Cognome Collaboratives
Knowledge Sharing and Problem Solving with Peers
Open Collaboratives for Patient Safety & Quality, Healthcare Operations and Clinical Research
Models, Implementation and Support services provided by Cognome
Collaboratives for AI & ML Practitioners
Focused Collaboratives for model development, testing, training & deployment. Available to AI & ML practitioners implementing or planning to implement models at their hospital or health system.
Collaboratives also provide an opportunity to join our Federated Training Data Network. Build highly intelligent and responsible ML models by training them on complementary data sets with other health system partners.
Join now for access to all Collaboratives For a limited time, join any Collaborative to gain access to all open Collaboratives
1) Patient Safety & Quality Improvement
2) Healthcare Operations & Revenue Cycle
3) Clinical Research & Data Science- Open access all models - including visibility into how the models are built and how they make their predictions.
- You do not have to use Cognome's models to join our Collaboratives...bring your own model or none at all, if you just want to learn and network with peers.
- Deployed models will be (re)trained on your data. You can also join our Federated Training Data Network to further optimize performance while creating more responsible models.
- Models are Large Language Models (LLMs) or predictive Machine Learning (ML) Models - and sometimes both.
- Peer & Partner Support: collaborate with your peers as well as Cognome's data scientists, informaticians and developers.
- Access to Cognome's professional services team who can assist with your install, integration and optimization efforts
A full list of all the models can be found here. And we are adding more models and collaboratives all the time.
Patient Safety & Patient Quality
These life saving models have a direct impact on your ability to improve safety and quality outcomes. Examples include:
Sepsis Early Warning Predictor: Provide your clinical staff with an early warning system with our Sepsis Prediction model designed to ensure the lowest false positive rates in the industry. This model is part of our Sepsis Solution Bundle which includes the Predictor, Sepsis Real-Time Surveillance model and a Sepsis LLM for regulatory reporting. Save lives, reduce alert fatigue, reduce manual reporting by 90%+ and increase clinician adoption. Peer reviewed here.
Predict Readiness for Surgery: This model assigns ASA-PS classifications uniformly across all patients so you can clearly evaluate whether or not your patient is ready for surgery. Peer reviewed in the Journal of Clinical Anesthesia.
Respiratory Failure Predictor: This early recognition model predicts ARDS allowing care teams to intervene sooner and prevent poor or worsening outcomes. Peer reviewed in the National Library of Medicine.
Clinical Research and Clinical Trials
Our research collaborative will show you how to take your data and make it available for research of all kinds. Both for your own team, and more broadly with others.
Cognome has two sets of products in this collaborative, the first, Cognome I/O takes data from across your ecosystem and makes it available in OMOP format. The second, Cognome Search allows cohorts to be created (or uploaded) and studies to be made based on an almost limitless number of inclusion and exclusion criteria.
These two tools come integrated with your institutions IRB; with built in NLP; the ability to de-identify data sets both permanently and in real time; and can limit data availability based on who is doing the research. This means that data can be fully identified for clinicians who are working with their own patients, or de-identified for local clinicians who are doing research, or shown in aggregate format only for Real World Evidence studies.
All data is hosted in your environment either on-premesis or in your cloud.
Clinical Trial Matching: Give this model access to your patient charts and it will show you which ones qualify for clinical trials and it will write up a description of why the patient fits giving a full analysis of their inclusion and exclusion criteria match.
Healthcare Operations & Revenue Cycle
Cost reduction and revenue retention without sacrificing quality or negatively affecting outcomes is an imperative across the healthcare ecosystem. The models in this collaborative reduce cost from manual and often redundant tasks, especially for nurses and doctors, while improving the health system's ability to retain or increase revenues and provide clinicians the ability to operate at the top of their license. Examples include:
Case Cancellation: This model predicts patients that are "no show" risks for an appointment or procedure. Take action to make sure that patients with a high "no show" score are able to make it to appointments or get additional services to continue their treatment. A single clinic at one of our clients saw $500K increase in revenue during the first 6 months alone.
AutoChart: is both a machine learning model plus an LLM. It is the foundation for several of our operational models. Train AutoChart to do any chart abstraction that is currently being performed manually. A collaborative member reduced nurse chart abstraction hours for American College of Surgeons' NSQIP registry by 90%. With the ability to work with structured and unstructured data, AutoChart not only saves time and money but as importantly removes a significant cause of burnout among clinical staff. This model can be utilized for a wide range of high ROI use cases.
Length-of-Stay (LoS) Predictor: Highly accurate ML model to predict the estimated length of stay for patients. More accurately plan for patient discharge while delivering a better patient and family experience. Proactively address medical, behavioral health, and SDOH issues. The algorithm utilizes insights and variables known to affect length of stay in inpatient, critical care and perioperative settings