Advance Directives among Nursing Home Residents with Mild,Moderate,and Advanced Dementia

Abstract

Objective:

To describe prevalence and content of AD documentation among NH residents by dementia stage.

Background:

The prevalence of advance directives (ADs) among nursing home (NH) residents with mild, moderate, and advanced dementia remains unclear.

Methods:

Population-based, cross-sectional study of all licensed NHs in five U.S. states. Subjects included all long-stay (>90 day) NH residents with dementia, aged ≥65 years, and a Cognitive Performance Scale (CPS) score ≥1 from the 2007 to 2008 Minimum Data Set 2.0 (n = 180,621). Dementia severity was classified as follows: mild (CPS 1–2), moderate (CPS 3–4), and advanced (CPS 5–6).

Measurements:

ADs were defined as the presence of a living will, do-not-resuscitate order, do-not-hospitalize order, medication restriction, or feeding restriction).

Results:

Overall, 59% of residents had any AD and 17% had a living will. Prevalence of any AD increased by dementia severity: mild (51.2%), moderate (58.2%), and advanced (61.5%) (p < 0.001). In adjusted analysis, resident characteristics associated with any AD documentation included older age, female gender, being white, and having more severe dementia. Having a living will was associated with higher education (≥high school graduate vs. some high school or less) and being married.

Discussion:

While dementia severity was associated with greater likelihood of having documented any AD, almost 4 in 10 residents with dementia lacked any AD. Effective outreach may focus efforts on subgroups with lower odds of any AD or living wills, including non-white, less educated, and unmarried NH residents. A greater understanding of how such factors impact care planning will help to address barriers to patient-centered care for this population.

Background

Advance care planning allows individuals to communicate their wishes regarding future medical care in the event that they are not able to speak for themselves. This is particularly important for adults with complex care needs, such as the nursing home (NH) population. In this setting, advance directives (ADs) provide essential guidance for both healthcare providers and family members who may be called to act as substitute decision makers.

Currently, there are over 1.3 million residents of long-term care NH facilities in the United States, almost half of whom have a diagnosis of Alzheimer's disease or other dementia.¹ As this population continues to grow,^2–4 it is increasingly important to understand and routinely document patient preferences regarding medical decision making and end-of-life care planning. Among residents with dementia, the subgroup of those with advanced dementia is at the highest risk for episodes of acute illness leading to hospitalization,^5,6 and is also the least able to articulate medical wishes. It is therefore important to capture goals of care information for patients with dementia while their ability to communicate is intact, which is typically preserved in early-stage disease.

Since the passage of the Patient Self-Determination Act of 1990, the prevalence and use of ADs in the general NH population have increased.^7,8 However, there is significant variation in estimates of the number of NH residents with ADs,^7,9 and the prevalence and variation in ADs by stage of dementia remain unclear. Documenting ADs among long-stay nursing facility residents with dementia carries particular importance as nursing facilities are the site of death for the majority of patients with dementia.¹⁰

Available studies about AD documentation among NH residents with dementia are small,¹¹ do not examine AD by dementia stage,^3,12,13 and do not focus on long-stay residents.^3,12–15 Furthermore, many prior studies focused solely on the presence of do-not-resuscitate (DNR) and do-not-hospitalize (DNH) orders, but did not describe other important ADs such as restrictions on feeding and medication use.

Given the growing reliance on NH care for dementia patients, and increasing support for initiating advance care planning discussions early in the dementia illness trajectory, it is timely to document what is known about ADs for NH residents with dementia by level of severity. This report is also timely, given the recent changes to the physician fee schedule in the United States under Medicare Part B allowing for Medicare reimbursement of 30-minute advance care planning discussions as of January 1, 2016.¹⁶ Therefore, we sought to describe ADs by stage of dementia in the nursing facility population from a large population-based sample of all licensed NH facilities in five U.S. states. We describe resident factors associated with having ADs in this report and hypothesize that later stages of dementia will have greater levels of ADs.

Methods

Data source and population

We used Minimum Data Set (MDS) 2.0 data from January 1, 2007, to December 31, 2008, to assemble a cohort of NH residents with dementia from all licensed facilities (N = 3371) in five states (Minnesota, Massachusetts, Pennsylvania, California, and Florida). The MDS is a standardized, clinically based instrument used in the U.S. NHs to assess resident condition for clinical and reimbursement purposes. While Centers for Medicare and Medicaid Services (CMS) introduced a newer version of the MDS on October 1, 2010, our data have relevance for large-scale descriptive epidemiology because the newer version of the MDS (v 3.0) does not capture any data on ADs. The MDS collects information on the demographic, functional, medical, psychological, and cognitive status of residents.¹⁷ A nurse interviews the resident, consults the medical record, and talks with other caregivers to collect information on the resident's care, cognitive and physical functioning, and behavior to complete the MDS assessment. The CMS require that each certified facility conduct an MDS assessment of all residents on admission, quarterly thereafter, and with significant changes in clinical status.

