Romain Pirracchio
University of California-San Francisco, CA·Medicine & Clinical Research
Lung infection is the most common cause of sepsis, a dangerous whole-body reaction to infection, and accounts for over 40% of acute respiratory failure cases. Every hour of delay in treatment raises the risk of respiratory failure and death, which is why guidelines recommend an "hour-1 bundle": blood cultures, a lactate measurement, antibiotics, intravenous fluids and, if needed, blood-pressure-raising drugs (vasopressors) within the first hour.
- Fields
Medicine & Clinical Research
Public Health & Epidemiology
Computer Science
- Topics
sepsis
pneumonia
acute respiratory failure
- Grants
- 1 known grant, through May 2029
What they work on
- Train machine learning models on electronic health record data from three health systems to estimate each patient's individual treatment effect from antibiotics, fluid resuscitation and vasopressors in suspected community-onset lung sepsis.
- Externally validate these individual treatment effect models across the participating integrated health systems to test whether predictions transfer between hospitals.
- Test the resulting clinical decision support system prospectively and silently in real time, so predictions are generated without influencing bedside care.
- Design an EHR-embedded user interface that presents a treatment recommendation with a confidence level, and refine it iteratively using structured clinician feedback to support a future trial.
More about the research
Hospitals follow this only about 60% of the time, and patients respond very differently to each element, so giving all of it to everyone with suspected lung sepsis carries its own risks. Clinician judgment and existing early warning or sepsis-prediction algorithms cannot reliably tell who will benefit from which piece.
Why it matters
If validated, this decision support tool could help emergency and critical care clinicians start antibiotics, fluids or vasopressors quickly in the patients most likely to benefit while avoiding unnecessary exposure in others. For patients with severe pneumonia and sepsis, that could mean less organ dysfunction, fewer cases of respiratory failure and lower mortality, and for hospitals, fewer inappropriate interventions.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Anna Krupp
University of Iowa, IA·Medicine & Clinical Research
Develop a clinical decision support system prototype in the electronic health record that identifies ICU patients suitable for physical activity, using clinician interviews and user-centered design.
Shamim Nemati
University of California-San Diego, CA·Biomedical Engineering
Compare a large language model–supported differential diagnosis system against the binary sepsis model COMPOSER for accuracy, with expert clinician review of the diagnosis lists, using retrospective hospital data.
Alexander Friedman
Columbia University, NY·Medicine & Clinical Research
Measure how closely care for severe hypertension, obstetric hemorrhage, sepsis and abnormal maternal vital signs follows safety bundle recommendations, using EHR data from eight consortium hospitals.
The school
University of California-San Francisco: 89 labs on PI Pages, $2.1B a year on research, 164 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.