Abstract
Background:
Scant research has examined the relationship between family characteristics and end-of-life (EOL) outcomes despite the importance of family at the EOL.
Objectives:
This study examined factors associated with the size and composition of family relationships on multiple EOL hospitalizations.
Design:
Retrospective analysis of the Utah Population Database, a statewide population database using linked administrative records.
Setting/subjects:
We identified adults who died of natural causes in Utah, United States (n = 216,913) between 1998 and 2016 and identified adult first-degree family members (n = 743,874; spouses = 13.2%; parents = 3.6%; children = 51.7%; siblings = 31.5%).
Measurements:
We compared demographic, socioeconomic, and death characteristics of decedents with and without first-degree family. Using logistic regression models adjusting for sex, age, race/ethnicity, marital status, comorbidity, and causes of death, we examined the association of first-degree family size and composition, on multiple hospitalizations in the last six months of life.
Results:
Among decedents without documented first-degree family members in Utah (16.0%), 57.7% were female and 7 in 10 were older than 70 years. Nonmarried (aOR = 0.90, 95% CI = 0.88–0.92) decedents and decedents with children (aOR = 0.97, 95% CI = 0.94–0.99) were less likely to have multiple EOL hospitalizations. Family size was not associated with multiple EOL hospitalizations.
Conclusions:
First-degree family characteristics vary at the EOL. EOL care utilization may be influenced by family characteristics—in particular, presence of a spouse. Future studies should explore how the quality of family networks, as well as extended family, impacts other EOL characteristics such as hospice and palliative care use to better understand the EOL care experience.
Keywords: caregivers, end-of-life, population health, socioeconomic factors
Introduction
An estimated 53 million informal caregivers care for adults annually in the United States, providing support to patients with chronic illness, disability, terminal diseases, mental illness, and cognitive impairment.1 Caregiving at the end of life (EOL) is burdensome and involves more complex symptom management,2 assistance with daily care, and decision making.3–5 EOL care is often borne by family caregivers—in the United States, ∼9 in 10 EOL caregivers are family members.2 Aside from being intensively involved in daily caregiving and care coordination, family caregivers are often also pillars of emotional, physical, and financial support at EOL.6–8
Hospitalizations are burdensome and costly during the last months of life9,10—the largest proportion of EOL hospitalizations are at EOL.11 The use of emergency department services, inpatient hospitalizations, and intensive care services increases toward the final days of life, particularly among decedents who are younger, from minority backgrounds, with lower income and education, or have more poorly controlled or chronic illness.12–18 Nearly 20% of Medicare spending among the older adult population is in the last year of life, with costs increasing closer to death.11
Larger families may offer greater caregiving resources, skills, and labor to care for dying patients at home and potentially provide better quality caregiving that reduce the need for individuals to be hospitalized for symptom changes or EOL crises. Families may have a beneficial impact on EOL decision making; Veterans with family members who were actively engaged in caregiving were more likely to have palliative care and documented preferences for less aggressive care at EOL.19 Despite the importance of families in EOL caregiving, most caregiving research focuses on a single or “primary caregiver” and neglects the contribution of family coordination and task sharing.20
In contrast, a lack of close family networks, such as first-degree relatives (parents, siblings, spouses, children), may contribute to social isolation, which is associated with poorer health outcomes.21,22 Social isolation in adults is associated with greater risk of mortality, hospitalization, and emergency department visits.23,24 Changes in demography such as higher rates of childlessness,25 lower rates of marriages, an aging society, and migration of family members across states and national borders26 may also affect the pool of available family caregivers for those with serious illnesses.
Social relationships are associated with commonly known risk factors (e.g., physical activity, smoking)21 and can directly and indirectly affect an individuals' uptake of care behaviors (e.g., medication adherence, hydration).27 Antonucci et al.'s28 convoy model of social relations purports that individuals create a network of relationships across the life span. During EOL when the dying person becomes more dependent, convoy members may become more activated,29 suggesting that the number and diversity of relationships in one's convoy, particularly among one's family, may allow for sharing of the social support, advocacy, coordination, and caregiving tasks.30 While population-based data cannot indicate quality, strength, or specific actions of family members, it can provide general information on the impact of families on costly EOL health care utilization.
Understanding the impact of families on patients' EOL hospitalization use in a U.S. population can inform policies and programs targeted at supporting EOL caregiving families and help identify subpopulations that may benefit the most from these programs. Analyses of the Health and Retirement Study in the United States show that decedents without close family members at the time of death are at risk for poorer EOL outcomes like nursing home death and fewer hours of caregiving.31 Unfortunately, much of the research on family impacts on EOL outcomes is limited by smaller studies and convenience samples with few population-based studies.5 With the exception of palliative care research conducted in Nordic countries,32–34 no such work exists in the United States that facilitates linkages of patient EOL outcomes with family data.31
The Utah Population Database (UPDB) is the only population database of its kind in the United States. The UPDB links administrative and health records at the individual and family level to create a rich set of demographic, genealogical, genetic, health, and administrative data for all Utah residents. While researchers have examined familial relationships in the context of longevity, genetics, cancer risk, and early life risk factors for later life comorbidity and dementia onset,32,34–37 the UPDB is yet to be tapped for EOL caregiving research. We used data from UPDB covering the period 1996 to 2016 to address two aims: (1) identify characteristics related to having no first-degree family and (2) examine the relationship between first-degree family characteristics and multiple hospitalizations during the last six months of life. We hypothesized that the number of first-degree family and type of first-degree family would be associated with multiple EOL hospitalizations.
Methods
This study was approved by the University of Utah's Resource for Genetic and Epidemiologic Research. As the registry data for this study were deidentified, this study was considered nonhuman subjects research by the University Institutional Review Board (IRB #00117703).
Data
The UPDB was established in the mid-1970s and links data sources such as vital records, voter registrations, family history records, statewide health care facility data, and electronic data warehouses for Utah residents. Demographic, genealogical, economic, geographic, and health care data from this study were collected from UPDB demographic records, vital record data (birth, marriage, and death certificates), American Community Survey, and statewide inpatient, ambulatory, and emergency department medical databases (Table 1). These linked data sources provided information on EOL inpatient hospitalization for each decedent, as well as information on the characteristics of each decedent's first-degree family (Fig. 1).
Table 1.
