Use of Feeding Tubes in Nursing Home Residents With Severe Cognitive Impairment

To the Editor: Many patients in the terminal phase of dementia have difficulty swallowing, with resulting weight loss and recurrent aspiration. Feeding tubes are often used for patients with advanced dementia in the hopes of delaying death or enhancing the quality of life, but the evidence supporting this practice is questionable.^{1 - 2} Although none of the 50 states prohibit the forgoing or withdrawal of a feeding tube, 15 states require a written advance directive to forgo insertion of a feeding tube.³ Aronheim et al,⁴ however, noted that variability in state laws did not fully explain the differences between states in use of feeding tubes. We sought other explanations for interstate variation in use of feeding tubes.

Using the 1999 National Repository of the Minimum Data Set (MDS), we determined the rate of feeding tube use among 385 741 US nursing home residents with severe cognitive impairment, as evidenced by a cognitive performance scale score of 5 or greater (the equivalent of a score of <5 on the Mini-Mental State Examination). All subjects had resided in a nursing home within 60 days of April 1, 1999. Using states as the unit of analysis, we examined the potential explanations for the state differences in the use of feeding tubes among severely cognitively impaired persons. Factors examined included (1) whether state law placed additional restrictions on forgoing feeding tube use based on a review conducted by the American Bar Association,³ (2) state daily Medicaid payment,⁵ and (3) rate of do-not-resuscitate (DNR) orders and orders to forgo artificial hydration and nutrition in severely demented nursing home residents as reported in the Resident Assessment Instrument that comprises the MDS. All these potential explanations of state variation in use of feeding tubes were entered into a multivariate linear regression model that also included a resource utilization groups (RUGS) case mix index to adjust for state differences in disease severity.

Nationwide, 18.1% of severely cognitively impaired persons had a feeding tube in 1999 (Figure 1). The rate varied from 3.8% (Nebraska) to 44.8% (District of Columbia). After adjustment for the RUGS case mix, state differences in the use of DNR orders among persons with severe cognitive impairment were associated with feeding tube use. (–0.45 parameter estimate, 95% confidence interval, –0.33 to –0.56). For each 10 percentage point increase in use of DNR orders, there was a 4.5 percentage point decrease in the use of feeding tubes. None of the other factors examined were associated with the states' rates of feeding tube use. Individual state reports of feeding tube use, use of DNR orders, and orders to forgo artificial hydration and nutrition are available online at http://www.chcr.brown.edu/dying/factsondying.htm.

Dementia often is a contributing cause of death for nursing home residents. We found nearly 10-fold differences in the states' rate of feeding tube use. Between-state differences in physician practices of writing DNR orders were related to feeding tube use, but other state policies were not. Given that the use of feeding tubes in patients with severe dementia may have questionable benefits, the wide variation that we found suggests the need for more consistent medical policies about how they are used in this situation.