Abstract
Background:
Studies examining end-of-life (EOL) care in older cancer patients are scarce, and prior studies have not accounted for gradients of cognitive impairment (COG-I). We examine EOL care patterns across COG-I gradients, hypothesizing that greater COG-I severity is associated with lower odds of receiving aggressive EOL care.
Methods:
Using data from the linked Surveillance Epidemiology and End Results (SEER) -Medicare: Minimum Data Set (MDS) 3.0, patients with nursing facility stays (NFS) were identified who died with metastatic cancer from 2013–2017. Markers of aggressive EOL care were: cancer-directed treatment, intensive care unit admission, >1 emergency department visit, or >1 hospitalization in the last 30 days of life, hospice enrollment in the last 3 days of life, and in-hospital death. In addition to descriptive analysis, we conducted multivariable logistic regression analysis to evaluate the independent association between COG-I severity and receipt of aggressive EOL care.
Results:
Of the 40,833 patients in our study population, 49.2% were cognitively intact; 24.4% had mild COG-I; 19.7% had moderate COG-I; and 6.7% had severe COG-I. The percent of patients who received aggressive EOL care was 62.6% and 74.2% among those who were cognitively intact and those with severe COG-I, respectively. Compared with cognitively intact patients, those with severe COG-I had 86% higher odds of receiving any type of aggressive EOL care (adjusted odds ratio (aOR): 1.86 (95% confidence interval: 1.70–2.04)), which were primarily associated with higher odds of in-hospital death. The odds of in-hospital death associated with severe COG-I were higher among those with short- than with long-term stays (aOR: 2.58 (2.35–2.84) and aOR: 1.40 (1.17–1.67), respectively).
Conclusions:
Contrary to our hypothesis, aggressive EOL care in patients with NFS with metastatic cancer was highest among those suffering severe COG-I. These findings can inform the development of interventions to help reduce aggressive EOL care in this patient population.
Keywords: metastatic cancer, aggressive end-of-life care, nursing home status
Introduction:
The number of adults over age 65 with incident cancer in the U.S. is projected to reach 2.3 million by 2030,1 and those experiencing dementia are forecasted to be at 10.5 million by 2050.2 End-of-life (EOL) care for persons with cancer and dementia has been characterized as suboptimal, with burdensome interventions and transfers to acute care settings.3,4 Decisions regarding cancer treatment are complicated by the high multimorbidity burden in people with dementia,5 which increases considerably across gradients of cognitive impairment (COG-I).5
Prior studies on cancer and EOL care in older adults with cancer with nursing facility stays (NFS) have largely not analyzed cognitive impairment.6,7 Studies using the linked Surveillance, Epidemiology, and End Results (SEER) and Medicare administrative data to study EOL care in patients with COG-I have identified dementia by the presence (or absence) of ICD-9 diagnosis codes;8,9 this approach is limited in identifying COG-I severity gradients. Updated ICD-10 codes may allow for greater granularity in capturing the severity of COG-I, if severity is coded in the administrative records.
The linked SEER-Medicare and nursing home Minimum Data Set (MDS) data offers an opportunity to study EOL care across gradients of COG-I rather than simply as the presence/absence of dementia. Although only 17% of older adults with moderate or severe incident dementia receive their care in a nursing facility,10 the linked SEER-Medicare-MDS database provides a glimpse of how EOL care varies across gradients of COG-I. In this study, we hypothesized that increased COG-I severity in older patients with metastatic cancer and NFS would be associated with lower odds of receiving aggressive EOL care.
Methods:
We examined aggressive EOL care receipt in patients with metastatic cancer and NFS across COG-I gradients. This study was approved by our university institutional review board (#20201463).
Data Sources:
We used data from the 2013–2017 linked SEER-Medicare, and MDS.11 The SEER registry captures incident cancer cases, covering 48% of the population across the US. It includes demographic variables and tumor characteristics (e.g., anatomic cancer site, date of cancer diagnosis, and cancer stage).
The Medicare Beneficiary Summary file allowed the identification of beneficiaries who received care through the traditional fee-for-service (FFS) system. We used claims data, including the Medicare Provider, Analysis, and Review (MedPAR), the Outpatient Standard Analytic File (SAF), the Carrier SAF, Durable Medical Equipment SAF, and Hospice SAF to characterize healthcare utilization, including indicators of aggressive EOL care, with a 6-month lookback period from death that also covered the latter half of 2012. We used diagnosis codes to flag comorbid conditions from Elixhauser’s list of number of comorbidities,12 and identify individuals with metastatic disease.