Characteristics of Decedents: Description of Key Variables and Data Sources
| Variable name | Variable description | Data source |
|---|---|---|
| Sociodemographic characteristics | ||
| Age group | 20–49, 50–59, 60–69, 70–79, 80–89, 90–100 | Vital Statistics (death certificate) |
| Sex | Binary—male, female | Vital Statistics (birth certificate) |
| Marital status | Married, married but separated, divorced, widowed, never married/single, other | Vital Statistics (death certificate) |
| Race/ethnicity | American Indian or Alaska Native, Asian, Black, Native Hawaiian or Other Pacific Islander, and White. Latinx, not Latinx | Vital statistics records, based on NIH race/ethnicity classification |
| Rural/urban | Urban (population >100 people per sq. mile), rural (population <99 and >6 people per sq. mile), frontier (<6 people/sq. mile) | Residential address on administrative records; address recoded to zip code |
| Education | Less than high school, high school graduate, some college, college graduate, postcollege | Max level of education as reported on death certificate or as reported on birth certificate of latest-born child |
| Income | Median household income of census tract (in dollars) | Residential data coded to census tract, assigned five-year estimate of median household income from 2010 to 2014 American Community Survey |
| Insurance coverage | Medicaid, Medicare, private insurance, other, self-pay | Inpatient/ambulatory/emergency department records |
| Context of death and EOL health care utilization characteristics | ||
| Primary cause of death | Cancer (any site), heart disease, cerebrovascular disease, COPD, other | Vital statistics (death certificate) defined with ICD-9/ICD-10 codes |
| Dementiaa | Primary or secondary cause of death | Vital statistics (death certificate) |
| Comorbidity | CCI, measured as 0, 1–2, 3–4, and 5+ indicating no/mild/moderate/severe comorbidities, respectively | CCI using UPDB's algorithm based on individuals' ICD-9 and ICD-10 diagnosis and procedure codes obtained from inpatient/ambulatory/emergency department records |
| Multiple EOL hospitalizations | Two or more inpatient stays within six months of death | Utah Department of Health |
| Decedent family characteristics | ||
| Size of family | No family member, 1–3 family members, 4+ family members | Identified by pedigree linkages of UPDB, created through Vital Statistics (birth, death, marital records) and measured within the two years before death |
| Type of family member | Parent, child, sibling, spouse | Identified by pedigree linkages of UPDB, created through Vital Statistics (birth, death, marital records) and measured within the two years before death |
The UPDB is fee based for Utah and non-Utah researchers and contains data on over 11 million individuals, including individuals living in rural and frontier regions of the state. Access, use, and data privacy and protection are tightly regulated and overseen by the University of Utah Institutional Review Board and Genetic and Epidemiologic Research. The data are maintained in secure storage by the Huntsman Cancer Institute Computing and Technology Groups behind industry strength firewalls in line with industry best practices. Access to the raw data is limited to information technology staff and analysts at the UPDB, while analytic dataset access is through Virtual Private Network.
Major cause of death in Utah is often a primary diagnosis but seldom the precipitating cause of death.
CCI, Charlson Comorbidity Index; COPD, chronic obstructive pulmonary disease; EOL, end of life; ICD, International Classification of Disease; UPDB, Utah Population Database.
FIG. 1.
Decedent and family cohort (n = 960,787) and linkages of the UPDB data sources. UPDB, Utah Population Database.
Sample
Eligible decedents (n = 216,913) were residents of Utah and at least 20 years old at the time of death. Based on the classification of cause of death on the death certificate, only individuals who died of natural causes between 1998 and 2016 were included (other excluded categories were suicide, homicide, accident, undetermined, and pending investigation—sudden death due to natural causes is captured; Supplementary Table S1). The main dataset was composed of available patient and first-degree family data collected two years before and first-degree family data collected up to two years after patient death. Family members were alive, 18 years and older, and residing in Utah at the time of decedent death. Birth certificates were used to identify biological first-generation family members (defined as parents, children, and siblings) of each decedent included in the analytic cohort. Decedents who did not have any identified living spouse/partner or biological first-degree family member identified in the UPDB records (i.e., no spouse/partner and no child, sibling, and parent records) were classified as having no first-degree family.
Measures
Table 1 shows the key variables and data sources used for the decedent demographic, socioeconomic, context of death, and EOL health care utilization characteristics. Sociodemographic, family, and health characteristics collected on an individual level (e.g., death certificate) were selected a priori, guided by the EOL caregiving literature.
Aim 1
Sex, age group, race/ethnicity, education, region, primary cause of death, hospitalizations in the last six months, and Charlson Comorbidity Index (CCI)38 groups within the last two years were examined in association with having first-degree family at the time of death (no family vs. any family).
Aim 2
Type (spouse, children, parents, siblings) and size of family network (0, 1–3, and 4+) were examined in association with multiple EOL hospitalizations (2+ hospitalizations in last six months). We controlled for sex, age group, race/ethnicity, education, region, comorbidities, and cause of death (including dementia as an associated cause of death).
Analysis
Descriptive statistics, chi-square tests, and t tests were performed to describe the decedent cohort (n = 216,913) and their first-degree family members (n = 743,874), as well as to identify differences among deceased individuals with no family versus individuals with any family. Logistic regression models were performed for Aim 2.
Results
Characteristics of decedents
n = 216,913 (48.8% male, 51.2% female) eligible decedents were identified (median age 80 years, IQR = 69–87 years; Table 2). The majority of deceased individuals were non-Latinx White (79.2%), and 24.4% resided in rural or frontier regions of the state. Heart disease (24.2%) and cancer (21.3%) were the most common primary causes of death. Of the decedents, 13.4% had dementia as a primary or secondary cause of death. Over a quarter (26.0%) had multiple EOL hospitalizations.
Table 2.
