What Avera McKennan Does

Avera McKennan Hospital & University Health Center, founded in 1911 and based in Sioux Falls, South Dakota, is a cornerstone of the Avera Health system. As a premier academic medical center, it integrates patient care, medical education, and research. Serving a vast regional population, including many rural communities, it provides a full spectrum of services from primary care to advanced tertiary and quaternary specialties. With over 10,000 employees, it operates at a scale that generates immense amounts of clinical, operational, and financial data, positioning it as a critical hub for healthcare in the Upper Midwest.

Why AI Matters at This Scale

For an organization of Avera McKennan's size and complexity, AI is not a futuristic concept but a practical tool for managing systemic pressures. Large hospitals face immense challenges: razor-thin margins, staffing shortages, regulatory penalties for readmissions and hospital-acquired conditions, and the constant drive to improve patient outcomes. AI offers the ability to move from reactive to proactive management. It can process the petabytes of data generated across the enterprise—from EHRs to imaging archives to supply chain logs—to uncover patterns invisible to human analysis. At this scale, even a 1-2% improvement in operational efficiency or a slight reduction in adverse clinical events translates into millions of dollars saved and, more importantly, countless lives improved. For an academic center, AI also aligns with its mission to advance medicine through research and the adoption of cutting-edge, evidence-based technologies.

Concrete AI Opportunities with ROI Framing

1. Predictive Analytics for Patient Deterioration: Implementing an AI model that continuously analyzes EMR vital signs, lab results, and nursing notes can predict sepsis or clinical deterioration 6-12 hours earlier. For a 500-bed hospital, this can reduce ICU transfers and mortality, potentially saving over $5 million annually in avoided complication costs and improved CMS quality metrics.

2. AI-Optimized Workforce Management: Machine learning can forecast patient admission rates and acuity with high accuracy. By dynamically aligning nurse and staff schedules with predicted demand, the hospital can reduce costly agency staff usage and overtime by an estimated 10-15%, directly improving labor margins—a major expense line—while boosting staff satisfaction.

3. Automated Clinical Documentation: Natural Language Processing (NLP) can listen to clinician-patient encounters and auto-generate draft clinical notes, reducing physician documentation burden by several hours per week. This addresses epidemic-level burnout and can increase effective clinical capacity by 3-5%, allowing providers to see more patients or spend more time on complex cases.

Deployment Risks Specific to Large Health Systems

Deploying AI in a 10,000+ employee health system carries unique risks. Integration Complexity is paramount; layering AI on top of legacy EHR systems like Epic or Cerner requires robust APIs and can create performance bottlenecks. Change Management at this scale is daunting; convincing thousands of clinicians and staff to trust and adopt AI-driven workflows requires extensive training and demonstrated reliability. Data Silos and Quality are exacerbated in large, multi-facility systems; creating a unified, clean data lake for AI training is a massive, costly IT project. Regulatory and Liability concerns are heightened; an AI recommendation that leads to a bad outcome creates novel medico-legal questions. Finally, Vendor Lock-In is a strategic risk; large investments in a single AI platform can limit future flexibility and innovation. A phased, use-case-specific pilot approach, coupled with strong governance from clinical and IT leadership, is essential to mitigate these risks.