Why Dirty Data Can Make Healthcare Organizations Chase the Wrong Problems

Data has never been more readily available. It helps guide decision making in any industry, and healthcare is no different. From quality reporting and value-based care programs to workforce planning, population health initiatives and emerging AI applications, leaders increasingly rely on ever-expanding and more complex data to identify opportunities, measure performance and allocate resources.

As healthcare organizations adopt more systems and data sources, it becomes more difficult to ensure the information guiding decisions is complete, consistent and accurate.

The real cost of poor-quality data goes beyond inaccurate reports or the efforts required to correct them. It limits a healthcare organization's ability to understand performance, identify opportunities and make confident decisions. When leaders can't trust the information in front of them, they devote time and resources to validating data instead of acting on insights. In practice, inaccurate, incomplete or poorly classified data can affect everything from performance improvement efforts and reimbursement to strategic planning and operational decision-making.

More importantly, poor data can make it difficult for healthcare organizations to distinguish meaningful opportunities from measurement noise. There is no shortage of sources from which dirty data can emanate: poor architecture, legacy systems that merge or update, redundant or duplicative reporting from competing sources, transposed numbers or letters due to human error, an absence of validation controls. 

Performance metrics, benchmarking data and operational reports drive countless decisions across healthcare. But when the underlying information doesn’t tell the full story, organizations risk focusing resources on the wrong priorities, pursuing unnecessary interventions or overlooking opportunities that could drive meaningful improvement.

In that regard, the greater cost of dirty data may reside in the decisions organizations make — or they fail to even see — because they lack confidence that the information in front of them might not reflect the realities they face.

When Data Noise Obscures Performance Problems 

Improve quality. Reduce variation. Strengthen financial performance. Achieve better outcomes. The pressure is constant. But before leaders can identify where improvement is needed, they first need confidence that their data accurately reflects what’s happening within their organization. That distinction is important because not every performance issue is actually a performance issue; sometimes the underlying problem lies in how  performance is being measured.

Ballad Health experienced this challenge as it worked to improve visibility into quality and performance data across its health system. At Premier's Breakthroughs Conference in 2024, Ballad leaders described an environment where quality teams spent significant time maintaining reports, troubleshooting broken dashboards and manually reconciling information. Many reporting processes were highly manual and time-intensive, creating a reactive environment that consumed valuable resources and made it difficult to uncover actionable insights.

As reporting capabilities matured and leaders gained greater visibility into performance measures, they began taking a closer look at mortality outcomes. The data initially appeared to suggest significant opportunities for improvement. But rather than immediately changing clinical workflows or implementing new interventions, leaders first asked a different question: Did the data accurately reflect the complexity of the patients being treated?

Working with Premier’s Quality Enterprise team, Ballad examined opportunities to strengthen documentation and coding practices, helping ensure patient comorbidities and severity were more accurately represented within performance measures. With a clearer understanding of patient complexity, leaders were better positioned to distinguish true clinical improvement opportunities from measurement noise.

Ballad’s experience highlights a broader challenge facing healthcare organizations. When data lacks accuracy, comprehensiveness or context, leaders may direct resources to problems that are not actually driving performance while overlooking opportunities that are. In some cases, organizations risk redesigning workflows, launching improvement initiatives or altering processes based on signals that do not accurately reflect reality.

By improving documentation, coding and visibility into performance measures, Ballad gained a more accurate understanding of performance and focused improvement efforts where they could have the greatest impact. However, the organization's experience also revealed another important reality: data integrity challenges can carry significant consequences beyond just performance measurement. 

The Hidden Labor Cost of Dirty Data 

Quality teams, finance departments, operational leaders and clinical stakeholders all rely on timely, accurate data to support decision-making. When that information is incomplete, inconsistent or difficult to access, organizations often spend significant time validating what they already have instead of using it. 

Teams may reconcile conflicting reports, maintain separate spreadsheets, perform duplicate analyses or spend hours validating performance metrics before they feel comfortable acting on the results. These challenges are common across healthcare, and the costs extend well beyond labor hours to  delayed decisions and slower improvement cycles.

