Analyzing Failures in Data for Healthcare

Engineering Reliability into the Systems Healthcare Leaders Already Trust

A thought-leadership brief on engineering reliability into healthcare data, from HFMEA to the Epic Sepsis Model and CMS accuracy standards

5 min read

By Roberto Hernandez

Introduction

Healthcare leaders already understand reliability when the subject is a ventilator, medication pump, or clinical protocol. The harder question is whether they apply the same discipline to the data now driving alerts, dashboards, quality scores, and regulatory submissions. The Epic Sepsis Model made that risk visible: an algorithm embedded in everyday hospital workflow was externally validated in 2021 and found to miss roughly two-thirds of true sepsis cases. It appeared to work until someone measured it. That is the central leadership challenge of modern healthcare data: the most expensive failures are often silent.

Where the Ideas Came From

Healthcare data should be managed with the same reliability discipline applied to clinical equipment and protocols. FMEA and HFMEA help teams identify, prioritize, and address potential failures before they affect patients. The Epic Sepsis Model demonstrates how silent failures can persist when analytics are deployed without rigorous validation and monitoring. Service Level Objectives and detection metrics can establish measurable standards for critical data, dashboards, alerts, and regulatory reporting. Leaders should consistently ask what could fail, how they would detect it, and what reliability standard they are prepared to govern against.

What It Looks Like in Practice

The framework is straightforward: use FMEA or HFMEA to identify how a data-dependent process can fail. Use Service Level Objectives to define the minimum acceptable performance for critical datasets, dashboards, alerts, and regulatory submissions. Use detection and resolution metrics to determine whether the organization can identify a problem quickly enough to act.

The Epic Sepsis Model illustrates why this matters. A traditional dashboard might show that the alert is live. A reliability review asks different questions: What failure mode matters most? Is it false alarms that fatigue clinicians, or missed cases that never appear? How would leaders know if sensitivity drifted over time? The external validation made clear that silence was the dangerous failure mode: missed sepsis cases carried high severity and low detectability.

HFMEA shows the prospective alternative. Before launching or expanding a clinical decision-support tool, a cross-functional team can map the workflow, score failure modes, and decide which controls must exist before go-live. The point is not to eliminate all risk; it is to make risk visible, prioritized, and owned.

CMS is moving the same discipline from best practice toward mandate. For the 2026 reporting performance period, CMS has posted updated electronic clinical quality measure specifications and implementation resources. Leadership teams should not wait for auditors to define reliability for them; they should set internal targets for the data behind board reports, quality measures, reimbursement, and patient-safety decisions.

Figure 1. Reflection Matrix: FMEA, Epic Sepsis Model, CMS Accuracy

Key Takeaways

·       Healthcare data now deserves the same reliability discipline hospitals already apply to clinical operations.

·       The Epic Sepsis Model shows the cost of silent failure when high-stakes analytics appear to work until independently measured.

·       HFMEA gives leaders a proven method to anticipate, score, and control failure modes before they reach patients.

·       CMS accuracy expectations make data reliability a governance and compliance issue, not only an operational best practice.

·       Boards and executives should ask three questions consistently:

o   What could go wrong?

o   How would we know?

o   What reliability standard are we willing to govern against?

About the Author

Roberto Hernandez is an Associate Consultant with over 10 years of experience in technology, academia, and construction. 5 of those years are in leadership roles that offer a new perspective on how reliability engineering plays out in day-to-day operations.

About Paracletes

Paracletes is a performance-based consulting firm that helps senior leaders close the gap between operations, strategy, and technology execution.

If you would like to discuss how Paracletes can help your organization reach a level of performance it has not yet considered attainable, contact us at contact@paracletes.com or visit www.paracletes.com.

References

Centers for Medicare & Medicaid Services. (2025, May 8). Updated eCQM specifications and implementation resources for 2026 reporting/performance period. eCQI Resource Center.

DeRosier, J., Stalhandske, E., Bagian, J. P., & Nudell, T. (2002). Using health care failure mode and effect analysis: The VA National Center for Patient Safety’s prospective risk analysis system. The Joint Commission Journal on Quality Improvement, 28(5), 248–267.

Wong, A., Otles, E., Donnelly, J. P., Krumm, A., McCullough, J., DeTroyer-Cooley, O., Pestrue, J., Phillips, M., Konye, J., Penoza, C., Ghous, M., & Singh, K. (2021). External validation of a widely implemented proprietary sepsis prediction model in hospitalized patients. JAMA Internal Medicine, 181(8), 1065–1070.