. 2014 Dec 1;14(16):1–78.

Table A12:

Study Characteristics and Adjustment Factors—Included Observational Studies on the Determinants of Nursing Home Versus Hospital Death

Author, Year Country Sample Size	Study Design Data Source Time Frame	Population	Sociodemographic Factors	Illness-Related Factors	Health Services Availability	Patient or Family Preference for Place of Death
	Study Characteristics			Adjustment Factors
Ikegami et al, 2012 (33) Japan N = 1,160	Retrospective cohort Retrospective survey answered by nurses Not specified	• Nursing home residents	• Not included	• Cause of death	Restricted to nursing home residents Physicians based in home care EoL care in facility	Family preference Agreement among family members
Levy et al, 2012 (34) United States N = 7,408	Retrospective cohort Administrative data set Not specified	Veterans Affairs nursing home residents Death 2005–2007	• Not included	Type of disease End stage, not hospice Length of stay Cognitive function	Restricted to Veterans Affairs nursing home residents Hospice care	• Advance directives
Houttekier et al, 2011 (19) Belgium N = 79,846	Retrospective cohort Death certificates Not specified	Nursing home residents ≥ 65 years Non-sudden deaths 1998–2007 Eligible for palliative care	Age Urbanization Living alone	• Cause of death	• Bed availability in hospital and care homes	• Not included
Houttekier et al, 2010 (14) Results for Belgium included in separate publication (19) N = 56,341 (Netherlands) N = 181,238 (England)	Retrospective cohort Death certificates Not specified	Adults Deaths in 2003	Age Sex Living arrangements^a Urban residence Income^a	• Disease type	Hospital beds Nursing home beds	• Not included
Houttekier et al, 2010 (13) Belgium N = 1,690	Retrospective cohort Retrospective data based on previous study Last 3 months of life	Adults and pediatrics Deaths 2005–2006	Age^a Sex^a Income Informal care during last 3 months of life Place of residence^a	• Cause of death^a	Hospital beds Care home beds^a GP involvement^a Home care involvement^a Palliative home care team	• Patient preference
Bell et al, 2009 (25) United States N = 1,352	Retrospective cohort Prospective cohort study Not specified	Adults Death 1991–1999 Japanese-American males	Age^a Restricted to males Restricted to Japanese-American	Cause of death Cognitive impairment^a Functional measures^a Interval between last examination and death^a	• Not included	• Not included
Kwak et al, 2008 (36) United States N = 30,765	Retrospective cohort Administrative databases Last year of life	≥ 65 years Nursing home residents Eligible for both Medicare and Medicaid Death 2000–2002 Excludes traumatic and sudden deaths	Age Sex Marital status Ethnicity Education Urban/rural residence	Cause of death Nursing home stay	Restricted to Medicare/Medicaid nursing home residents Hospice use	• Not included
Takezako et al, 2007 (37) Japan N = 86	Retrospective cohort Chart review Not specified	Deaths: 1999–2004 Died at nursing home or after transfer to hospital Excludes sudden deaths	Age Sex Marital status Living at home before nursing home	• Functional status	Restricted to nursing home residents Full-time physician presence	• Family preference for nursing home care
Motiwala, et al 2006 (28) Canada N = 58,689	Retrospective cohort Health claims databases Last year of life	>65 years Death 2001–2002	Age^a Sex^a Social deprivation Immigration status	Disease type Comorbidities Acute care conditions	• Hospital bed availability	• Not included
Levy et al, 2004 (35) United States N = 152,494	Retrospective cohort Administrative databases Not specified	Adults Deaths at nursing home or hospital Medicare admissions to nursing homes in 2001	Age Ethnicity Rural/urban residence	Disease type Function Cognitive performance	Restricted to Medicare nursing home residents Type of hospital (profit/non-profit) Size of hospital	DNR order DNH order

Abbreviations: DNH, do-not-hospitalize; DNR, do-not-resuscitate; EoL, end-of-life; GP, general practitioner; N, number of patients included.

Variables included in the multivariate model but odds ratios not provided.