MACHINE LEARNING

Between clinical, behavioral, IOT devices, and genomics, data is growing at much faster rate than humans can comprehend.  We use advanced data analysis and machine learning techniques to transform data into actionable insight.

Pathway similarities

Compute scores for how similar/dissimilar two patients care pathways are based on specific criteria.  Cluster patients based on similarities

Risk Prediction

Compute risk scores for various conditions and incidents then identify key risk factors

Diagnosis & Problem Prediction

Predict likelihood of various diagnosis or problems weeks if not months in advance of condition

Program
Recommendations

Recommend health and wellness programs and services for patients based on historical data

Treatment Outcome Predictions

Predict likelihood of improved condition based on assigned treatment plan and adherence

anomaly Detection

Identify trends and criteria-based outliers to proactively focus your attention and mitigation plans

Text Medical Chatbot

Build context-aware intelligent conversation that leverages machine learning for conversation flow and buildup

SOLUTIONS PLATFORM

HOSTING
Our SaaS HIPAA-compliant platform can be hosted on AWS or Azure and supports multi-tenancy.
WEB 
Our web client supports rich interface using JavaScripts, ReactJS, KnockoutJS, and JQuery as well as D3JS and HighCharts
VOICE
We use chatbots and virtual assistants to augment our user experience for optimal performance

cloud environment

Our platform uses micro-services architecture to support high-scale and distributed workloads over large volume of data.

Domain name system for custom URL
lookup
Edge servers (CDN) for cached content and
faster access
Elastic content store on Amazon S3 or Azure Blobs
Load balancer to distribute and fail-over traffic
Web front-end and services back-end nodes
supporting multi-tenancy
Automatically scale front-end and back-end based on load
Multi-shard transactional databases with
high availability and disaster recovery
Big data store for searching, machine learning, and aggregations
1

Domain name system for custom URL lookup

2

Edge servers (CDN) for cached content and faster access

3

Elastic content store on Amazon S3 or Azure Blobs

4

Load balancer to distribute and fail-over traffic

5

Web front-end and services back-end nodes supporting multi-tenancy

6

Automatically scale front-end and back-end based on load

7

Multi-shard transactional databases with high availability and disaster recovery

8

Big data store for searching, machine learning, and aggregations

modeling and late binding

In order to achieve maximum flexibility, rapid solution design and development, and a high-degree of solution customization, we opted for a late-binding modeling approach that is based multiple layers of abstraction.

STORAGE MODELING

The first tier is a storage model the defines that physical structure of our healthcare data store.

ENTITY MODELING

The second tier is an entity model.  This is a logical over physical representation of all the key healthcare related entities used by solutions on our platform including their properties, relationships, and CRUD methods.

ANALYSIS MODELING

The third tier is a data expression, correlation, and filtering model.  This is a logical flow of data capturing key business concepts and rules.

VISUAL MODELING

The fourth tier is a visual representation leveraging the logical flow and rules through binding to physical data via the logical entity model.

SOLUTION MODELING

The fifth and final tier encapsulates the user experience, flow, and business logic of the application.  This includes auditing and role-based access control to views, operations, and content.

KEY FEATURES

Our platform provides a common set of capabilities to all solutions on our platform.

VOICE, MOBILE, and WEB INTERFACE

Build a highly flexible and configurable solution with a responsive and adaptable user experience based on situation and available medium.

WORKFLOWS AND TASKS

Define advanced processes and schedules with tasks that can be manual (human) or automated (system).  Tasks includes invoking methods as well as sending emails, texts, and voice notifications.

ALERTS

Define visual rules on various data elements to generate conditional events which can trigger mitigation workflows.

NOTIFICATIONS

Elevate various conditional events generated by the solution into notifications that is surfaced to the user for awareness or to take manual actions.

AUTHENTICATION, AUTHORIZATION, AND AUDITING

Perform advanced data redaction and interface restriction based on individuals and roles. Support single sign-on and audit all data access.

FEEDS AND CHANNELS

Share knowledge among users through public channels and private (group-based) feeds.

CHATS

Secure one on one or group-based chats and notes between users as well as care team.

our model

Identifying high-risk patients allows a providers to prioritize these patients for care coordination and targeted interventions. Predictive analytics avoids unnecessary hospitalizations incidents, and adverse health events. 

Using modern technology it's now easier to aggregate and correlate clinical data with other data sources, then apply deep learning models to predict risk and improve outcomes.

SIZE OF DATA SET

our models

Large sets of digital health care data

other models

Large sets of digital health care data

QUALITY OF DATA

OUR MODELS

Excellent – cleansed, normalized, validated
Natural language processing (NLP) to access unstructured data

OTHER MODELS

Poor data fidelity
No way to access unstructured data

SOURCES OF DATA

OUR MODELS

Clinical feeds from EMRs to lab results
Reference and research data/publications
Behavioral, activity, nutrition plans, and sensory IOT data

OTHER MODELS

Usually claims data
Usually inpatient data only
Manual input via assessments

HEALTHCARE MARKETPLACE

Marketplace will focus entirely on PHM and bioinformatics apps and services.

Users would be able to contribute new clinical content and models such as dashboards, reports, workflows, clinical pathways, risk measures and opportunity measures which can be plugged into new or existing solutions.

Developers would be able to contribute new solutions, new data visualizations, new solution services (APIs), deep learning modules, performance models, machine learning models, and new research data sets.

STARTUP ACCELERATOR

Entrepreneurs and students can leverage our hosted (AWS or Azure) platform and services to build web and mobile solutions leveraging our extensible application framework, data analytics, rich visualization, and data management capabilities.

Business model may be a combination of revenue share and/or equity stake.  Funding grants are also available for select ventures.