Demographic, Socioeconomic, Context of Death, and End-of-Life Health Care Utilization Characteristics of Decedents
| Characteristic of decedents | Decedents |
Malesa |
Femalesa |
|---|---|---|---|
| (n = 216,913) n (%) | (n = 105,924) n (%) | (n = 110,989) n (%) | |
| Demographic characteristics | |||
| Age at decedents' deathb | |||
| Median (Q1–Q3) | 80 (69–87) | 78 (66–86) | 82 (72–89) |
| Decedent age groupsb | |||
| 20–49 | 12,498 (5.8) | 6,999 (6.6) | 5,499 (5.0) |
| 50–59 | 17,327 (8.0) | 10,113 (9.5) | 7,214 (6.5) |
| 60–69 | 28,875 (13.3) | 16,571 (15.6) | 12,304 (11.1) |
| 70–79 | 49,001 (22.6) | 26,140 (24.7) | 22,861 (20.6) |
| 80–89 | 73,599 (33.9) | 33,742 (31.9) | 39,857 (35.9) |
| 90+ | 35,613 (16.4) | 12,359 (11.7) | 23,254 (21.0) |
| Race/ethnicityc | |||
| White, Non-Latinx | 171,806 (79.2) | 84,282 (79.6) | 87,524 (78.9) |
| Non-White or Latinx | 45,107 (20.8) | 21,642 (20.4) | 23,465 (21.1) |
| Marital statusd | |||
| Married | 97,539 (45.0) | 64,044 (60.5) | 33,495 (30.2) |
| Nonmarried | 119,374 (55.0) | 41,880 (39.5) | 77,494 (69.8) |
| Widowed | 80,893 (37.3) | 21,553 (20.3) | 59,340 (53.5) |
| Divorced | 26,827 (12.4) | 13,668 (12.9) | 13,159 (11.9) |
| Separated, single/never married and others | 11,654 (5.3) | 6,659 (6.3) | 4,995 (4.5) |
| Socioeconomic characteristics | |||
| Maximum educatione | |||
| Less than high school | 62,239 (28.7) | 26,170 (24.7) | 36,069 (32.5) |
| High school graduate | 77,201 (35.6) | 34,695 (32.8) | 42,506 (38.3) |
| Some college | 49,250 (22.7) | 26,455 (25.0) | 22,795 (20.5) |
| College graduate | 15,832 (7.3) | 9,213 (8.7) | 6,619 (6.0) |
| Postcollege | 12,391 (5.7) | 9,391 (8.9) | 3,000 (2.7) |
| Median household incomef (in $) | |||
| Median (Q1–Q3) | 56,989 (45,208–70,441) | 56,989 (45,357–70,476) | 56,989 (45,139–70,441) |
| Missing | 11,200 (5.2) | 5,578 (5.3) | 5,622 (5.1) |
| Regiong | |||
| Urban | 163,984 (75.6) | 79,100 (74.7) | 84,884 (76.5) |
| Rural | 42,055 (19.4) | 21,2280 (20.1) | 20,775 (18.7) |
| Frontier | 10,874 (5.0) | 5,544 (5.2) | 5,330 (4.8) |
| Context of death and EOL health care utilization | |||
| Primary cause of deathh | |||
| Heart disease | 52,478 (24.2) | 27,084 (25.6) | 25,394 (22.9) |
| Cancer, any site | 46,138 (21.3) | 24,725 (23.3) | 21,413 (19.3) |
| Cerebrovascular disease | 14,070 (6.5) | 5,702 (5.4) | 8,368 (7.5) |
| COPD | 10,971 (5.1) | 6,035 (5.7) | 4,936 (4.4) |
| Other | 93,256 (43.0) | 42,378 (40.0) | 50,878 (45.8) |
| Dementia as cause of deathi | 29,037 (13.4) | 10,595 (10.0) | 18,442 (16.6) |
| Dementia from all sourcesi | 49,115 (22.6) | 19,508 (18.4) | 29,607 (26.7) |
| Insurance statusj | |||
| Medicaid | 24,386 (11.2) | 10,337 (9.8) | 14,049 (12.7) |
| Medicare | 140,342 (64.7) | 66,753 (63.0) | 73,589 (66.3) |
| Private | 116,028 (53.5) | 59,154 (55.8) | 56,874 (51.2) |
| Other | 10,601 (4.9) | 6,556 (6.2) | 4,045 (3.6) |
| Self-pay | 46,339 (21.4) | 23,842 (22.5) | 22,497 (20.3) |
| Unknowne | 42,991 (19.8) | 19,253 (18.2) | 23,738 (21.4) |
| Comorbidities—CCIk | |||
| 0 | 26,168 (12.1) | 11,869 (11.2) | 14,299 (12.9) |
| 1–2 | 69,454 (32.0) | 31,497 (29.7) | 37,957 (34.2) |
| 3–4 | 52,288 (24.1) | 26,415 (24.9) | 25,873 (23.3) |
| 5+ | 42,420 (19.6) | 24,192 (22.8) | 18,228 (16.4) |
| No record available/unknownl | 26,583 (12.3) | 11,951 (11.3) | 14,632 (13.2) |
| Inpatient Hospitalization at EOL | |||
| None within two years | 57,827 (26.7) | 27,075 (25.6) | 30,752 (27.7) |
| None within six months | 33,048 (15.2) | 15,403 (14.5) | 17,645 (15.9) |
| Any within last month | 84,278 (38.9) | 43,172 (40.8) | 41,106 (37.0) |
| Greater than 2 within six months | 56,463 (26.0) | 29,129 (27.5) | 27,334 (24.6) |
Sex as reported from UPDB. Unknown sex was excluded due to small numbers.
Age of decedent was referenced at the approximate time of the decedents' death.
Race and Ethnicity were based on the National Institute of Health definitions.45 Unknown race/ethnicity was excluded due to small numbers (79).
Marital status was defined as having a legal relationship with the decedent and was obtained from the death certificate. Unknown marital status was excluded due to small numbers (200).
Education—maximum level of education for decedents as reported from the UPDB was obtained. Unknown education was excluded due to small numbers (2340).
Median household income was based on residence at death and 2010 to 2014 American Community Survey five-year estimates.
Region—region was derived from zip code of decedents' residence as reported in death certificates and was categorized using 2018 Utah Department of Health classifications. Urban counties have a population density of >100 people per sq. mile; rural counties have a population density of <99 and >6 people per sq. mile, and Frontier counties have <6 people per sq. mile.46
Primary cause of death was obtained from death records and defined with ICD-9 and ICD-10 codes (Supplementary Table 1).
Dementia as a cause of death includes primary and secondary causes of death as indicated in death certificates (i.e., a decedent with cancer and dementia are listed in both categories) separately, as dementia is often a primary diagnosis but seldom the precipitating cause of death. Other sources of dementia diagnosis were inpatient, ambulatory surgery, and emergency department records. Please note that inpatient and ambulatory surgery records were from 1996 to 2016, and emergency department records were from 2000 to 2016.
Insurance coverage was only available for individuals who had an inpatient or ambulatory surgery or emergency record in the UPDB within the last year of decedents' death. Please note that inpatient and ambulatory surgery records were from 1996 to 2016 and emergency department records were from 2000 to 2016.
Comorbidity—CCI (Supplementary Table S2)37 was based on diagnosis data obtained from inpatient, ambulatory surgery, and emergency department records within the last two years of decedents' death. Please note that inpatient and ambulatory surgery records were from 1996 to 2016, and emergency department records were from 2000 to 2016.
Decedents did not have any inpatient/ambulatory surgery/emergency records in UPDB within two years before death.
Decedents' first-degree family characteristics
n = 743,874 first-degree family members were identified (n = 97,997, 13.2% partners/spouses; n = 26,993, 3.6% parents; n = 384,457, 51.7% adult children; n = 234,427, 31.5% siblings; Fig. 1). The median number of first-degree family was four (IQR = 6–2). Of the decedents, 16.0% did not have any identified first-degree family members at the time of death. Under half (45.2%) had a spouse (Table 3). A larger proportion of decedents without any first-degree family were female compared to male at the oldest age groups (80–89 years: 36% vs. 26%; 90+ years: 27% vs. 13%; Fig. 2).
