Objective Sepsis Surveillance Using Electronic Clinical Data.

Area of Science:

• Critical care medicine
• Health informatics
• Epidemiology

Background:

• Severe sepsis surveillance is crucial for public health.
• Traditional methods using claims data may lack accuracy and stability.
• Electronic health records (EHRs) offer a potential alternative for real-time data capture.

Purpose of the Study:

• To compare the accuracy of severe sepsis surveillance using EHR clinical data versus claims data.
• To evaluate incidence and mortality trends of severe sepsis using both surveillance methods.

Main Methods:

• Developed an EHR-based severe sepsis definition using clinical indicators (infection, organ dysfunction).
• Validated the EHR definition's accuracy and stability against a claims-based definition via chart review.
• Compared incidence and mortality trends from 2003-2012 using both EHR and claims data.

Main Results:

• The EHR clinical data definition demonstrated stable and high sensitivity over time (77-80%).
• Claims data sensitivity increased but remained lower (52-67%).
• EHR data showed modest incidence increase (7.7%) and mortality decline (1.7%), contrasting sharply with claims data (72% incidence rise, 5.4% mortality decline).

Conclusions:

• EHR clinical data provides a more sensitive and stable method for severe sepsis surveillance compared to claims data.
• Electronic surveillance enables more reliable estimation of severe sepsis burden and trends.
• This approach can improve understanding and management of sepsis at a population level.