A Trial of Automated Outbreak Detection to Reduce Hospital Pathogen Spread

Area of Science:

• Healthcare epidemiology
• Infectious disease surveillance
• Public health informatics

Background:

• Current hospital outbreak detection relies on labor-intensive methods.
• The effectiveness of automated statistical surveillance in reducing outbreak size is not well-established.

Purpose of the Study:

• To evaluate if automated statistical surveillance of microbiology data can lead to earlier containment of healthcare-associated pathogen outbreaks.
• To assess the impact of automated surveillance on reducing the number of additional outbreak cases.

Main Methods:

• A cluster-randomized trial involving 82 community hospitals was conducted.
• Hospitals were randomized to either standard outbreak response or standard response plus automated statistical surveillance.
• The primary outcome was the number of additional cases post-outbreak detection, comparing intervention and baseline periods, including a pre-pandemic and pandemic analysis.

Main Results:

• Automated surveillance did not significantly reduce additional outbreak cases overall (difference-in-differences: 0.78, P=0.46).
• A significant reduction in additional cases was observed during the pre-pandemic period (difference-in-differences: 0.36, P=0.13).
• No significant reduction was observed during the COVID-19 pandemic period (difference-in-differences: 0.94).

Conclusions:

• Automated statistical surveillance did not reduce overall hospital outbreak size during the COVID-19 pandemic.
• The system demonstrated potential for reducing outbreak size in a pre-pandemic setting.
• Further research may be needed to optimize automated surveillance systems in diverse healthcare settings.