Table 3.
Characteristics of Decedents' First-Degree Family
| Characteristic of decedents | Decedents n (%) | Males n (%) | Females n (%) |
|---|---|---|---|
| Size of first-degree family (parent, child, sibling, spouses) | |||
| 0 | 34,774 (16.0) | 14,695 (13.9) | 20,079 (18.1) |
| 1 | 34,167 (15.8) | 17,277 (16.3) | 16,890 (15.2) |
| 2 | 26,399 (12.2) | 12,151 (11.5) | 14,248 (12.8) |
| 3 | 26,739 (12.3) | 12,656 (11.9) | 14,083 (12.7) |
| 4–5 | 46,426 (21.4) | 23,002 (21.7) | 23,424 (21.1) |
| 6–10 | 44,534 (20.5) | 23,946 (22.6) | 20,588 (18.5) |
| +10 | 3,874 (1.8) | 2,197 (2.1) | 1,677 (1.5) |
| Median (Q1–Q3) | 4 (2–6) | 4 (2–6) | 4 (2–5) |
| Type and number of first-degree family members | |||
| At least one spouse, n (%) | 97,949 (45.2) | 62,022 (58.6) | 35,927 (32.4) |
| At least one parent, n (%) | 19,963 (9.2) | 11,438 (10.8) | 8,525 (7.7) |
| Child, median (Q1–Q3) | 3 (2–4) | 3 (2–4) | 2 (2–4) |
| Sibling, median (Q1–Q3) | 2 (1–3) | 2 (1–3) | 2 (1–3) |
As family size was an independent variable in the analyses, we excluded polygamous decedents from the data due to large numbers of linked children per decedent (7 identified decedents and 25 decedents with unclear polygamous status were removed; without excluding these individuals, the range of adult children was 0–26).
FIG. 2.
Size of first-degree family.
Aim 1: Identifying Characteristics Related to Having no First-Degree Family
Characteristics of decedents with no first-degree family
Chi-square tests of independence showed that proportions of those without family versus family differed by demographic, socioeconomic, context of death, and EOL health care utilization characteristics (p < 0.001; Table 4).
Table 4.
Bivariate (Chi-Square) Analysis of Demographic, Socioeconomics, Context of Death, and End-of-Life Health Care Utilization Differences between Decedents with and without First-Degree Family
| Variables | Decedents (n = 216,913) |
||
|---|---|---|---|
| No first-degree family n = 34,774 |
Any first-degree family n = 182,139 |
p a | |
| n (%) | n (%) | ||
| Sex | <0.001 | ||
| Male | 14,695 (42.3) | 91,229 (50.1) | |
| Female | 20,079 (57.7) | 90,910 (49.9) | |
| Age group | <0.001 | ||
| 20–49 | 2,059 (5.9) | 10,439 (5.7) | |
| 50–59 | 2,785 (8.0) | 14,542 (8.0) | |
| 60–69 | 4,495 (12.9) | 24,380 (13.4) | |
| 70–79 | 6,888 (19.8) | 42,113 (23.1) | |
| 80–89 | 11,198 (32.2) | 62,401 (34.3) | |
| 90–100 | 7,349 (21.1) | 28,264 (15.5) | |
| Race/ethnicity | <0.001 | ||
| White, Latinx | 25,726 (74.0) | 146,080 (80.2) | |
| Non-White or Latinx | 9,048 (26.0) | 36,059 (19.8) | |
| Maximum education | <0.001 | ||
| Less than high school | 12,945 (37.2) | 49,294 (27.1) | |
| High school | 11,308 (32.5) | 65,893 (36.2) | |
| More than high school | 10,521 (30.3) | 66,952 (36.8) | |
| Region | <0.001 | ||
| Urban | 25,532 (73.4) | 138,452 (76.0) | |
| Rural | 7,491 (21.5) | 34,564 (19.0) | |
| Frontier | 1,751 (5.0) | 9,123 (5.0) | |
| Primary cause of death | <0.001 | ||
| Cancer | 6,535 (18.8) | 39,603 (21.7) | |
| Heart disease | 8,157 (23.5) | 44,321 (24.3) | |
| COPD | 2,107 (6.1) | 8,864 (4.9) | |
| Cerebrovascular disease | 2,244 (6.4) | 11,826 (6.5) | |
| Other | 15,731 (45.2) | 77,525 (42.6) | |
| Dementia as cause of death | <0.001 | ||
| Yes | 5,271 (15.2) | 23,766 (13.0) | |
| No | 29,503 (84.8) | 158,373 (87.0) | |
| Inpatient hospitalizations in last six months | |||
| Two or more hospitalizations | 8,580 (24.7) | 47,883 (26.3) | <0.001 |
| None within two years before death | 15,329 (44.1) | 75,546 (41.5) | <0.001 |
| CCIb,c | <0.001 | ||
| 0 | 4,403 (12.7) | 21,765 (11.9) | |
| 1–2 | 10,980 (31.6) | 58,474 (32.1) | |
| 3–4 | 7,857 (22.6) | 44,431 (24.4) | |
| 5+ | 6,608 (19.0) | 35,812 (19.7) | |
| No records within two years before deathb,c | 4,926 (14.2) | 21,657 (11.9) | |
p-Values were calculated from chi-square tests for categorical variables and t tests for continuous variables.
Based on statewide inpatient and ambulatory surgery and emergency department records within two years before decedents' death.
Decedents did not have any inpatient/ambulatory surgery/emergency department records in UPDB within two years before death.
Demographic and socioeconomic characteristics
There were greater proportions of females (57.7% vs. 49.9%), decedents aged 90 to 100 (21.1% vs. 15.5%), decedents with less than a high school education (37.2% vs. 27.1%), and rural decedents (21.5% vs. 19.0%; all at p < 0.001) without first-degree family compared to those with first-degree family.
Context of death and EOL health care utilization
Proportions of decedents with and without at least one first-degree family member differed by primary cause of death, although these differences were small (Table 4). Having no first-degree family was associated with a lack of inpatient or ambulatory surgery records (proportions of decedents without first-degree family versus with first-degree family: 44.1% vs. 41.5%, p < 0.001).
Aim 2: Examine the Relationship Between First-Degree Family Characteristics and Multiple Hospitalizations During the Last Six Months of Life
Table 5 and Table 6 report analyses from the logistic regression models. In the first model, we examined the contributions of family size on the outcome of multiple EOL hospitalizations, while the second model examined the influence of types of family members within the first-degree family network.
Table 5.
