Case Study

How CMS used AI to cut hospital inquiry turnaround time by 35%

A custom GPT solution helped the agency deliver faster, clearer SEP-1 guidance—reducing backlog, improving consistency, and easing staff burden

Health Mental Health and Substance Use Clinical Quality Measures Federal Analytics & Operations Data Solutions
Client
CMS: Center for Medicare and Medicaid Services

AI helped CMS respond faster and more consistently to hospital questions—freeing staff time and helping providers focus on patient care.

18%
increase in timely resolutions, with more questions answered within the same month they were submitted
Client Need

Managing a surge of hospital questions on SEP-1

The Centers for Medicare & Medicaid Services (CMS) relies on quality measures to help hospitals improve patient care. For the Severe Sepsis and Septic Shock (SEP-1) measure—an assessment of whether hospitals follow evidence-based protocols—providers frequently seek clarification on how to apply the standard to complex clinical situations, submitting nearly 60 questions per day. Each inquiry requires 10–20 minutes of expert review, creating a daily backlog of more than 300 unanswered questions and delaying the support hospitals need to deliver high-quality care.

Our Approach

Testing AI to support faster, consistent responses

To help CMS respond more quickly, Mathematica and its partner designed and built a custom ChatGPT-based solution that generates standardized, formatted responses in seconds. Our approach included:

  • Developing an AI-powered draft response tool to handle routine questions in 10–15 seconds.
  • Conducting multiple rounds of testing on randomly selected questions to confirm accuracy and completeness.
  • Integrating AI into CMS’s workflow so experts could review and finalize responses for complex cases.
Key Outcomes

Faster answers, clearer guidance, and less administrative burden for CMS

Integrating AI into the response workflow helped CMS deliver timelier, more consistent support to hospitals seeking clarification on SEP-1. The improvements strengthened hospitals’ ability to apply the measure confidently and focus more time on patient care. As a result, hospitals reported:

  • 35 percent decrease in overall turnaround time, enabling faster resolution of hospital inquiries.
  • 18 percent increase in timely resolutions, with more questions answered within the same month they were submitted.
  • 2 percent decrease in reopened questions, reflecting clearer and more complete responses.

AI-generated drafts accelerated the process, while expert, human review ensured the accuracy and reliability hospitals depend on for compliance and care delivery.

Our Takeaway

AI can help federal agencies manage high-volume inquiries without sacrificing accuracy. By pairing automation with expert oversight, CMS can deliver timely, consistent guidance to hospitals—supporting better patient care and reducing administrative burden.

Efficiency Meets Impact.
That's Progress Together.

To solve their most pressing challenges, organizations turn to Mathematica for deeply integrated expertise. We bring together subject matter and policy experts, data scientists, methodologists, and technologists who work across topics and sectors to help our partners design, improve, and scale evidence-based solutions.

Work With Us