Transcript IMPROVE HF Primary Results

Associations Between
Outpatient Heart Failure Process
of Care Measures and Mortality
Gregg C. Fonarow, Nancy M. Albert, Anne B. Curtis,
Mihai Gheorghiade, J. Thomas Heywood, Mark L. McBride,
Patches Johnson Inge, Mandeep R. Mehra, Christopher M. O'Connor,
Dwight Reynolds, Mary N. Walsh, Clyde W. Yancy
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Disclosures
• The IMPROVE HF registry is sponsored by Medtronic
• The sponsor had no role or input into the selection of endpoints or
quality measures used in the study
• Outcome Sciences, Inc, a contract research organization,
independently performed the practice site chart abstractions for
IMPROVE HF, stored the data, and provided benchmarked quality of
care reports to practice sites. Outcome Sciences received funding from
Medtronic.
• Individually identifiable practice site data were not shared with either
the steering committee or the sponsor
• Individual author disclosures are provided in the manuscript
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Heart Failure Care in the Outpatient
Cardiology Practice Setting
• There are well documented gaps, variations, and disparities in
the use of evidence-based, guideline recommended therapies
for heart failure (HF) in inpatient and outpatient care settings.
• IMPROVE HF showed a performance improvement program
can increase the use of guideline recommended HF therapies
in the outpatient setting.
• It is assumed that use of process based performance measures
are associated with improved clinical outcomes; however that
has not been evaluated for current or emerging outpatient HF
measures.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
IMPROVE HF Outpatient Process Measures
Yancy CW, et al. Circulation. 2005;112:154-e235.
Bonow RO, et al. J Am Coll Cardiol. 2005;46:1144-1178.
Study Objectives
To examine associations between patient level process
measures and patient level survival for each performance
measure and two summary measures (total composite score
and all-or-none care).
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Statistical Methods
• For primary analysis, patients who were eligible for treatment but not treated at
baseline and who crossed over within first 12 months of the performance initiative
were excluded from each measure
• The composite score for each patient was calculated as the sum of individual
quality measure numerators divided by the sum of individual quality measure
denominators for which patient was eligible
• The all-or-none care measure was defined for each patient in terms of whether
they received all individual measures for which they were eligible
• Process-of-care measure conformity at baseline stratified by vital status at 24
months was summarized and compared in patients alive vs. those who died by
Chi-square test or t test
• For each quality measure, composite score and all-or-none care measure,
Generalized Estimating Equation (GEE) methodology was used to estimate
unadjusted and adjusted relationships between each process measure and
patient-level mortality in first 24 months. The GEE models accounted for
correlation of patients within practices.
• Appropriate clinical and/or statistical meaningful baseline patient demographic and
clinical characteristics and practice characteristics were included in the
multivariate GEE model for adjusted odds ratio estimation
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Patient Population
• To be enrolled in IMPROVE HF, patients had to have heart
failure or post-myocardial infarction left ventricular dysfunction
with left ventricular ejection fraction of 35% or less.
• There were 15,177 patients from 167 cardiology and
multispecialty practices in the US evaluated at baseline and
enrolled in the longitudinal cohort.
• At the 24 month follow-up 11,621 of the 15,177 patients (76.6%)
had documentation of vital status.
• A total of 2507 patients (16.5%) were lost to follow-up and 1048
(6.9%) were seen in practices (n=12) that did not complete the
follow-up assessment.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
IMPROVE HF Patient Characteristics
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Patient Characteristics (Continued)
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
IMPROVE HF Practice Characteristics
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Patient 24 Month Follow-up Mortality
• At 24 months, 2569 of the 11,621 patients (22.1%) with complete
vital status had died.
• Patients who died were significantly more likely to have ischemic
HF origin and comorbidities, including diabetes mellitus,
hypertension, chronic obstructive pulmonary disease, peripheral
vascular disease, and depression.
• Statistically significant differences were also evident for
laboratory evaluations, including BUN and creatinine, with higher
levels in patients who had died during the 24-month follow-up.
• The baseline process measure conformity was significantly lower
among patients who died compared with those who survived for
5 of the 7 individual measures.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Baseline Measure Conformity:
Alive vs. Dead at 24-Month Follow-Up
The baseline process measure conformity was significantly lower among patients who died
compared with those who survived for 5 of 7 individual measures.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Association Between HF Process Measures and
Mortality: Univariate and Multivariate GEE Models
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
6 of 7 Process Measures Associated with Reduced
Mortality
Mortality
Adjusted Odds Ratios with 95% CI Displayed
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Composite Measures Associated with
Reduced 24 Month Mortality
• Each 10% improvement in composite care was associated with a
13% lower odds of 24-month mortality (adjusted odds ratio, 0.87;
95% confidence interval, 0.84 to 0.90; P<0.0001).
• The adjusted odds for mortality risk for patients with conformity to
each measure for which they were eligible was 38% lower than for
those whose care did not conform for 1 or more measures for which
they were eligible (adjusted odds ratio, 0.62; 95% confidence
interval, 0.52 to 0.75; P<0.0001).
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Results Summary
• Baseline process measure conformity was significantly lower
among patients who died compared to those who survived for
5 of 7 measures (ACEI/ARB, beta-blockers, anticoagulation for
atrial fibrillation, ICD, CRT).
• Baseline process measure composite score was 70.0% for
patients alive at 24 months compared to 63.4% for those who
died (p < 0.0001).
• Adjusted odds ratio for mortality risk for patients with
conformity to all eligible measures was 38% lower than those
without conformity to one or more eligible measure.