Relationships between Sex, Age, Race, Cause of Death, Comorbidities, and Size of First-Degree Family on the Odds of Multiple Hospitalizations in the Last Six Months of Life
| Variables | Multiple hospitalizations Total n = 56,463 |
|
|---|---|---|
| aOR (95% CI) | p a | |
| Sex | ||
| Male (ref) | 1.00 | |
| Female | 1.09 (1.06–1.11) | <0.001 |
| Age group | ||
| 20–49 (ref) | 1.00 | |
| 50–59 | 0.78 (0.74–0.82) | <0.001 |
| 60–69 | 0.78 (0.74–0.82) | <0.001 |
| 70–79 | 0.75 (0.72–0.79) | <0.001 |
| 80–89 | 0.63 (0.60–0.66) | <0.001 |
| 90–100 | 0.44 (0.42–0.46) | <0.001 |
| Race | ||
| Non-Latinx (ref) | 1.00 | |
| Non-White or Latinx | 1.06 (1.03–1.09) | <0.001 |
| Education | ||
| Less than high school (ref) | 1.00 | |
| High school | 1.04 (1.01–1.07) | 0.003 |
| More than high school | 1.06 (1.03–1.09) | <0.001 |
| Region | ||
| Urban (ref) | 1.00 | |
| Rural | 0.93 (0.91–0.96) | <0.001 |
| Frontier | 1.02 (0.97–1.08) | 0.354 |
| Size of first-degree family | ||
| No first-degree family (ref) | 1.00 | |
| 1–3 first-degree family | 1.00 (0.96–1.03) | 0.781 |
| 4+ first-degree family | 0.98 (0.95–1.01) | 0.240 |
| Marital status | ||
| Married (ref) | 1.00 | |
| Nonmarried | 0.90 (0.88–0.92) | <0.001 |
| Primary cause of death | ||
| Other (ref) | 1.00 | |
| Cancer | 0.58 (0.57–0.60) | <0.001 |
| Heart disease | 0.76 (0.73–0.78) | <0.001 |
| COPD | 1.00 (0.95–1.05) | 0.991 |
| Cerebrovascular disease | 0.77 (0.74–0.81) | <0.001 |
| Dementia as cause of death | ||
| No (ref) | 1.00 | |
| Yes | 0.48 (0.46–0.50) | <0.001 |
| CCIb,c | ||
| 0 (ref) | 1.00 | |
| 1–2 | 3.86 (3.66–4.08) | <0.001 |
| 3–4 | 8.50 (8.06–8.97) | <0.001 |
| 5+ | 16.18 (15.33–17.09) | <0.001 |
| No records in two years | — | — |
Bold values indicate signficance at the p < 0.05 level.
p-Values were calculated from multivariable logistic regression models.
Based on statewide inpatient and ambulatory surgery and emergency department records within two years before decedents' death.
Decedents did not have any inpatient/ambulatory surgery/emergency department records in UPDB within two years before death.
Table 6.
Relationships between Sex, Age, Race, Cause of Death, Comorbidities, and Type of First-Degree Family on the Odds of Repeat Hospitalizations in the Last Six Months of Life
| Variablesa | Multiple hospitalizations Total n = 56,463 |
|
|---|---|---|
| aOR (95% CI) | p | |
| Sex | ||
| Male (ref) | 1.00 | |
| Female | 1.08 (1.05–1.10) | <0.001 |
| Age group | ||
| 20–49 (ref) | 1.00 | |
| 50–59 | 0.79 (0.75–0.83) | <0.001 |
| 60–69 | 0.80 (0.76–0.84) | <0.001 |
| 70–79 | 0.76 (0.72–0.81) | <0.001 |
| 80–89 | 0.64 (0.61–0.67) | <0.001 |
| 90–100 | 0.45 (0.42–0.47) | <0.001 |
| Race | ||
| Non-Latinx (Ref) | 1.00 | |
| Non-White or Latinx | 1.06 (1.03–1.09) | <0.001 |
| Education | ||
| Less than high school (ref) | 1.00 | |
| High school | 1.04 (1.02–1.07) | 0.002 |
| More than high school | 1.07 (1.04–1.10) | <0.001 |
| Region | ||
| Urban (ref) | 1.00 | |
| Rural | 0.94 (0.91–0.96) | <0.001 |
| Frontier | 1.02 (0.98–1.08) | 0.333 |
| Type of family network | ||
| Spouse | ||
| No (ref) | 1.00 | <0.001 |
| Yes | 1.08 (1.06–1.11) | |
| Any parents | ||
| No (ref) | 1.00 | 0.394 |
| Yes | 1.01 (0.97–1.06) | |
| Any children | ||
| No (ref) | 1.00 | 0.007 |
| Yes | 0.97 (0.94–0.99) | |
| Any siblings | 0.987 | |
| No (ref) | 1.00 | |
| Yes | 1.00 (0.98–1.02) | |
| Primary cause of death | ||
| Other (ref) | 1.00 | |
| Cancer | 0.59 (0.57–0.60) | <0.001 |
| Heart disease | 0.76 (0.74–0.78) | <0.001 |
| COPD | 1.00 (0.95–1.05) | 0.966 |
| Cerebrovascular disease | 0.77 (0.74–0.81) | <0.001 |
| Dementia as cause of death | ||
| No (ref) | 1.00 | |
| Yes | 0.48 (0.52–0.56) | <0.001 |
| CCIb,c | ||
| 0 (ref) | 1.00 | |
| 1–2 | 3.87 (3.67–4.08) | <0.001 |
| 3–4 | 8.52 (8.07–8.99) | <0.001 |
| 5+ | 16.22 (15.36–17.12) | <0.001 |
| No records in two years | — | — |
Bold values indicate signficance at the p < 0.05 level.
p-Values were calculated from multivariable logistic regression models.
Based on statewide inpatient and ambulatory surgery and emergency department records within two years before decedents' death.
Decedents did not have any inpatient/ambulatory surgery/emergency department records in UPDB within two years before death.
