CDM Data Analytics Using a RIM approach Diane Gutiw PhD SAIC Canada Health Solutions Business Unit.

CDM Data AnalyticsUsing a RIM approach

Diane Gutiw PhD

SAIC Canada

Health Solutions Business Unit

CDM Data Analytics Discussion … The Business Problem Overview of CDM Workflow Overview of the need for Data Analytics Support in V3 messaging and RIM How to implement BioSense as potential solution for CDM

The CDM Business Problem

Multiple disease conditions Diabetes, CHF, COPD, HBP, Asthma, Renal, Depression, Arthritis,

Alzheimer's, Epilepsy, Hepatitis, CF, Eating Disorders, Obesity, Cancer, Anemia, AIDS/HIV etc.

May be overlap in conditions Treatment focus is disease management not cure CDM Clinician collaboratives Data sharing based on treatments and outcomes rather

than POS data Challenges of data collection Challenges of data standardization/normalization

Potential Data Sources

SHR RIM

SHR RIM

EMREMR

EMREMR

EMREMR

PharmacyPharmacy

PharmacyPharmacy

ADTADT

ADTADT

Drug Repository

Drug Repository

Lab Repository

Lab Repository

DI Repository

DI RepositoryLocation

RepositoryLocation

RepositoryProvider

RepositoryProvider

RepositoryClient

RegistryClient

RegistryImmunization

RepositoryImmunization

Repository

CDM Workflow

Support in Messaging

Supporting Messages: Lab Diagnostic Imaging Immunization Prescription ADT … Gaps identified in existing reporting for

CDM

Gaps in HL7 v3 Messaging

Use existing Registries/Repositories as “source of truth” CDM elements available in RIM Clinical data specific to CDM not in any existing RMIM Data required for CDM Analytics outside of existing repositories:

Clinical background (smoking cessation, diet, weight loss) all available in SNOMED

Concept of Collaborative Teams Includes combined assessment, referrals, treatment, hospitalizations

and adjusted treatment pattern for CDM Treatment refusals, testing negation indicators and lifestyle

adjustments/refusals critical to data analytics And then how to consolidate data and extract for “meaningful use”?

Benefits of a RIM-based solution Supports data modeling requirements Supports standardization Supports normalization Extends to multiple conditions, treatments,

outcomes Supports reporting needs

BioSense Solution

BioSense Solution:

To provide a technology platform which simplifies the enablement of public health reporting in a healthcare-centric environment, satisfying the requirements of HITECH and “meaningful use” objectives relative to Public Health.

BioSense Architecture

A toolset that can be run standalone, or can be integrated into an existing system or system interface.

An interface that allows loose coupling.  This allows it to be integrated into proprietary platforms such as Microsoft .NET.

A design to support Public Health reporting, and flexible enough to support future health reporting requirements.

Interoperable Hardware and software vendor independent. Able to collect appropriate and targeted

hospital/jurisdictional data.

Retail Pharmacy Rx

POS

Hospitals & Clinics

Secure Data NetworkProvides Certificate based authentication

services

BioSense System

12

IntegratorRhapsody

The Integrator “listens” to the source’s HL7 traffic and places

needed data on-line

Data Marts

Application specific, de-normalized &

decentralized data

Data BrokerRhapsody

Receives source data, un-batches and

transforms HL7 files to XML & send XML to

queue

Extract, Transform, Load (ETL)

InformaticaApply business rules to data and load Data

Warehouse

Data Warehouse

Centralized Data Repository

Secure Messaging

Provides a guaranteed message delivery service across the

internet

ApplicationSAS

WebLogicESRI

Bio-Intelligence Users

Public Health UsersAnd CDM Users

SAS Analytics

POS Labs

LifeRay Portal

Example: Stand-alone Solution for Surveillance

13

BIOSENSE

NATIONAL HEALTH SAFETY NETWORK

BIOSENSE ELR

Public Health Open Reporting System (PHORS)

PHORS leverages SAIC’s healthcare and HL7 messaging knowledge. The PHORS architecture is designed to allow Infection Control Practitioners (ICPs) to define detection rules which are then stored in a rules based engine and used by web services to identify patient data to assist ICPs in managing and reporting of Hospital Acquired Infections (HAI)

BioSense Features

14

Configurable by the end user You select the characteristics of the

records

Controllable by the end user Can require user approval prior to

reporting

Additional reporting available Can add other characteristics or reports for local needs

Supports a multitude of data feeds

Next Steps….

BioSense currently in production in US for Public HealthBioSense currently receives > 4 million clinical records dailyBatch (Not Real Time)

913 Veterans Affairs (VA) facilities 375 Department of Defense (DOD) facilities

Real Time117 facilities from 24 Hospital Systems536 facilities from 9 Syndromic Systems 2 National Reference Labs (LabCorp and Quest)>35,000 retail pharmacies (PerSe, McKesson RelayHealth)

BioSense to be presented to Canadian Public Health and CDM business users

BioSense to be adapted to meet CDM requirements as POC

Questions or follow up …

Diane Gutiw [email protected]