• Every 10% improvement in composite care was associated
with a 13% lower odds of 24 month mortality risk.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Study Limitations
• Patient eligibility and utilization rates were determined by accuracy and
completeness of medical records and their abstraction
– Reasons for preventing treatment may not have been documented
– NYHA was not quantified in many records and instead was based on qualitative
description of the patient’s functional status
• Potential for ascertainment bias
– Self-selected cardiology practices, primary care setting not included
• Not randomized – secular trends may have influenced results
• Follow-up not available for all patients
• Study analyzed medications prescribed rather than patient adherence
• Associations between care processes and outcomes do not determine causality
• Did not assess health-related quality of life, symptom control, functional capacity,
patient satisfaction, hospitalization rates, or other clinical outcomes that may be
of interest
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Conclusions
• These data are among the first to demonstrate that
adherence to HF process measures for ACEI/ARB, beta
blocker, anticoagulation for atrial fibrillation, and HF
education is significantly associated with survival in
outpatients with heart failure.
• Process measures for ICD use and CRT use could also
be considered for inclusion in HF outpatient performance
measure sets.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Clinical Implications
• These HF process measures appear to discriminate the
quality of HF care at the patient level and may be useful
for assessing and improving HF care.
Fonarow GC, et al. Circulation. 2011;123(15):1601-1610.
Back-Up Slides
IMPROVE HF Study Overview
• Largest, most comprehensive performance improvement study
for HF patients in the outpatient setting
• Designed to enhance quality of care of HF patients by facilitating
adoption of evidence-based, guideline-recommended therapies:
– Evaluate utilization rates of evidence-based, guideline-recommended
HF therapies at baseline and over the course of the performance
improvement intervention (chart audit and feedback; use of
performance measures)
– Multifaceted, practice-specific performance improvement toolkit
including clinical decision support tools (reminder systems)
– Sites attended an educational workshop to set treatment goals and
develop a customized clinical care pathway (educational outreach)
Fonarow GC, et al. Am Heart J. 2007;154:12-38.
Methods: Guideline-Recommended
Quality Measures
• Seven quality measures with strong evidence prospectively selected:
– Angiotensin-converting enzyme inhibitor (ACEI) or angiotensin II receptor blocker
(ARB)*
– ß-blocker*
– Aldosterone antagonist
– Anticoagulation therapy for atrial fibrillation/flutter (AF)*
– Cardiac resynchronization therapy with or without ICD (CRT)
– Implantable cardioverter defibrillator with or without CRT (ICD)
– Heart failure (HF) education*
• Patients deemed eligible for individual quality measure based on
meeting guideline criteria, without contraindications, intolerance, or
other documented reasons for non-treatment
• Steering committee selected quality measures based on potential to
improve patient outcomes, definition precision, construct and content
validity, feasibility
* Included as ACC/AHA outpatient HF performance measure, endorsed by National Quality Forum.
Fonarow GC, et al. Circulation. 2010;122:585-596.
Methods: Patient Selection, Practice
Selection, Data Collection and Management
• Patient Inclusion:
– Clinical diagnosis of HF or prior MI
with at least 2 prior clinic visits within
2 years
– LVEF ≤ 35% or moderate to severe
left ventricular dysfunction
• Patient Exclusion:
– Cardiac transplantation
– Estimated survival < 1 year from
non-cardiovascular condition
• Average of 90 eligible patients per
practice randomly selected for each
of 3 study cohorts
• Practices: Outpatient cardiology
(single specialty or multi-specialty)
practices from all regions of the
country
Fonarow GC, et al. Circulation. 2010;122:585-596.
• Data quality measures
– 34 trained, tested chart review
specialists
– Training oversight by study steering
committee members
– Monthly quality reports
– Automated data field range, format,
unit checks
• Chart abstraction quality
– Interrater reliability averaged 0.82
(kappa statistic)
– Source documentation audit sample
concordance rate range of 92.3% to
96.3%
• Coordinating center: Outcome
Sciences, Inc. (Cambridge, MA)
– Individual practice data not shared
with sponsor or steering committee
Methods: Study Design and
Patient Disposition
• Patients who were eligible for treatment but not
treated at baseline and who crossed over
within the first 12 months of the intervention
were excluded from each measure
Fonarow GC, et al. Circulation. 2010;122:585-596.
Methods: Practice Specific Performance
Improvement Intervention
* Use or participation was encouraged but not mandatory. Practices could adopt or modify tools.
Practice Survey:
• 96% adopted one or more performance improvement strategies
• 85% used benchmarked quality reports
• 60% employed one or more IMPROVE HF tools
Fonarow GC, et al. Circulation. 2010;122:585-596.
IMPROVE HF Performance Intervention:
Benchmarked Practice Profile Report
IMPROVE HF
Performance Improvement Tools
• As part of an enhanced treatment plan, IMPROVE HF
provided evidence-based best-practices algorithms,
clinical pathways, standardized encounter forms,
checklists, pocket cards, chart stickers, and patient
education and other materials to facilitate improved
management of outpatients with HF
• The materials can be downloaded from
www.improvehf.com
• The materials are also included in the Circulation onlineonly Data Supplement
Fonarow GC, et al. Circulation. 2010;122:585-596.
IMPROVE HF Practice Specific Education
and Implementation Tools
Evidence Based Algorithms
and Pocket Cards
Clinical Trials and
Current Guidelines
Clinical Assessment and
Management Forms
www.improvehf.com
Patient Education
Materials
Dissemination of best practices:
• Webcasts
• Online Education
• Newsletters