While the size of the first-degree family network was not predictive of multiple EOL hospitalizations, being female (aOR = 1.09, 95% CI = 1.06–1.11), non-White or Latinx (aOR = 1.06, 95% CI = 1.03–1.09), having a greater than high school education (aOR = 1.06, 95% CI = 1.03–1.09), and having more comorbidities (CCI 1–2, aOR = 3.86, 95% CI = 3.66–4.08; CCI 5+, aOR = 16.18, 95% CI = 15.33–17.09) were significantly associated with greater likelihoods of multiple EOL hospitalizations compared with males, non-Latinx White, decedents with less than a high school education and decedents with no documented comorbidities. Odds for multiple EOL hospitalizations were lower for older (50–59 years, aOR = 0.78, 95% CI = 0.74–0.82; 90–100 years, aOR = 0.44, 95% CI = 0.42–0.46), rural (aOR = 0.93, 95% CI = 0.91–0.96), nonmarried (aOR = 0.90, 95% CI = 0.88–0.92) and cancer (aOR = 0.58, 95% CI = 0.57–0.60), heart disease (aOR = 0.76, 95% CI = 0.74–0.78), cerebrovascular disease (aOR = 0.77, 95% CI = 0.74–0.81) or dementia (aOR = 0.48, 95% CI = 0.46–0.50; Table 5) decedents compared with decedents 20 to 49 years, urban dwelling, nonmarried decedents, or decedents with no dementia or causes of death other than cancer, heart disease, chronic obstructive pulmonary disease, and cerebrovascular disease.
Odds of multiple EOL hospitalizations were significantly higher for those with spouses (aOR = 1.08, 95% CI = 1.06–1.11). The odds for multiple EOL hospitalizations were lower for those with adult children; however, the magnitude was small (Table 6). The odds of the other predictors remained consistent in the second model. Sensitivity analyses examining multiple EOL hospitalizations in the final month of life confirmed the association between higher odds of hospitalizations among married decedents (Supplementary Table ST3 and ST4).
Discussion
Understanding the effects of family networks on decedent outcomes is of growing interest.39–41 This is the first study to examine relationships between first-degree family characteristics of decedents and EOL outcomes in a U.S. population dataset and among the few that examines multiple EOL hospitalizations, making comparisons to other studies difficult. The lack of consistency in operationalization of what constitutes a burdensome EOL hospitalization across studies affects comparisons—while we found that 26% of our sample had multiple EOL hospitalizations, others have reported 8% to 68.6%42 in other population-based studies in other countries (a U.S. study has reported that 15% of avoidable readmissions were due to EOL issues).43
Decedents without close family at EOL
In our study, we found that 16.0% (n = 34,774) had no first-degree family. However, as extended family was not included, the proportion of deceased individuals who were kinless is likely lower. Population-based examination of decedents in Denmark that included grandchildren and great-grandchildren reported that ∼10% had no family,44 suggesting that different strategies may be needed to understand EOL outcomes among decedents with no family.
The impact of first-degree family size and composition on multiple EOL hospitalizations
Results from Aim 2 did not support that family size was associated with greater risk of multiple EOL hospitalizations. The null finding of size of family and multiple EOL hospitalizations could be due to our inability to examine the wider family beyond the first-degree network, nonfamily networks of support (e.g., friendships, religious organizations, paid support), or families with multiple generations coresiding in a household (e.g., parents, decedents, adult children), which could affect findings.
Our findings suggest that having a spouse impacts multiple EOL hospitalizations. This phenomenon warrants further research as EOL care has implications for bereaved spouse health care utilization—data from the Health and Retirement Study found that Medicare spending one year postdeath among bereaved spouses of decedents who died in hospital was $3,106 higher than spouses of patients who died in other settings.45 Family caregivers are disproportionately spouses,46 and more than half of spousal caregivers are the sole provider of care at EOL.47 Our findings suggest that spousal caregivers may require more support, particularly those caring for younger, racial/ethnic minority, or female patients.
Our analyses accounted for the influence of age, family characteristics, and comorbidities on multiple EOL hospitalizations. The results strongly suggest that sex, race/ethnicity, geography, marital status, and health context contribute to the likelihood of EOL hospitalizations. In addition, we found that higher education was positively associated with multiple EOL hospitalizations. While nationally representative studies find that higher income and insured decedents are more likely to avoid hospital deaths and utilize hospice,48 socioeconomic status may also impact ability of decedents to access various avenues of health care, including inpatient care at EOL. These factors are important to explore in future analyses to provide insight into more equitable EOL care; however, it is outside the scope of this article to examine interactions of age, family composition, socioeconomic status, and race. Nevertheless, this data resource allows for future examinations of social determinants of health pertaining to late life outcomes.
The AARP and the National Alliance for Caregiving's report, Caregiving in the United States 2020, report that rural caregivers face greater strains, including caring for multiple care recipients.1 Our findings suggest that those living in rural regions had lower odds of multiple EOL hospitalizations; however, our sensitivity analyses found that those living in the most underpopulated regions of the state (i.e., frontier regions) had increased odds of multiple hospitalizations in the last month of life. This finding has implications for further investigation of EOL transitions in care, particularly for those with difficulties in accessing hospice care.
Strengths and limitations
While population registry data are less vulnerable to inadequately powered samples, there are some key limitations. Over 90% of decedents in our cohort were white, and 79.2% were non-Latinx White (compared to 60.1% non-Latinx White in the overall population). While the population of Utah is diversifying, this demographic is reflective of the population that is 65 and above in Utah.49
The proportion of rural and frontier decedents in the cohort is higher than the rural dwelling population in the United States (24.4% vs. 19.5%). However, older adults are more likely to live in rural areas and are a challenging population to include in research.50 The representation of rural and frontier decedents is thus a strength of this research.
Given the nature of the UPDB as a repository of linked data for Utah, family members outside Utah during the study period or who temporarily relocate without officially updating their residential address are unavailable. Data from the American Community Survey between 2005 and 2009 indicate that ∼18% of Utah residents move from other states to Utah and 14% move from Utah to other states annually51—these family members are not represented in this research.
Chronic illness may be a risk factor for death by suicide52 or accidental overdoses particularly for opioids.53 An estimated 663 suicides and 650 overdose deaths occur annually in Utah54,55; however, chronically ill patients within this population are not represented in the outcome of multiple EOL hospitalizations in this study, limiting the generalizability of findings for these at-risk populations.
Although we are able to study the structural features of first-degree family (size and composition), we are unable to examine the functional aspects of family support, such as caregiving involvement or quality of family relationships. Moreover, our data are limited to first-degree biological family members and spouses and may overlook other key caregiving relationships such as nonbiological family members (e.g., families of choice, stepchildren), extended family and multigenerational households, and nonfamily relationships such as friends, coworkers, or religious networks.
In addition, our ability to accurately account for hospice and palliative care use with these existing sources is limited to individuals who had an inpatient or ambulatory record. Other services such as home care are unavailable as yet. Sources such as electronic health documentation for social support data may be limited by inconsistencies in documentation; nevertheless, characterizing the EOL caregiving network with other sources of data is a needed next step to advance population caregiving research. Other strategies such as mortality follow-back surveys and extending linkages to second- and third-degree family may be needed to better contextualize the caregiving relationship.
Conclusion
The relationship between family networks on EOL care utilization has not been widely explored, with much of the existing literature using convenience samples5 and focusing on spouse or dyadic caregiving relationships.56 This study provides a population-based analysis of the associations between family structure and size and EOL outcomes, which provides the groundwork for future exploration of the palliative care needs of families in EOL care.