We identified all residents with dementia using Section I of the MDS 2.0, with associated Part D file and a length of stay >90 days (n = 226,936). We used the Cognitive Performance Scale (CPS) score to classify dementia severity. CPS scores correspond to levels of impairment: 1 (borderline intact); 2 (mild impairment); 3 (moderate impairment); 4 (moderately severe impairment); 5 (severe impairment); and 6 (very severe impairment with eating problems).¹⁸ For this study, we limited the sample to those with a CPS ≥1 for a final sample size of 180,621, and classified residents with CPS 1–2 as having mild dementia, those with CPS 3–4 as having moderate dementia, and those with CPS 5–6 as having advanced dementia.

Study variables

Advance directives

ADs were defined as the presence of a living will, DNR order, DNH order, medication restriction (i.e., resident or responsible party does not wish the resident to receive life-sustaining medications such as antibiotics or chemotherapy), or feeding restrictions (i.e., resident or responsible party does not wish the resident to be fed by artificial means if unable to be nourished orally). For analysis, we created three outcome variables: living will only; any other AD (not including living will); and any AD.

Other variables

Resident characteristics

Resident characteristics were obtained from the baseline MDS assessment and included the following: sociodemographics (age; sex; race/ethnicity [Hispanic; white, not Hispanic; black, not Hispanic; Asian/Pacific Islander; American Indian/Alaskan Native]; marital status [married; single (widowed, separated, divorced, never married)]; education [some high school or less; high school graduate or greater]); whether Medicaid was the primary payer; enrollment in hospice; residence in a special care unit for dementia; and geographic location of the NH (California, Florida, Massachusetts, Minnesota, and Pennsylvania).

Data analyses

We used descriptive statistics, including χ² tests for categorical variables, to describe resident characteristics and the proportion of residents having any AD, a living will only, and any other AD, and to characterize the proportion of residents with mild, moderate, and advanced dementia who have specific types of ADs (i.e., DNR; living will; do not hospitalize order; feeding restrictions; and medication restrictions).

We used logistic regression to compare residents with ADs to residents without any ADs. We stratified analyses by whether residents had a living will only, other ADs only, and any AD, because living wills require the resident to participate while other ADs are typically decided by a responsible party (i.e., not the resident). Unadjusted analysis was followed by adjusted logistic regression that included all factors associated with outcomes at the p < 0.20 level in unadjusted analysis. We used complete case analysis when values were missing for key predictor variables. All analyses were performed using STATA 10.1 (StataCorp, College Station, TX). The Institutional Review Board of the University of Massachusetts Medical School approved this study.

Results

The study population included 180,621 residents from five states; 72.7% were female and 19.6% were nonwhite. Overall, 22.1% had mild dementia, 53.3% had moderate dementia, and 24.6% had advanced dementia (Table 1).

Table 1.

Characteristics of Residents with Dementia (N = 180,621)

