Contribute to prospective research
Help study ARDS phenotypes and outcomes across different healthcare settings.
An international research network & prospective observatory
Understanding ARDS
across different
patients and settings.
POLARIS brings together ICU teams to understand ARDS across different patients and healthcare systems. We collect standardized clinical data and develop collaborative research on phenotypes, care practices, and patient outcomes.
for Patient and Disease Phenotyping
For ICU teams
Why join POLARIS?
Build research together
Participating centers can propose studies and choose which projects to join.
Represent your patients and setting
Bring your ICU's experience into an international network of academic, community and private hospitals.
Getting started
How to join
- Express interestTell us about your ICU using the registration form.
- Speak with the coordinating teamDiscuss the study workflow and the right local approval pathway.
- Prepare locallyComplete required approvals and arrange data collection at your site.
- Begin recruitmentRecruit once your site is ready and open to enrollment.
Questions from new centers
Which ICUs can join?
Adult academic, community and private ICUs can express interest, regardless of size or resource level.
What happens to clinical care?
POLARIS is observational. The study does not allocate treatment; clinical decisions remain with the care team.
Who helps with local approvals?
The coordinating team discusses the approval pathway and data collection workflow with each interested center. Contact the team →
November 2026
Beginning at BIDMC.
Growing together.
POLARIS recruitment is open at Beth Israel Deaconess Medical Center in Boston. Other centers join progressively as their local approvals and data collection systems become ready.
Follow recruitment across the network →Our first prospective study
One clinical syndrome.
Different biological responses.
Acute respiratory distress syndrome brings together patients with different injuries, inflammatory responses and clinical trajectories. This heterogeneity may help explain why the same treatment does not have the same effect in every patient.
Hyperinflammatory and hypoinflammatory phenotypes describe differences in systemic inflammation. Clinical classifiers offer a way to estimate these phenotypes using vital signs, laboratory results and ventilatory variables available in routine care.
These measurements also reflect local practice. A classifier developed in one setting must be tested prospectively in others, where patients, treatments and missing data differ.
18 routine variables
Clinical and laboratory measurements
Inflammatory phenotypes
Estimated with AI-CLARITY
International validation
Prospective evaluation across ICU settings
Prospective phenotyping
From retrospective findings
to prospective evaluation.
Our first study asks whether prospectively assigned inflammatory phenotypes are associated with 28-day mortality.
It also assesses transportability and examines whether associations with early intravenous corticosteroids differ by phenotype within a prespecified causal framework.
Read the scientific approachObservation within routine care
- Day 0
- Baseline data and phenotype assignment
- Days 3 & 7
- Repeated assessment and clinical course
- Day 28
- Vital status
No treatment is allocated by POLARIS.
An investigator-initiated program
A network built for shared research.
Academic, community and private adult ICUs are welcome. Join the network, contribute to prospective data collection, and help develop future collaborative projects.