Supplementary Material
Acknowledgments
The authors thank the Pedigree and Population Resource of Huntsman Cancer Institute, University of Utah (funded, in part, by the Huntsman Cancer Foundation) for its role in the ongoing collection, maintenance, and support of the Utah Population Database (UPDB). The authors also acknowledge partial support for the UPDB and all datasets within the UPDB through grant P30 CA2014 from the National Cancer Institute, University of Utah and from the University of Utah's program in Personalized Health and Center for Clinical and Translational Science.
Authors' Contributions
D.L.T. contributed to the acquisition and interpretation of the data, drafting, revising, and providing final approval of the article, and accepts accountability for all aspects of the work. K.A.O. contributed to the conception and design, acquisition, and interpretation of the data, revising and final approval of the article, and accepts accountability for all aspects of the work. H.M. contributed to the acquisition and analysis of the data, revising and final approval of the article, and accepts accountability for all aspects of the work. R.L.U. contributed to the interpretation of the data, revising and providing final approval of the article, and accepts accountability for all aspects of the work. K.R.S. contributed to the conception and design of the work, interpretation of the data, revising and providing final approval of the article, and accepts accountability for all aspects of the work. C.S. contributed to the interpretation of the data, revising and providing final approval of the article, and accepts accountability for all aspects of the work. M.H contributed to the acquisition and interpretation of the data, revising and providing final approval of the article, and accepts accountability for all aspects of the work. L.E. contributed to the conception and design of the work, acquisition and interpretation of the data, drafting, revising, and providing final approval of the article, and accepts accountability for all aspects of the work.
Funding Statement
This study was funded by the University of Utah's College of Nursing led interdisciplinary Family Caregiving Collaborative and Mount Sinai's Claude D. Pepper Older Americans Independence Center (NIH/NIA P30AG028741 and NIH/NIA 5K07AG060270). The first author is supported by the University of Utah Vice President's Clinical and Translational Research Scholars Program and the College of Nursing Family Caregiving Collaborative Caregiving Scholar program. Research was supported by the NCRR grant, “Sharing Statewide Health Data for Genetic Research” (R01 RR021746, G. Mineau, PI), with additional support from the Utah State Department of Health and the University of Utah.
Author Disclosure Statement
No competing financial interests exist.
Supplementary Material
References
- 1. AARP and National Alliance for Caregiving. Caregiving in the United States 2020. Washington, DC: AARP; 2020. [Google Scholar]
- 2. Ornstein KA, Kelley AS, Bollens-Lund E, Wolff JL: A national profile of end-of-life caregiving in the United States. Health Aff 2017;36:1184–1192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Angus DC, Barnato AE, Linde-Zwirble WT, et al. : Use of intensive care at the end of life in the United States: An epidemiologic study. Crit Care Med 2004;32:638–643. [DOI] [PubMed] [Google Scholar]
- 4. Institute of Medicine: Dying in America: Improving Quality and Honoring Individual Preferences Near the End of Life. Washington, DC: The National Academies Press, 2015. [PubMed] [Google Scholar]
- 5. Stajduhar KI, Funk L, Toye C, et al. : Part 1: Home-based family caregiving at the end of life: A comprehensive review of published quantitative research (1998–2008). Palliat Med 2010;24:573–593. [DOI] [PubMed] [Google Scholar]
- 6. Kelley AS, McGarry K, Bollens-Lund E, et al. : Residential setting and the cumulative financial burden of dementia in the 7 years before death. J Am Geriatr Soc 2020;68:1319–1324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. LaValley SA: End-of-life caregiver social support activation: The roles of hospice clinicians and professionals. Qual Health Res 2018;28:87–97. [DOI] [PubMed] [Google Scholar]
- 8. National Academies of Sciences, Engineering, and Medicine: Families Caring for an Aging America. Washington, DC: National Academies of Sciences, Engineering, and Medicine, 2016. [PubMed] [Google Scholar]
- 9. Gomes B, Calanzani N, Gysels M, et al. : Heterogeneity and changes in preferences for dying at home: A systematic review. BMC Palliat Care 2013;12:7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Lage DE, DuMontier C, Lee Y, et al. : Potentially burdensome end-of-life transitions among nursing home residents with poor-prognosis cancer. Cancer 2020;126:1322–1329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Duncan I, Ahmed T, Dove H, et al. : Medicare cost at end of life. Am J Hosp Palliat Care 2019;36:705–710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Aldridge MD, Kelley AS: The myth regarding the high cost of end-of-life care. Am J Public Health 2015;105:2411–2415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Cardona-Morrell M, Kim JCH, Brabrand M, et al. : What is inappropriate hospital use for elderly people near the end of life? A systematic review. Eur J Intern Med 2017;42:39–50. [DOI] [PubMed] [Google Scholar]
- 14. Cohen-Mansfield J, Skornick-Bouchbinder M, Brill S: Trajectories of end of life: A systematic review. J Gerontol B Sci Soc Sci 2017;73:564–572. [DOI] [PubMed] [Google Scholar]
- 15. Karanth S, Rajan SS, Sharma G, et al. : Racial-ethnic disparities in end-of-life Care quality among lung cancer patients: SEER-Medicare based study. J Thorac Oncol 2018;13:1083–1093. [DOI] [PubMed] [Google Scholar]
- 16. Miesfeldt S, Murray K, Lucas L, et al. : Association of age, gender, and race with intensity of end-of-life care for Medicare beneficiaries with cancer. J Palliat Med 2012;15:548–554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Miller SC, Lima JC, Intrator O, et al. : Palliative care consultations in nursing homes and reductions in acute care use and potentially burdensome end-of-life transitions. J Am Geriatr Soc 2016;64:2280–2287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wagner E, Patrick DL, Khandelwal N, et al. : The influence of multimorbidity on health care utilization at the end of life for patients with chronic conditions. J Palliat Med 2019;22:1260–1265. [DOI] [PubMed] [Google Scholar]
- 19. Sudore RL, Casarett D, Smith D, et al. : Family involvement at the end-of-life and receipt of quality care. J Pain Symptom Manage 2014;48:1108–1116. [DOI] [PubMed] [Google Scholar]
- 20. Spillman BC, Freedman VA, Kasper JD, Wolff JL: Change over time in caregiving networks for older adults with and without dementia. J Gerontol B Psychol Sci Soc Sci 2020;75:1563–1572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Holt-Lunstad J: The potential public health relevance of social isolation and loneliness: Prevalence, epidemiology, and risk factors. Public Policy Aging Rep 2018;27:127–130. [Google Scholar]