		Any advance directive^a (n = 103,772)			Living will (n = 31,786)			Other advance directives (n = 94,174)
Characteristic	N	n	%	P	n	%	p	n	%	p
Sex
Male	49,355	24,450	49.5	<0.001	7270	14.7	<0.001	21,912	44.4	<0.001
Female	131,209	79,295	60.4	<0.001	24,509	18.7	<0.001	72,238	55.1	<0.001
Age category
<80 years old	45,558	19,229	42.2	<0.001	5336	11.7	<0.001	17,151	37.7	<0.001
80–89 years old	88,566	50,904	57.5		15,878	17.9		45,827	51.7
≥90 years old	46,497	33,639	72.5		10,572	22.7		31,196	67.1
Race/ethnicity
Hispanic	12,519	4435	35.4	<0.001	517	4.1	<0.001	4212	33.6	<0.001
White, not of Hispanic origin	144,010	91,202	63.3		29,938	20.8		82,439	57.3
Black, not of Hispanic origin	15,812	4407	27.9		1050	6.6		3905	24.7
Asian/Pacific Islander	6215	2732	44.0		103	1.7		2694	43.4
American Indian/Alaskan Native	683	326	47.7		48	7.0		304	44.5
Education
<High school	33,909	16,320	48.1	<0.001	4731	14.0	<0.001	14,726	43.4	<0.001
High school or greater	77,619	40,445	52.1	<0.001	13,517	17.4	<0.001	35,224	45.4	<0.001
Marital status
Married	40, 699	21,615	53.1	<0.001	7020	17.3	0.04	19,121	47.0	<0.001
Not married (widowed, separated, divorced, never married)	139,922	82,157	58.7	<0.001	24,766	17.7	0.04	75,053	53.6	<0.001
Medicaid
Yes	99,564	60,684	60.9	<0.001	17,425	17.5	0.231	56,188	56.4	<0.001
No	81,057	43,088	53.2	<0.001	14,361	17.7	0.231	37,986	46.9	<0.001
Dementia severity
Mild	39,917	20,442	51.2	<0.001	6939	17.4	0.273	17,913	44.9	<0.001
Moderate	96,271	55,981	58.2		17,066	17.7		50,908	52.9
Severe	44,433	27,349	61.5		7781	17.5		25,353	57.0
Hospice enrollment
Yes	5462	4880	89.3	<0.001	1611	29.5	<0.001	4733	86.7	<0.001
No	175,159	98,892	56.5	<0.001	30,175	17.2	<0.001	89,441	51.1	<0.001
Special care unit
Yes	16,029	10,556	65.9	<0.001	3155	16.7	<0.001	9714	60.6	<0.001
No	164,592	93,316	56.6	<0.001	28,631	17.4	<0.001	84,460	51.3	<0.001
Geographic location
California	54,136	28,462	52.6	<0.001	630	1.2	<0.001	28,240	52.2	<.001
Pennsylvania	42,983	26,277	61.1		14,272	33.2		22,468	52.3
Minnesota	16,120	12,287	76.2		3200	19.9		11,745	72.9
Florida	42,698	20,915	49.0		12,749	29.9		16,060	37.6
Massachusetts	24,684	15,831	64.1		935	3.8		15,661	63.5

Any advance directive defined as living will or other advance directives (DNR, DNH, feeding restrictions, medication restrictions)

DNR, do-not-resuscitate; DNH, do-not-hospitalize.

For the entire population of residents with dementia, 57.5% (n = 103,772) had documentation of any AD, 17.6% (n = 31,786) had a living will, and 52.1% (n = 94,174) had other ADs. Dementia severity was associated with the documentation of any AD and other ADs, but not living wills (Table 1). Specifically, residents with mild dementia had the lowest prevalence of any AD at 51.2%, moderate dementia 58.2%, and advanced dementia 61.5% (p < 0.001); the prevalence of living will ranged from 17.4% to 17.7%, with no difference by dementia severity (p = 0.273). The prevalence of each specific type of AD order, not including living wills, increased with greater severity of dementia (Table 2).

Table 2.

Advance Directive Type by Severity of Dementia

	Overall N = 180,621		Mild dementia residents (CPS 1–2) n = 39,917		Moderate dementia residents (CPS 3–4) n = 96,271		Advanced dementia residents (CPS 5–6) n = 44,433
Type of advance directive	n	%	n	%	N	%	n	%	p
Do-not-resuscitate	89,994	49.8	16,822	42.2	48,749	50.6	24,423	54.9	<0.001
Living will	31,786	17.6	6939	17.4	17,066	17.7	7781	17.5	0.27
Do-not-hospitalize	8051	4.5	966	2.4	4171	4.3	2914	6.6	<0.001
Feeding restrictions	27,191	15.1	5348	13.4	14,533	15.1	7310	16.5	<0.001
Medication restrictions	7976	4.4	1537	3.9	4164	4.3	2275	5.1	<0.001

CPS, Cognitive Performance Scale.

There was significant geographic variation with respect to AD documentation, with the prevalence of any AD ranging from 49% for residents in Florida nursing facilities to 76% in Minnesota. Living wills also varied by state, from a low of 1.2% in California and 3.8% in Massachusetts to a high of 33.2% in Pennsylvania.

In adjusted analysis, resident characteristics, including older age, being female, and white, were associated with having each AD outcome (living will only; any other AD (not including living will); and any AD) (Table 3). Having moderate and advanced dementia was uniquely and strongly associated with having other ADs (i.e., ADs that did not include living wills) (adjusted odds ratio [OR] 1.34, 95% confidence interval [CI] 1.31, 1.38; adjusted OR 1.59, 95% CI 1.54, 1.65, respectively). Having a living will was uniquely associated with having higher education (adjusted OR 1.30; 95% CI 1.25, 1.35) and being married (adjusted OR 1.09; 95% CI 1.05, 1.14). Geographic differences persisted in adjusted analysis, with all states having greater odds of any AD documentation compared to Florida, and residents of California, Massachusetts, and Minnesota having lower odds for having a living will than Florida (Table 3).

