this seems obvious, the political climate often complicates remedying this risk. However, it can help inform the type of alignment needed to reduce the risk.
Determine the current method for requesting reports and analytics.
Without a formal intake, triage, prioritization, and assignment of work, the reporting and analytic environment becomes the “wild, Wild West.” Without a disciplined process, there is increased risk that staff are not focused on the highest priority work.
Identify current data governance processes and ownership within the organization. Current data governance might be performed through numerous, disconnected committees so dig around.
Data governance focuses on managing data from initial capture in a transactional system, such as an EHR or laboratory system, through its aggregation into reporting structures and data stores and enterprise data warehouses. The intent is to make sure that each step of the health data management process is controlled and the effects of processes on the data are well documented and understood.

This initial assessment will clarify the current the level of risk and potential inefficiencies that exist in the analytic organization. This information can be shared with executive leadership to make the case for changing to a new organization model.

There are several analytic organizational models available, including Business Intelligence Competency Center or Analytics Center of Excellence. The most comprehensive model, however, is to align the analysts organizationally. All analysts are members of one analytics department, reporting to a single leader. They provide cross-organizational support, working for various departments as needed.

It can be difficult to implement this model as it can be incredibly disruptive to the organization. The political capital needed to build the organization may also be lacking.

An alternative, less dramatic solution, is to create a central business intelligence group to serve as a consulting group. This group will offer specialized services to functional areas and produce findings to help determine organizational risk and its readiness for change.

3. Create an EDW as the Framework for Clinical Data Management

Create or purchase an enterprise data warehouse (EDW) as a critical step toward a robust analytic infrastructure. The EDW becomes a safe, central repository of data that is organized and optimized for measurement, analysis, and reporting. Sure, this is a large effort, but the payoff is significant and long-term.

The true value of the data warehouse is to organize data, provide links across disparate data sources (so the analysts don’t have to), and provide access so analysts and clinicians can “fish for themselves.” Aligning the analysts and developing clear clinical data governance and management policies will strengthen the entire analytics environment.

For example, the patient identifier provides a valuable link between the EMR data, departmental sources, and patient satisfaction data. After the data is pre-linked, it can be organized in a way that allows for fast reporting and visualization. When used correctly, the EDW becomes the one place analysts rely on for information, creating a single source of truth for the entire enterprise.

 

LB EDW

Figure 3 – Late-Binding™ Enterprise Data Warehouse

Once an EDW is in place, align the analysts and develop some important clinical data governance and management policies to strengthen the entire analytics environment and ensure the greatest return on investment.

Create a New Model to Improve Care and Deliver Better Outcomes

Creating a new model that supports effective clinical data management requires an EDW to assemble and coordinate data from across the organization, providing the foundation to improve care and deliver better outcomes.

This streamlined model greatly reduces lead time, enabling analysts to provide strategic insight from the disparate data collected across various systems. Healthcare organizations that embrace the value of data analytics, and harness it to create an effective clinical data management model, will be in the best position to survive and thrive in the new era of healthcare.

 

Which of the following common analytics problems represents the biggest roadblock in your organization?

    View Results

Loading ... Loading ...

PowerPoint Slides

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

Loading next article...