Using the AHRQ National Quality Measures

Clearinghouse to Assess Diabetes Quality

Measurement

Brian Leas, MS, MAUniversity of Pennsylvania Health System

AHRQ Annual ConferenceSeptember 19, 2011

Research Objectiveso Evaluate the state of diabetes quality measurement, utilization & impacto Determine key strengths, weaknesses, gapso Develop strategies to enhance value and impact of diabetes measures

Research Teamo Thomas Jefferson University School of Population Health

David Nash, Principal Investigator Kathryn Kash, Qualitative InterviewerNeil Goldfarb, Co-Investigator Rich Toner, Research CoordinatorAl Crawford, Measurement Analyst Brian Leas, Project ManagerBettina Berman, Measurement Analyst

Timeline: Spring 2008 – Summer 2009

Fundingo National Changing Diabetes Program of Novo Nordisk

Background

Environmental scan of diabetes quality measures◦ Measure developers: Accreditation agencies, payers,

physician groups, QI organizations, state/local initiatives◦ All types of measures: Process, outcome, safety, etc.◦ Specifications: Definitions, inclusion/exclusion, time frame◦ Scope: Broadly representative of key sectors, rather than

fully comprehensive

Key informant interviews◦ Measure developers, physicians, payers, employers, QI orgs◦ Representation from all relevant stakeholder groups◦ Interviewees expert in both technical development and

practical utilization of quality measures ◦ Semi-structured, recorded

Methodology

Step 1: National Quality Measures Clearinghouse (NQMC)◦ Keywords: “diabetes” and “diabetes mellitus”◦ 98 measures identified

Step 2: Supplemental search◦ Roster compiled of key organizations & known measure

sets◦ Reviewed websites, contacted organizations for

information◦ 48 additional measures identified

Step 3: Data abstracted into Excel spreadsheet

Search Strategy

Descriptive characteristics◦ Brief description◦ Clinical topic (e.g. HgbA1c mgmt, lipid profile, foot exam)◦ Developer◦ Release and revision dates◦ Broader measurement set◦ Website link◦ NQMC identification #

Specifications◦ Level of measurement (e.g. clinician, health plan)◦ Source of data (e.g. administrative, medical record, pharmacy,

lab, registry)◦ Numerator / denominator◦ Inclusion / exclusion◦ Time frame

Measure Abstraction

Validity◦ Supporting evidence◦ Extent of measure testing◦ Composition of development group◦ Endorsement by NQF and others

Policy-relevant domains*◦ AHRQ: Process, outcome, access, structure, use of

services, patient experience, population health◦ IOM: Effectiveness, efficiency, equity, patient-

centeredness, safety, timeliness

*Dimensions as of 2008

Measure Abstraction

Measure development, selection, endorsement

◦ National Committee for Quality Assurance◦ National Quality Forum◦ American Medical Association

◦ Centers for Medicare and Medicaid Services◦ Ambulatory Quality Alliance◦ Pharmacy Quality Alliance◦ American Board of Medical Specialties◦ American Board of Internal Medicine◦ American College of Physicians◦ Wisconsin Collaborative for

Healthcare Quality

Measure utilization and value

◦ Blue Cross Blue Shield Association◦ Centers for Disease Control and Prevention◦ Institute for Healthcare Improvement◦ America’s Health Insurance Plans◦ American Academy of Family Physicians◦ Nat’l Business Coalition on Health◦ American Diabetes Association◦ American Association of Clinical Endocrinologists◦ American Association of Diabetes Educators

Key Informant Interviews

Scope of Measuremento Nearly 150 measures in > 25 distinct clinical categorieso Primarily process measureso Substantial redundancyo Widely varying specificationso Evidence base often weak or unclear

Measurement Gapso Primary preventiono Outcomes of careo Patient perspectiveso Coordination of careo Access to care

Findings

Highly useful for identifying relevant measures◦ Although 1/3 of measures in our study had not been included

in NQMC, many were later submitted by their respective authors

◦ Sufficiently specific and sensitive◦ Capability to stratify by numerous categories◦ Only known resource for identifying and comparing measure

sets from different sources

Limitations◦ Time lag from development to submission to posting◦ Relies on self-reporting◦ Until recently, limited data available on evidence base,

validity, conflicts of interest

Perspectives on NQMC

Brian LeasResearch Analyst

Center for Evidence-based PracticeUniversity of Pennsylvania Health System

Brian.leas@uphs.upenn.edu

Questions?