As was the case with Ballad, by reducing manual reporting activities and increasing access to trusted information, the health system was able to shift more focus toward analysis, improvement and clinical decision-making, rather than report maintenance.

Every hour spent validating data is an hour not devoted to improving patient outcomes, reducing costs or advancing strategic priorities. The true labor cost of dirty data is not simply the effort required to correct it; it’s the value organizations forfeit when talented teams are forced to spend their time managing information rather than acting on it.

When Data Errors Create Real Financial Consequences

Some data issues can have far more direct financial implications. 

As Ballad continued examining the accuracy of its data, leaders identified another area where information reliability mattered: administrative data used to support quality reporting and reimbursement calculations.

Healthcare reimbursement and quality programs rely on administrative data elements such as admission type, point of origin and present-on-admission indicators to determine measure eligibility, exclusions and performance calculations. Even small errors can affect how performance is measured and how organizations are evaluated.

Ballad, like many health systems, was navigating technology and reporting changes while working to improve enterprise-wide data quality. Leaders identified opportunities to strengthen how certain administrative data elements were captured, classified and mapped within reporting systems.

On the surface, performance metrics appeared to suggest worsening outcomes in some areas. However, further analysis revealed that some of the variation was tied not to care delivery but to how administrative data was being categorized and interpreted within performance calculations. When Premier helped identify and address those issues, Ballad estimated it avoided more than $2 million in readmission and HAC penalties.

While documentation and coding improvements provided Ballad with a clearer understanding of patient complexity and performance opportunities, improvements to administrative data integrity helped ensure quality scores and reimbursement calculations more accurately reflected the care being delivered. It also revealed that some costly data issues are not caused by missing information; they are caused by information that appears accurate until closer examination.

Healthcare organizations will continue investing in new technologies, which places increased importance on maintaining data integrity from integrating acquired facilities and modernizing data infrastructure. A single mapping issue, registration error or classification discrepancy may seem insignificant on its own. When applied across thousands of patients and multiple performance programs, the financial implications can become substantial . 

In this sense, dirty data manifests itself as more than just another operational challenge. It is increasingly a business risk that can influence reimbursement, regulatory performance and organizational financial health. If data is not standardized across organizations, it becomes an apples to oranges comparison.

When Organizations Stop Trusting the Data

The most difficult cost of dirty data to measure may be the impact it can have on confidence, especially when that data helps guide so many decisions every day. Leaders use performance reports to identify improvement opportunities. Clinical teams monitor quality measures to evaluate outcomes. Finance departments depend on accurate information to forecast performance and manage risk. And while organizations are also looking to advanced analytics and AI to support decision-making across the enterprise, at the same time, these technologies can amplify the importance of data integrity.

These efforts, though, depend on a common assumption: that the underlying data has merit and can be trusted. When trust and confidence in the data begins to erode, the effects extend beyond reporting and can create organizational friction: 

• Teams may spend additional time validating information before taking action. 
• Leaders may delay decisions while seeking confirmation from multiple sources.
• Departments may create their own disparate reports or maintain separate spreadsheets to reconcile perceived discrepancies. 
• Conversations that should focus on solutions can instead shift and become about which numbers are correct.

Advanced analytics cannot compensate for inaccurate information. AI can identify patterns, summarize insights and accelerate analysis, but it cannot reliably distinguish between a meaningful signal and flawed data if the underlying information is incomplete, inconsistent or incorrect – an unfortunate and costly realization of the familiar Garbage In, Garbage Out concept. 

Leaders who are confident that their data accurately reflects performance are free to move more quickly, prioritize resources more effectively and focus improvement efforts where they’re most needed. 

Seeing the Full Picture

More data alone does not automatically lead to better decisions. The impact of inaccurate, incomplete or poorly classified information is about delivery and performance, and not simply a technical issue. 

Data that creates noise, not clarifies it, makes true improvement opportunities indistinguishable from measurement artifacts. Organizations that establish trusted, accessible and accurate data foundations are better positioned to identify meaningful opportunities, prioritize resources effectively and act with greater confidence. Before healthcare leaders can solve the right problems, they first need confidence that they’re seeing them clearly.