Feb 6, 2008
Published in: Journal of the American Geriatrics Society, v. 55, no. S2, Oct. 2007, p. S457-S463
Posted on RAND.org on January 01, 2007
OBJECTIVES: To use a formal decision-making strategy to reach clinically appropriate, internally consistent decisions on the application of quality indicators (QIs) to vulnerable elders (VEs) with advanced dementia (AD) or poor prognosis (PP). DESIGN: Using a conceptual model that classifies QIs principally by aim and burden of the care process, 12 clinical experts rated whether each Assessing Care of Vulnerable Elders-3 (ACOVE-3) QI should be applied in evaluating quality of care for older persons with AD or PP. QI exclusions were assessed for each of the 26 conditions and by whether these conditions were mainly medical (e.g., diabetes mellitus), geriatric (e.g., falls), or crosscutting processes of care (e.g., pain management). QI exclusions were also identified for older persons who decided against hospitalization or surgery. RESULTS: Of 392 ACOVE-3 QIs, 140 (36%) were excluded for patients with AD and 135 (34%) for patients with PP; 57% of QIs focusing on medical conditions were excluded from patients with AD and 53% from patients with PP, whereas only 20% of QIs for geriatric conditions were excluded from AD and 15% from PP. All QIs with care processes judged to carry a heavy burden were excluded; 86% of moderate-burden QIs were excluded from AD and 92% from PP. All QIs aimed at long-term goals were excluded; 83% of intermediate-term goal QIs were excluded from AD and 98% from PP. Individuals holding a preference to forgo hospitalization or surgery would be excluded from 7% of potentially applicable QIs. CONCLUSION: Measurement of quality of care for VEs with AD, PP, and less-aggressive care preferences should include only a subset of the ACOVE-3 QIs, largely those whose burden is light and whose goal is continuity or short-term improvement or prevention.