Table 3.

Association between Resident Characteristics and Having Any Advance Directive, by Type

	Any advance directive ^a		Living will		Other advance directives
Patient characteristics	Adjusted odds ratio	95% Confidence interval	Adjusted odds ratio	95% Confidence interval	Adjusted odds ratio	95% Confidence interval
Age
<80 years (reference)	1.00		1.00		1.00
80–89 years	1.65	1.61, 1.70	1.54	1.48, 1.60	1.61	1.57, 1.66
≥90 years	2.85	2.75, 2.94	2.08	1.98, 2.17	2.76	2.67, 2.86
Sex
Male (reference)	1.00		1.00		1.00
Female	1.17	1.14, 1.21	1.11	1.07, 1.15	1.17	1.14, 1.21
Race/ethnicity
White (reference)	1.00		1.00		1.00
American Indian/Alaskan native	0.57	0.47, 0.70	0.65	0.43, 0.97	0.57	0.47, 0.69
Asian/Pacific islander	0.46	0.43, 0.49	0.36	0.28, 0.48	0.47	0.44, 0.51
Black	0.26	0.25, 0.27	0.24	0.22, 0.26	0.29	0.27, 0.30
Hispanic	0.35	0.33, 0.36	0.17	0.15, 0.19	0.41	0.39, 0.43
Education
Less than high school (reference)	1.00		1.00		1.00
High school or greater	1.07	1.04, 1.09	1.28	1.24, 1.33	0.99	0.97, 1.02
Marital Status
Single (widowed, separated, divorced, never married) (reference)	1.00		1.00		1.00
Married	0.99	0.96, 1.02	1.08	1.04, 1.13	0.96	0.97, 1.02
Medicaid
No (reference)	1.00		—	—	1.00
Yes	1.44	1.41, 1.48	—	—	1.51	1.48, 1.55
Dementia severity
CPS 1–2 (reference)	1.00		—	—	1.00
CPS 3–4	1.31	1.27, 1.35	—	—	1.34	1.31, 1.38
CPS 5–6	1.48	1.43, 1.53	—	—	1.59	1.54, 1.65
Hospice
Not enrolled (reference)	1.00		1.00		1.00
Enrolled	7.65	6.88, 8.51	1.67	1.54, 1.82	8.46	7.67, 9.32
Special care unit
No (reference)	1.00		1.00		1.00
Yes	1.46	1.41, 1.53	1.25	1.18, 1.31	1.42	1.37, 1.49
State
FL (reference)	1.00		1.00		1.00
PA	1.67	1.62, 1.72	1.03	1.00, 1.07	2.00	1.94, 2.07
CA	1.38	1.34, 1.42	0.03	0.028, 0.033	2.45	2.37, 2.53
MA	1.52	1.46, 1.57	0.08	0.075, 0.088	2.69	2.59, 2.79
MN	2.29	2.18, 2.42	0.44	0.41, 0.47	3.47	3.30, 3.65

Any advance directive defined as living will or other advance directives (DNR, DNH, feeding restrictions, medication restrictions).

MN, Minnesota; MA, Massachusetts; PA, Pennsylvania; CA, California; FL, Florida.

Discussion

No prior study has described prevalence of ADs among NH residents by dementia severity. In this large, population-based, study sample in the NH setting, we found that living wills were less common than other ADs that may involve substituted judgment from a proxy decision maker (i.e., DNR and DNH orders, feeding restrictions, and medication restrictions). Dementia severity was associated with greater likelihood of such ADs, but not living wills.

Overall, we found that only 58% of residents with dementia had any AD. This is a cause for concern since the risk for acute conditions requiring more aggressive care, such as infections, increases as dementia progresses to more severe disease stages.⁶ For example, in the last 3 months of life, the likelihood of residents with advanced dementia undergoing at least one burdensome intervention such as a hospitalization, emergency room visit, parenteral therapy, or tube feeding is around 40%.⁶ Without an AD, there is increased burden on loved ones to make decisions on behalf of the resident, and increased risk that the decisions made may not be in keeping with the resident's wishes. Assuming that substitute decision makers and healthcare teams honor the wishes expressed by residents, ADs can reduce the risk of undue and burdensome care in this population.

