Appropriateness of quality indicators for older patients with advanced dementia and poor prognosis.

OBJECTIVES

To evaluate the applicability of process-of-care quality indicators (QIs) to vulnerable elders and to measure the effect of excluding indicators based on patients' preferences and for advanced dementia and poor prognosis.

DESIGN

The Assessing Care of Vulnerable Elders (ACOVE) project employed 203 QIs for care of 22 conditions (including six geriatric syndromes and 11 age-associated diseases) for community-based persons aged 65 and older at increased risk of functional decline or death. Relevant QIs were excluded for persons deciding against hospitalization or surgery. A 12-member clinical committee (CC) of geriatric experts rated whether each QI should be applied in scoring quality of care for persons with advanced dementia (AdvDem) or poor prognosis (PoorProg). Using content analysis, CC ratings were formulated into a model of QI exclusion. Quality scores with and without excluded QIs were compared.

SETTING

Enrollees in two senior managed care plans, one in the northeast United States and the other in the southwest.

PARTICIPANTS

CC members evaluated applicability of QIs. QIs were applied to 372 vulnerable elders in two senior managed care plans.

MEASUREMENTS

Frequency and type of QIs excluded and the effect of excluding QIs on quality of care scores.

RESULTS

Of the 203 QIs, a patient's preference against hospitalization or surgery excluded 10 and eight QIs, respectively. The CC voted to exclude 81.5 QIs (40%) for patients with AdvDem and 70 QIs (34%) for patients with PoorProg. Content analysis of the CC votes revealed that QIs aimed at care coordination, safety or prevention of decline, or short-term clinical improvement or prevention with nonburdensome interventions were usually voted for inclusion (90% and 98% included for AdvDem and PoorProg, respectively), but QIs directed at long-term benefit or requiring interventions of moderate to heavy burden were usually excluded (16% and 19% included, respectively). About half of QIs aimed at age-associated diseases were voted for exclusion, whereas fewer than one-quarter of QIs for geriatric syndromes were excluded. Thirty-nine patients (10%) in our field trial held preferences or had clinical conditions that would have excluded 68 QIs. This accounted for 5% of all QIs triggered by these 39 patients and 0.6% of QIs overall. The quality score without exclusion was 0.57 and with exclusion was 0.58 (P =.89).

CONCLUSION

Caution is required in applying QIs to vulnerable elders. QIs for geriatric syndromes are more likely to be applicable to these individuals than are QIs for age-associated diseases. The objectives of care, intervention burdens, and interval before anticipated benefit affect QI applicability. At least for patients with AdvDem and PoorProg, identification of applicable or inapplicable QIs is feasible. In a community-based sample of vulnerable elders, few QIs are excluded.