Measuring Quality of Care for Nursing Home Residents — Considering Unintended Consequences

BACKGROUND

Public accountability has been called the “third revolution” of medical care.¹ At face value, measuring the quality of medical care seems an important fundamental step to improve quality. Omnibus Budget Reconciliation Act 1987 required all nursing homes that receive federal funding to complete the Resident Assessment Instrument (RAI).² The RAI composes the Minimum Data Set, which, as of 1998, must be sent in electronic form to each state health department, which forwards those files to the Centers for Medicare and Medicaid Services (formerly known as the Health Care Financing Administration). In response to concerns about suboptimal care in nursing homes, outcome indicators are now being used to guide state inspection of nursing homes and to provide consumers with information on nursing home choice at an internet site entitled, “Nursing Home Compare” (see www.medicare.gov).

The American Geriatrics Society strongly endorses efforts to enhance the quality of care for this vulnerable population. Fundamental to the definition of a quality indicator is that the standard of care being measured is applicable to the individual to whom it is applied. For example, a quality indicator examining usage of mammograms would not be applicable to women under the age of 40. Nursing homes provide services to a heterogeneous population with varied health care service needs and goals of care. Subsets of nursing home residents include persons admitted for subacute or rehabilitative care following an acute illness, persons requiring long-term residential care for psychiatric, neurologic, or physical illness, and persons who have progressive incurable illness in its terminal phase. The majority of quality indicators are appropriate for nursing home residents whose goals for health care include maintenance of function and restoration of health. Such indicators are inappropriate when applied to dying persons, who usually prefer health care designed to enhance physical comfort and quality of life in the context of an inevitable decline in function and health.

Increasingly, nursing homes are the place of terminal care and site of death for Americans dying of chronic illnesses.³ Between 1989 and 1997, nursing homes became the final place of care for increasing numbers of older Americans. In 1989, 18.6% of persons who died of non-traumatic causes died in a nursing home. This increased to 24.1% in 1997. States such as Minnesota, Nebraska, Wisconsin, South Dakota, Washington, and Rhode Island in 1997 had more than one in three persons dying in a nursing home. By 2020, Brock and Foley estimate that 40% of deaths will occur in a nursing home.⁴ This change in demographics emphasizes the need to apply appropriate quality indicators for nursing home care for this population.

POSITION 1
Quality indicators must be appropriate to the usual health care needs and goals of a specified population.

Rationale: Nursing homes are increasingly the site of death for many older and frail adults. With the exception of pain, the proposed Centers for Medicare and Medicaid Services (CMS)⁵ quality indicators predominately focus on restoration of health and do not recognize that many nursing home residents have progressive incurable diseases. For example, indicators reported by CMS include worsening functioning and weight loss, which are expected aspects of terminal illnesses.

The numerator for the weight loss quality indicator is nursing home residents with weight loss greater than 10% in 180 days or 5% in 30 days as noted on the MDS and the denominator does not exclude all persons who are dying. The development of this quality indicator is based, in part, on important concerns as noted by ethnographic research by Kayser-Jones and colleagues that suggest that nursing homes were not providing adequate staff time for the feeding of nursing home residents.⁶ These are important concerns. Physical comfort during active dying may be promoted by natural decrease in intake of food and fluids, resulting in decreased shortness of breath, nausea, vomiting, and edema. Application of quality indicators such as weight loss to this population may result in unintended consequences such as hospitalizations, intravenous therapy or insertion of feeding tube to avoid sanctions. Furthermore, nursing facilities that agree to treat a disproportionate share of dying persons might be discouraged from admitting persons with known terminal illnesses.

Decisions to limit use of artificial hydration and nutrition ought to be based on patients’ and families’ informed preferences (as reflected in prior AGS position statements on decisions at the end of life).^7,8 Legal precedent and professional organization guidelines have endorsed the right of competent persons or their surrogate decision maker to withhold or withdraw artificial nutrition and hydration. Because information is missing on patient preferences and prognosis in the MDS, the current formulation of quality indicators cannot exclude dying persons from the denominator. The validity and clinical usefulness of quality indicators will improve when a more appropriate denominator population is used.

POSITION 2
Quality indicators need to be developed and validated that apply to care of those persons where the nursing home is the final place of care.

Rationale: Now that nursing homes are increasingly the place of terminal care and site of death for older Americans, indicators that examine the quality of end of life care are essential. These quality indicators, like other nursing home quality indicators, should be evaluated for their reliability, validity, and responsiveness.

POSITION 3
Prior to the nationwide implementation of a quality indicator, there must be evidence that the indicator is reliable, valid, under control of that health care institution, and that consideration has to be given to response burden to the patient and institution.

Rationale: The American Geriatrics Society encourages a public dialogue on what thresholds should be set for the reliability, validity, and responsiveness of proposed quality indicators. If nursing homes are subject to sanctions, or if quality indicators are reported publicly, the psychometric properties of proposed indicators should be held to high standards. There should be empirical or strong normative evidence that the proposed quality indicators are under the control of providers within the nursing homes to which they are applied.

The burdens and costs associated with data collection should also be considered as quality indicators are developed. Imposing long data collection efforts on a frail and vulnerable population is an intrusion. In addition, consideration must be given to the indirect impact of imposing costly and burdensome data collection efforts on health care institutions, such as nursing homes. Increasingly, health care institutions are being asked to provide more services with limited reimbursement. Burdensome data collection efforts can have the unintended consequence of directing staff away from direct patient care.

This AGS position statement was developed and reviewed by the AGS Ethics Committee and approved November, 2002 by the AGS Board of Directors. The AGS thanks Joan M. Teno, MD, MS for her work on this position statement. Address correspondence to: American Geriatrics Society, The Empire State Building, 350 Fifth Avenue, Suite 801, New York, NY 10118, Tel: 212-308-1414; Fax: 212-832-8646; E-mail: info.amger@americangeriatrics.org.