Fewer than one in five residents in our population had a living will. Since living wills are initiated by individuals themselves, typically before developing serious illness, it is not surprising that the prevalence of living wills did not increase with increasing dementia severity (i.e., when they lacked the cognitive ability to initiate or participate in the writing of a living will). This suggests that <20% of NH residents have documented evidence of conversations related to ADs early in their course of dementia. Interestingly, living wills were more prevalent among residents who were married and had greater education. Our findings are consistent with other studies describing that ADs are more likely in advanced dementia residents who are better educated and married.¹¹ This suggests an opportunity for increasing the awareness of early advance care planning and ADs among less educated individuals.

We found evidence of significant difference in the prevalence of any AD by increasing age, female sex, race/ethnicity, and geography. These findings are consistent with other studies showing variation in AD among advance dementia residents by race,¹¹ in the general population by older age, sex, race/ethnicity, and education,¹⁹ and in geographic variation in the care of seriously ill adults.²⁰ Geographic variation in living wills is not surprising given that there are states that do not officially recognize living wills, including Massachusetts. Our findings suggest that facilities and healthcare providers may have the greatest impact by improving advance care planning outreach to subgroups of residents who have consistently shown lower likelihood of AD completion—NH residents who are less-educated, nonwhite, or unmarried.

This is a particularly important time to refocus attention on advance care planning, with the January 1, 2016, amendment of the Physician Fee Schedule and Medicare's introduction of Current Procedural Terminology (CPT) Code 99497 for advance care planning. This code allows for recognition and billing of the first 30 minutes of advance care planning discussion, face-to-face with the patient, family member(s), and/or surrogates with an additional 30 minutes available through CPT Code 99498.²¹ As stated in the CY 2016 Physician Fee Schedule (PFS) final rule (80 Fed. Reg. 70956), advance care planning (ACP) services may be appropriately furnished in a variety of settings depending on the needs and condition of the beneficiary. While there are no place of service limitations on the ACP codes, theoretically allowing the use of these codes for services rendered in the nursing facility, there is not yet empiric evidence that this is being used. Furthermore, there are nuances of NH practice that may actually make implementation of such a CPT code challenging and distinctly different from an outpatient/ambulatory setting. For example, factors that may constrain use in the NH include the limited time clinicians spend on site,^22,23 the increasing use of the “Physician-Order for Life Sustaining Treatment” paradigm,²⁴ and the fact that telephone conversations with surrogates do not meet the billing criteria (when much of the family communication from NH clinicians and family occurs by phone and not face-to-face). Going forward, NH facilities need to explore and develop models for increasing the frequency of advance care planning discussions, particularly among residents with dementia who still have cognitive ability to articulate their preference.

Our findings must be considered in the context of the study limitations. First, we used data from 2007 because the newer version of the MDS (v 3.0) does not capture any data on ADs. Given the recent surge in interest in advance care planning, it is possible that available data may be underestimating AD levels in the United States. However, we believe the age of the data does not substantially mitigate the importance of our findings, particularly given the recent continued evidence of concerning rates of hospitalizations and burdensome interventions in this population.^5,6

We also note that our outcome measure describes ADs with restrictions for care, including DNR, DNH, and limitations on feeding and medications. It is possible that some NH residents without an AD engaged in advance care planning conversations with their providers and elected to accept all resuscitative efforts; our current data cannot differentiate between residents who actively decided against an AD and those for whom the lack of AD represents an omission of advance care planning. Finally, our data do not capture information regarding the choice of substitute decision makers in this population. However, since designation of a surrogate decision maker is a customary part of the NH admission process, we do not believe that this detracts from our findings.

Conclusions

There is room for interventions to increase the documentation of ADs in NH residents with dementia. More importantly, there is a need to reach these residents at earlier stages of their cognitive impairment, while they are still able to state their preferences. Furthermore, when ADs are documented, it remains important for facilities and clinicians to honor ADs. The Medicare benefit may provide one possible strategy for increasing advance care planning discussion in the NH, but further work needs to be done to understand how to address the particular challenges of billing for such conversations in this setting. In the end, we need to better understand how to reach populations that have traditionally had lower levels of AD completion.

Footnotes

Acknowledgments

The authors acknowledge the contributions of Daniel Peterson, MA, who assisted with data analysis. Dr. J.T. was supported by a Sojourns Scholar Leadership Award from the Cambia Health Foundation.

