Using Clinical Decision Support to Measure Quality for Special Populations T. Bruce Ferguson Jr. MD East Carolina Heart Institute Chair, Dept. of Cardiovascular.

Using Clinical Decision Support to Measure Quality for Special

Populations

T. Bruce Ferguson Jr. MDEast Carolina Heart Institute

Chair, Dept. of Cardiovascular Sciences

Brody School of Medicine at ECU

• design for implementation a new IT platform to support a Longitudinal CVD Information System (LCIS)• address disparities in CVD, in the unique safety-net Charity population in Louisiana.

PURPOSE

AHRQ THQIT Grant, 2004-2005

SCOPE• multi-institutional project within 8-hospital Charity Hospital system in LA• Collaboration Partnership between:

• LSU Schools of Medicine and Public Health• Tulane University Schools of Medicine and Public Health• ARMUS Corporation (IT provider)• LA Department of Health and Hospitals, Office of Public Health

METHODS•Our focus was directed at augmenting existing resources and creating an LCIS for collecting, analyzing, and coupling clinical and financial data to assess the medical and financial care effectiveness in this CHF population.

•Technology issues centered on:• longitudinal data collection methodology across

system• Integration of clinical longitudinal data into

repository/registry for analysis

•This LCIS would enable us addressing the significant care-delivery and patient-related disparities within this population and setting.

RESULTS

A model prototype of the LCIS was developed, including testing and evaluation components.

A patient centric prototype enabled multiple providers to collect longitudinal CHF data from patient encounters within multiple care-delivery settings.

Prototype testing and refinement was being undertaken at the time hurricane Katrina devastated LSU and Tulane Schools of Medicine, and disrupted forever the Charity safety-net population and system.

RESULTSSecured ASP

CHF Clinical Registry

(Outcomes)

Point of Care Patient Centric Data Collection(Nota Medica)

NM and Outcomes Integration

CLINICAL DECISION SUPPORTIn CVD, long history with retrospective analysis of clinical data:

Ferguson TB Jr. et al. Ann Thorac Surg 2002; 73:480-490

IMPROVEMENTIN

QUALITYOF

CARE

AHRQ HS 10403-07: CQI in Medicine

Ferguson, TB Jr. JAMA 2004

SUSTAINABILITY

USED CLINICAL DATA SYSTEMS TO ESTABLISH THE INFRASTRUCTURE FOR CONTINUOUS QUALITY IMPROVEMENT WITH PROCESS AND OUTCOMES DATA

Center for Health Services Research and DevelopmentEast Carolina University

1998, All-Cause, Age-Adjusted, per 10,000 Population

YLL Due To CVD:YLL Due To CVD: Eastern North Carolina would rank Eastern North Carolina would rank

50th50th

825.9 – 922.7

648.6 – 740.3

740.4 – 825.8

1,067.1 – 3,254.3

0.0 – 648.5

YLL-75 per 10,000 < 75

922.8 – 1,067.0

• Morbid ObesityMorbid Obesity• DiabetesDiabetes• CholesterolCholesterol• High Blood PressureHigh Blood Pressure• Heart AttacksHeart Attacks• StrokesStrokes

Care Delivery OrganizationUniversity Health Systems

of Eastern North Carolina

Pitt County Memorial Hospital

745 Bed private hospital

95 – 100% Occupied

Regional affiliated hospitals (11)

Brody School of Medicine at ECU

3rd largest component of UNC

System

East Carolina Heart Institute@ PCMH & BSOM

ECHI Clinical Cardiovascular Information System

Critical drivers (IOM, payers) to force information into longitudinal, patient-centric, value-based focus

CV Information – Current Statusat BSOM and PCMH

CrusadeDX

STS NCD ACC/PCI ACC/ICD Vascular CHF Others

Web SMS EncompassQMI NRMI

Clinical Databases

Hospital Information Systems Registries

Logician

Outpatient

Stand-alone systems, Requires local IT Support and Backup, Limited analytical, reporting capabilities

Integration of domains of clinical information silos into common platform interfaced with common analytic engine

East Carolina Heart Institute

Analytical platform encompasses all major cardiovascular data through common Web Access portal and entry mechanism, agnostic of the individual “Database” or “Registry”

Integrated Clinical CV Dataset provides for cross-DB analysis and reporting

CLINICAL DECISION SUPPORT

Technologies: Data Repository • Data set independent design • Meta data• Flexible and Expandable

Applications • Independent of any data set and database

engine• Object Oriented Design• Visual query, data and summary tables• Ad-hoc analysis• Matching Algorithm (99.7%)

Presentation Layer

Repository

Clinical Data

(STS, Cath/PCI, ICD, etc)

Financial Data

(UB 92, Claims, etc)

PreOPRisk Analysis

PreOPConsultation

Performance Reviews, CQI Measures, Clinical and Financial Reporting

Data C

ollectio

nA

nalysis

Presen

tation

Match

ing

CLINICAL DECISION SUPPORT

CLINICAL DECISION SUPPORTAnalyze Cost, Complications and Mortality by Predicted Risk Groups

CLINICAL DECISION SUPPORT Clinical Outcomes Risk Calculation

Clinical Decision Support in CVD