Artificial Intelligence

The Ibis AI Contextualizes Large Volumes of Data to Facilitate Decision-Making

The AI in Ibis is key to the program’s commercial success. The purpose of the AI is to streamline population health management. The AI’s job is to optimize care plan implementations to achieve the best health outcomes, for individuals and populations.

Plans: Planning, defined as sequencing strategies and actions to achieve specific goals, is a uniquely human function. Planning for healthcare has always been a human function. But as populations get older and more people need carefully designed care plans, there is a shortage of professionals available to create, implement and monitor plans. The AI in Ibis is designed to automate some aspects of care planning and implementation. The AI identifies action steps and strategies, for both individuals and populations, to achieve two goals: decrease hospital stays and lengthen aging in community.
On a daily basis, the AI plans the actions of the Member and the Ibis Care Team. The AI prompts for the actions, monitors the actions and learns their impact. The Member’s actions are displayed and tracked through the IbisHub.
A web portal displays prioritized action lists for all roles in the Ibis Care Team. All actions taken by the Care Team are documented and tracked through the web portal.

Prompts: The next step in the AI continuum process is to prompt for new observations (active or passive) and for actions from either the patient or members of the care team.

Observes: The AI gathers data from all the sources needed to anticipate health deteriorations based on prescribed protocols. What data is collected, when and how, is situational and personalized to each individual by the care team. A broad array of data is gathered to provide a uniquely comprehensive "snapshot" of the health of each member.

Learns: From the gathered data, the AI uses the baseline health parameters as determined by the care team correlated with the patterns of the Member's self-management. Deviations from expected patterns then trigger chains of clinically advised logic to mitigate downward trends. With this model, logic guides learning and learning triggers logic. The AI ascertains how often these patterns
arise, if persistent sequences exist, and if medication adherence had an impact.

Discovers: Once the baseline patterns of health and self-management are learned, the AI discovers the correlation between self-management and health. The AI enables the clinical team to assess how quickly health deteriorate when self-management declines and how quickly health recovers when self-management improves. The same correlations are learned from the patterns of health changes followed by patterns of intervention (e.g., outpatient visits, inpatient stays, coaching calls, socialization). This enables the clinical team to adjust individual thresholds and care plans.

Anticipates: Once the correlations between self-management, health, and interventions are known, the AI will produce alerts based on the quality of self-management based on the prescribed care plan.

Spectrum of Observations

Vital Signs	Symptoms	Actions
Blood Pressure	Coughing	Medications
Blood Glucose	Breathing	Meals
Pulse	Feeling	Execise
Temperature	Palpitation	Vitals Monitoring
O2 Saturation	Swelling	Symptom Checking
FEV1	Mood	Appointments
Weight	Depression Screening	Sleep
Spirometer Reading	Movement	Activities of Daily Living

A Bold Step in the Future

Spectrum of Observations