- 22. Valtorta NK, Moore DC, Barron L, et al. : Older adults' social relationships and health care utilization: A systematic review. Am J Public Health 2018;108:e1–e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Manemann SM, Chamberlain AM, Roger VL, et al. : Perceived social isolation and outcomes in patients with heart failure. J Am Heart Assoc 2018;7:e008069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Rico-Uribe LA, Caballero FF, Martín-María N, et al. : Association of loneliness with all-cause mortality: A meta-analysis. PLoS One 2018;13:e0190033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hayford SR: Marriage (still) matters: The contribution of demographic change to trends in childlessness in the United States. Demography 2013;50:1641–1661. [DOI] [PubMed] [Google Scholar]
- 26. Cherlin AJ: Demographic trends in the United States: A review of research in the 2000s. J Marriage Fam 2010;72:403–419. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Pietromonaco PR, Uchino B, Dunkel Schetter C: Close relationship processes and health: Implications of attachment theory for health and disease. Health Psychol 2013;32:499–513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Antonucci TC, Ajrouch KJ, Birditt KS: The convoy model: Explaining social relations from a multidisciplinary perspective. Gerontologist 2014;54:82–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Fuller HR, Ajrouch KJ, Antonucci TC: The convoy model and later-life family relationships. J Fam Theory Rev 2020;12:126–146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Leonard R, Horsfall D, Rosenberg J, Noonan K: Carer experience of end-of-life service provision: A social network analysis. BMJ Support Palliat Care 2020;10:e20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Plick NP, Ankuda CK, Mair CA, et al. : A national profile of kinlessness at the end of life among older adults: Findings from the Health and Retirement Study. J Am Geriatr Soc 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Arbeeva LS, Hanson HA, Arbeev KG, et al. : How well does the family longevity selection score work: A validation test using the Utah Population Database. Front Public Health 2018;6:277–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Kristensen MS, Thygesen LC, Tay DL, et al. : Size and composition of family networks of decedents: A nationwide register-based study. Palliat Med 2021:0269216321998602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Lupo PJ, Danysh HE, Plon SE, et al. : Family history of cancer and childhood rhabdomyosarcoma: A report from the Children's Oncology Group and the Utah Population Database. Cancer Med 2015;4:781–790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Norton MC, Hatch DJ, Munger RG, Smith KR: Family member deaths in childhood predict systemic inflammation in late life. Biodemography Soc Biol 2017;63:104–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Norton MC, Smith KR, Østbye T, et al. Early parental death and remarriage of widowed parents as risk factors for Alzheimer disease: The Cache County study. Am J Geriatr Psychiatry 2011;19:814–824. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. van den Berg N, Rodríguez-Girondo M, van Dijk IK, et al. : Longevity defined as top 10% survivors and beyond is transmitted as a quantitative genetic trait. Nat Commun 2019;10:35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Charlson ME, Pompei P, Ales KL, MacKenzie CR: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40:373–383. [DOI] [PubMed] [Google Scholar]
- 39. Koehly LM, Ashida S, Schafer EJ, Ludden A: Caregiving networks-Using a network approach to identify missed opportunities. J Gerontol B Psychol Sci Soc Sci 2015;70:143–154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Waldrop DP: Caregiving systems at the end of life: How informal caregivers and formal providers collaborate. Fam Soc 2006;87:427–437. [Google Scholar]
- 41. Kumar V, Ankuda CK, Aldridge MD, et al. : Family caregiving at the end of life and hospice use: A national study of Medicare beneficiaries. J Am Geriatr Soc 2020;68:2288–2296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Pivodic L, Pardon K, Miccinesi G, et al. : Hospitalisations at the end of life in four European countries: A population-based study via epidemiological surveillance networks. J Epidemiol Community Health 2016;70:430–436. [DOI] [PubMed] [Google Scholar]
- 43. Donzé J, Lipsitz S, Schnipper JL: Risk factors for potentially avoidable readmissions due to end-of-life care issues. J Hosp Med 2014;9:310–314. [DOI] [PubMed] [Google Scholar]
- 44. Ornstein KA, Mair C, Kristensen MS, et al. : Kinlessness at the time of death: Implications for end-of-life decision-making. Innov Aging 2019;3(Suppl 1):S237–S237. [Google Scholar]
- 45. Ornstein KA, Garrido MM, Siu AL, et al. : Impact of in-hospital death on spending for bereaved spouses. Health Serv Res 2018;53(Suppl Suppl 1):2696–2717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Abernethy A, Burns C, Wheeler J, Currow D: Defining distinct caregiver subpopulations by intensity of end-of-life care provided. Palliat Med 2009;23:66–79. [DOI] [PubMed] [Google Scholar]
- 47. Ornstein KA, Wolff JL, Bollens-Lund E, et al. : Spousal caregivers are caregiving alone in the last years of life. Health Aff 2019;38:964–972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Orlovic M, Smith K, Mossialos E: Racial and ethnic differences in end-of-life care in the United States: Evidence from the Health and Retirement Study (HRS). SSM Popul Health 2019;7:100331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Kem C: Gardner Policy Institute. Utah state and county annual population estimates by single-year of age, sex, and race/ethnicity: 2010–2019. 2020. www.aacnnursing.org/Portals/42/News/Factsheets/Nursing-Shortage-Factsheet-2017.pdf (Last accessed May 1, 2021).
- 50. Explore Census Data: https://data.census.gov/cedsci/table?q=US+rural&tid=DECENNIALAIAN2010.HCT1 (Last accessed May 1, 2021).
- 51. United States Census Bureau: State-to-State Migration Flows. https://www.census.gov/data/tables/time-series/demo/geographic-mobility/state-to-state-migration.html (Last accessed May 1, 2021).
- 52. Gürhan N, Beşer NG, Polat Ü, Koç. M: Suicide risk and depression in individuals with chronic illness. Community Ment Health J 2019;55:840–848. [DOI] [PubMed] [Google Scholar]
- 53. Wilder CM, Miller SC, Tiffany E, et al. : Risk factors for opioid overdose and awareness of overdose risk among veterans prescribed chronic opioids for addiction or pain. J Addict Dis 2016;35:42–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Department of Health: Complete health indicator report of suicide. IBIS. https://ibis.health.utah.gov/ibisph-view/indicator/complete_profile/SuicDth.html (Last accessed May 1, 2021).
- 55. Centers for Disease Control and Prevention: Stats of the State of Utah. https://www.cdc.gov/nchs/pressroom/states/utah/utah.htm (Last accessed May 1, 2021).
- 56. Wolff JL, Dy SM, Frick KD, Kasper JD: End-of-life care: Findings from a national survey of informal caregivers. Arch Intern Med 2007;167:40–46. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.