There is an overwhelming amount of data that is available to caregivers in an Intensive Care Unit (ICU). There are dozens of sensors and devices continuously streaming data about a patient’s vital organ systems and the EMRs that provide a steady stream of updates in the form of lab test results, clinical notes, etc. In even the most advanced ICUs, it is mostly up to the caregiver to analyze all this information to arrive at important clinical decisions.

Unscrambl has taken the next step. Unscrambl’s PULSE platform, which comes with pre-built adapters, a comprehensive data model, pluggable analytics and a decisioning framework with real-time visualization capabilities is already serving the real-time analytics needs of several leading hospitals in the United States. The ability to plugin new analytical models and share them with other institutions is paving the way for having a repository of analytic models that can augment clinical decisioning.


Raised Alert based on rules

Optimize alerting to minimize false positives


Predict Sepsis

Predict AFib


Phillips/GE Connectors


Historical Data
Reference Data
Streaming Data

Bedside Monitors


Continuous Feedback

There is an abundance of alerts and alarms in ICUs. This leads to the phenomenon of alarm fatigue, where the caregivers just ignore most of the alarms, sometimes leading to poor outcomes. We surely do not want to add to this noise. At the decisioning-level, we include as much context (e.g. patient’s existing condition, medication, surgery, etc.) required to reduce the number of alerts that our system generates. When it comes to recommending an action by means of an alert, the healthcare alert optimization robot acts as a guard. We utilize the feedback – both from the caregivers and the outcomes – into our decision-making process. As time passes by, the robot trains itself to issue only the alerts that will most likely be acted upon but at the same time ensures that critical alerts are always raised.

Key Features

^ Back to top