Author Disclosure Statement

No competing financial interests exist.

References

Harris-Kojetin

, Sengupta

, Park-Lee

, et al.: Long-term care providers and services users in the United States: Data from the National Study of Long-Term Care Providers, 2013–2014. National Center for Health Statistics. Vital Health Stat, 2016:x–xii, 1–105.

Vincent

, Velko

: The Next Four Decades: The older population in the United States: 2010 to 2050. Current Population Reports 2010;P-25-1138. Washington, DC: U.S. Census Bureau.

Hirschman

, Abbott

, Hanlon

, et al.: What factors are associated with having an advance directive among older adults who are new to long term care services?. J Am Med Dir Assoc, 2012; 13:82e7–82e11.

Plassman

, Langa

, Fisher

, et al.: Prevalence of dementia in the United States: The aging, demographics, and memory study. Neuroepidemiology, 2007; 29:125–132.

Gozalo

, Teno

, Mitchell

, et al.: End of life transitions among nursing home residents with cognitive issues. N Engl J Med, 2011; 365:1212–1221.

Mitchell

, Teno

, Kiely

, et al.: The clinical course of advanced dementia. N Engl J Med, 2009; 361:1529–1538.

Resnick

, Schuur

, Heineman

, et al.: Advance directives in nursing home residents aged ≥65 years: United States 2004. Am J Hosp Palliat Care, 2008; 25:476–482.

Allen

, DeLaine

, Chaplin

, et al.: Advance care planning in nursing homes: Correlates of capacity and possession of advance directives. Gerontologist, 2003; 43:309–317.

McAuley

, Buchanan

, Travis

, et al.: Recent trends in advance directives at nursing home admission and one year after admission. Gerontologist, 2006; 46:377–381.

10.

Mitchell

, Teno

, Miller

, et al.: A national study of the location of death for older persons with dementia. J Am Geriatr Soc, 2005; 53:299–305.

11.

Triplett

, Black

, Phillips

, et al.: Content of advance directives for individuals with advanced dementia. J Aging Health, 2008; 20:583–596.

12.

Levy

, Fish

, Kramer

: Do-not-resuscitate and do-not-hospitalize directives of persons admitted to skilled nursing facilities under the Medicare benefit. J Am Geriatr Soc, 2005; 53:2060–2068.

13.

Brink

: Examining do-not-resuscitate orders among newly admitted residents of long-term care facilities. Palliat Care Res Treat, 2014; 8:1–6.

14.

Dobalien

: Advance care planning documents in nursing facilities: Results from a nationally representative survey. Arch Gerontol Geriatr, 2006; 43:193–212.

15.

Buchanan

, Bolin

, Wang

, et al.: Urban/rural differences in decision making and the use of advance directives among nursing home residents at admission. J Rural Health, 2004; 20:131–135.

16.

Centers for Medicare and Medicaid Services: Proposed policy, payment, and quality provisions changes to the Medicare Physician Fee Schedule for Calendar Year 2016. www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2015-Fact-sheets-items/2015-10-30-2.html (last accessed July 28, 2017).

17.

Hawes

, Phillips

, Mor

, et al.: MDS data should be used for research. Gerontologist, 1992; 32:563–564.

18.

Hartmaier

, Sloane

, Guess

, et al.: The MDS cognition scale: A valid instrument for identifying and staging nursing home residents with dementia using the minimum data set. J Am Geriatr Soc, 1994; 42:1173–1179.

19.

Rao

, Anderson

, Lin

F-C

, Laux

: Completion of advance directives among U.S. consumers. Am J Prev Med, 2014; 46:65–70.

20.

Mitchell

, Teno

, Roy

, et al.: Clinical and organizational factors associated with feeding tube use among nursing home residents with advanced cognitive impairment. JAMA, 2003; 290:73–80.

21.

Centers for Medicare and Medicaid Services: Frequently asked questions about billing the physician fee schedule for advance care planning services. www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/PhysicianFeeSched/Downloads/FAQ-Advance-Care-Planning.pdf (last accessed July 28, 2017).

22.

Perkins

, Gagnon

, deGruy

: A comparison of after-hours telephone calls concerning ambulatory and nursing home patients. J Fam Pract, 1993; 37:247–250.

23.

Fowkes

, Christenson

, McKay

: An analysis of the use of the telephone in the management of patients in skilled nursing facilities. J Am Geriatr Soc, 1997; 45:67–70.

24.

Kapp

: The nursing home as part of the POLST paradigm. Hamline Law Rev, 2014; 36:151–175.