Adding Clinical Data to Administrative Datasets: Overview of AHRQ Pilots

Roxanne Andrews, Anne Elixhauser

13th Annual Healthcare Cost and Utilization Project (HCUP) Partners Meeting

November 18, 2009

Limitations of Administrative Data for Quality Measurement

• Lacks clinically important information
• Limited to ICD-9-CM diagnosis codes
• Missing physiological data (lab values and vital signs)
• May not include present on admission (POA) indicator for diagnoses
• POA distinguishes conditions at admission (comorbidities) from conditions that arose during a stay (complications)

AHRQ's Adding Clinical Data Project

• Purpose:
• Establish the feasibility of linking clinical and administrative data
• Develop a reproducible approach for joining clinical and administrative data
• Enhance capabilities of state data organizations
• Set the stage for future integration of clinical and administrative data streams

Two Types of Contracts

In-depth Pilots

• Add or link hospital clinical information to administrative data
• September 2007 - September 2009

Planning Contract

• Not ready for pilot but want to investigate linkage
• September 2007 - March 2009

Awards to Statewide Data Organizations

Pilots

• Florida Agency for Health Care Administration
• Minnesota Hospital Association (MHA)
• Virginia Health Information(VHI)

Planning

• Washington Center for Health Statistics, State Department of Health

AHRQ Adding Clinical Data Pilot & Planning Awards

Map of the United States. Florida, Minnesota, and Virginia are highlighted as pilot states. Washington is highlighted as a planning state.

Project Requirements

• Identify and select clinical data elements
• Add POA if not already collected
• Extract clinical data from electronic format
• Transfer data from at least 5 hospitals
• Link clinical and administrative data
• Create a multi-hospital database
• Collaborate with stakeholders

Peer-to-Peer Learning Network

• Monthly conference calls
• Supported by Thomson-Reuters and NASHP
• Included California staff working on regs
• Annual in-person meet
• Purpose:
• Learn from and collaborate with peers
• Understand, anticipate, and resolve implementation hurdles
• Forge solutions for other data organizations
• Share tools and materials

Pilot Project Activities

• Developed project informational mat
• Recruited hospitals
• Identified clinical data elements to add
• Standardized lab data using Logical Observation Identifiers Names and Codes (LOINC)
• Provided education and feedback on quality of POA coding
• Developed processes for data transmission, lab coding, data collection, and linking
• Provided technical assistance to hospitals

Project Successes

Challenges and Lessons Learned

• Project initiation
• Data standards and transmission
• Communication
• Stakeholders
• Hospitals
• Data analysis

Next Steps for Pilot Organizations

• Continue analyzing data
• Develop reports and materials for volunteer hospitals
• Identify staffing, time, and funding needed for continuation of collection and analysis (sustainability)

Next Steps for AHRQ

• Disseminate information about the pilots
• Develop online toolkit based on the pilots' expereinces
• Obtain Partner input at lunch today
• Obtain Partner input through an information-gathering conference call
• Create materials and templates for Partners

Some Suggested Next Steps for AHRQ

• Develop materials on the business case for adding clinical data
• Estimate potential costs of adding clinical data
• Facilitate Partner education on LOINC
• Host vendor roundtable on LOINC

What suggestions do you have?