Tube-Feeding in US Nursing Home Residents with Advanced Dementia, 2000–2014

To the Editor

Over the last two decades, research has failed to demonstrate benefits of tube-feeding in advanced dementia.^1,2 Expert opinion and position statements by national organizations increasingly advocate against this practice.³ We describe feeding tube insertion rates from 2000–2014 among US nursing home residents with advanced dementia. Racial disparities were examined because Black race has been strongly associated with greater feeding tube use.⁴

Methods

Brown University Institutional Review Board approved this study with a waiver of informed consent. Data were derived from federally-mandated Minimum DataSet (MDS) assessments completed quarterly, as required, on all residents in US nursing homes between January 1, 2000 and October 31, 2015. First, residents with the following criteria on an initial MDS assessment completed nearest to April 1 (+/− 60 days) in each year from 2000–2014 were selected: i. Alzheimer’s disease or other dementia, ii. severe cognitive impairment,⁵ iii. not dependent for eating, and iv. no feeding tube. Next, residents meeting these criteria who became totally dependent for eating on any MDS assessment completed within 120 days after their initial assessment were identified. Finally, the proportion of advanced dementia residents with recent eating dependency who had a new feeding tube indicated on any MDS assessment during the next 12 months was determined. Poisson regression models were used to estimate the linear time trend of annual changes in insertion rates using all years of data, controlling for age, gender, race (white, black, and other), and stroke. The cumulative change over 14 years was calculated from annual changes, and presented as a comparison of insertion rates between 2000 and 2014 with adjusted risk ratios (ARRs) and 95% confidence intervals (CIs). Separate models were generated for White and Black residents controlling for age, gender, and stroke. Data were analyzed using SAS v9.4 and Stata v14.

Results

Between 2000 and 2014, 71,251 advanced dementia residents with recent dependence for eating were identified with the following characteristics: mean age, 84.04 years ± 8.79 (standard deviation); 76.4% female; White, 85.6%; Black, 9.5%; and prior stroke, 13.6%. These characteristics were similar across years. The proportion of residents receiving feeding tubes over the next 12 months declined from 11.7% in 2000 to 5.7% in 2014 (ARR, 0.41, 95% CI, 0.38–0.45) (Tables 1 and 2). Insertion rates declined between 2000 and 2014 among White (8.6 to 3.1%; ARR 0.37, 95% CI; 0.33–0.41), and Black residents (37.6% to 17.5%, ARR 0.47 (95% CI; 0.41–0.55). However, Black (versus White) residents were more likely to get tube-fed in 2000 (ARR 4.4, 95%CI; 4.0–4.7) and 2014 (ARR 5.6, 95% CI; 5.0–6.2).

Table 1.

Feeding-tube insertion rates among US nursing home residents with advanced dementia in 2000 to 2014

Year	No. advanced dementia residents with recent onset of total dependence for eating	No. of residents with feeding tubes over subsequent 12 months	12-month feeding tube insertion rates (%)
2000	7,029	820	11.7
2001	6,738	774	11.5
2002	6,239	701	11.4
2003	5,518	577	10.5
2004	5,194	462	8.9
2005	4,628	398	8.6
2006	4,389	393	9.0
2007	4,110	357	8.7
2008	3,890	331	8.5
2009	3,842	297	7.7
2010	3,794	283	7.5
2011	4,538	264	5.8
2012	4,246	235	5.5
2013	3,685	207	5.6
2014	3,411	193	5.7

Table 2.

Comparison of feeding-tube insertion rates among US nursing home residents with advanced dementia in 2014 and 2000

	% advanced dementia residents receiving a feeding tube within 12 months becoming dependent for eating		Difference between 2000 and 2014 (95% confidence interval)	Unadjusted risk ratio (95% confidence interval)	Adjusted risk ratio (95% confidence interval)
	2000	2014
All	11.7	5.7	−6.0 (−7.6 – −4.5)	0.45 (0.41–0.50)	0.41 (0.38–0.45)2
White1	8.6	3.1	−5.5 (−7.0 – −4.3)	0.37 (0.31–0.43)	0.37 (0.33–0.41)3
Black1	37.5	17.5	−20.1 (−25.5 – −10.2)	0.47 (0.40–0.55)	0.47 (0.41–0.55)3

¹Race data was obtained from the Minimum DataSet where it is collected as follows: White, Black or African-American, Asian, American Indian or Alaska Native, Native Hawaiian or other Pacific Islander, Other, and Unable to Determine. For these analyses race was categorized as White, Black, or other.

²Adjusted for age, gender, race and prior stroke

³Adjusted for age, gender, and prior stroke

Discussion

The proportion of US nursing home residents with advanced dementia and eating dependency receiving feeding tubes decreased by approximately 50% between 2000 and 2014. This decline parallels the emergence of research,^1,2 expert opinion, and recommendations by national organizations,³ discouraging this practice. Feeding-tube use decreased across racial groups, but remained relatively higher among Black residents, consistent with prior research.^2,4

This study has limitations. The number of reported feeding tube insertions are specific to the cohort definitions. The actual number of tubes inserted in all U.S. advanced dementia residents is likely much higher. Moreover, the reported number of residents with advanced dementia and eating problems declined from 2000 to 2014; reflecting the shifting composition of US nursing homes, such that patients with chronic illnesses, including dementia, are being increasingly maintained in the community with greater access to services.⁶ However, given the analyses applied the same definitions to the numerator and denominator in each year, the comparison of annual insertion rates is reasonable. Power was inadequate to examine factors associated with tube-feeding use.

There are many drivers of practice change in medicine. The next step to ensure the message from existing evidence and expert recommendations are disseminated and disparities are reduced, fiscal and regulatory policies are needed that discourage tube-feeding and promote a palliative approach to feeding problems in advanced dementia patients.