The MDS consists of clinical assessment data for individuals admitted to a nursing facility, both with short- and long-term stays. Clinical assessment is conducted upon admission, quarterly, at discharge, and when there is a change in health status. The broad array of MDS variables include chronic conditions, functional status, and COG-I, which we used to derive severity, using a validated algorithm.13
Study Population:
The study population included FFS beneficiaries residing in SEER registry areas aged ≥66 years at time of cancer diagnosis; with at least one comprehensive MDS assessment in the last 6 months of life, and with all-cause death in years 2013–2017. The study population was limited to patients with NFS with some of the most common cancers, and with varying case-fatality rates: breast, colorectal, lung, pancreas, or prostate cancer (n=117,214). We limited our study to individuals with diagnoses for metastatic cancer in the last 6 months of life (n=55,149 excluded) using the Elixhauser list of comorbidities to identify the presence of metastatic cancer.12 We excluded individuals with missing data for COG-I (n=21,100) and excluded non-Hispanic Indigenous residents due to small sample size (n=132) resulting in our study population (n=40,833).
Study Variables:
Outcome Variables:
Primary outcomes were claims-based variables indicating aggressive EOL care in the last 30 days of life: cancer-directed treatment; intensive care (ICU) admission, >1 emergency department (ED) visit, >1 hospitalization; hospice enrollment in the last 3 days of life; and in-hospital death. A binary variable was created to indicate receipt of any of the aggressive EOL care indicators.
Independent Variables:
Our main independent variable of interest was COG-I, derived from the validated Cognitive Function Scale (CFS).13 The CFS integrates measures of cognition using variables from the MDS 3.0, including the Brief Interview for Mental Status (BIMS), which is a short, standardized performance-based cognitive screener, and the former Cognitive Performance Score (CPS) scale used under MDS 2.0. The CFS is comprised of four levels: “Cognitively Intact” (BIMS scores 13–15); “Mildly Impaired” (BIMS scores 8–12 and 0–2 on CPS); “Moderately Impaired” (BIMS scores 0–7 and 3–4 on CPS); and “Severely Impaired” (unable to complete BIMS and 5–6 on CPS). In the case of patients who were unable to participate or complete the interview questions in a BIMS assessment, they were instead assessed by nursing staff in measures used towards calculating their CPS score.
Demographic variables included age at death (66–74, 75–84, and 85+); sex (male, female); and race/ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic Asian or Pacific Islander, and Hispanic). Categorization of variables reflects coding in SEER and MDS documentation. Cancer type was based on anatomic cancer site. Additional variables included dual eligibility status (yes/no) as an additional marker for socioeconomic disadvantage and heightened vulnerability for adverse health outcomes;14 and the number of Elixhauser comorbidities, which was dichotomized as ≤4 or ≥5. To account for previously documented differences in EOL care by length of stay,15,16 we identified patients with short- (<= 100 days in a nursing facility) and long-term stays (more than 100 days) as defined by the Centers for Medicare & Medicaid Services and MDS guidelines.17
Analytic Approach:
In addition to descriptive analyses, we developed multivariable logistic regression models to assess the independent association between COG-I severity and each aggressive EOL indicator, as well as receipt of any aggressive EOL care. We conducted sensitivity analysis using stratified models to determine whether the results differed between patients with short- and long-NFS. Results are presented as adjusted odds ratios (aOR) with 95% confidence intervals.
Analyses were performed using SAS 9.4 and R version 4.1.1.
Results:
Our study population is described in Table 1. Of the 40,833 patients with metastatic cancer and NFS, 49.2% were cognitively intact, 24.4% had mild COG-I, 19.7% had moderate COG-I, and 6.7% had severe COG-I. Overall, 82.7% had short-term NFS—ranging from 75.6% among those with severe COG-I to 87.0% among cognitively intact patients.
Table 1.
Patient demographics stratified by gradients of cognitive impairment among Medicare FFS enrollees with nursing facility stays, 66 years of age or older, who died between 2013–2017, and had active metastatic cancer in their last 6 months of life.
| Total, n (%)
(n=40,833) |
Degree of Cognitive Impairment, n (% of total) | ||||
|---|---|---|---|---|---|
| No COG-I (n=20,109) |
Mild COG-I (n=9,952) |
Moderate COG-I (n=8,048) |
Severe COG-I (n=2,724) |
||
| Age at Death, Years, n (%) | |||||
| 66–74 | 11,698 (28.6) | 6,555 (32.6) | 2,645 (26.6) | 1,815 (22.6) | 683 (25.0) |
| 75–84 | 17,906 (43.9) | 8,898 (44.2) | 4,348 (43.7) | 3,479 (43.2) | 1,181 (43.4) |
| 85+ | 11,229 (27.5) | 4,656 (23.2) | 2,959 (29.7) | 2,754 (34.2) | 860 (31.6) |
| Mean (SD) | 79.6 (7.4) | 78.7 (7.2) | 80 (7.4) | 80.9 (7.5) | 80.4 (7.6) |
| Median (Q1-Q3) | 79 (74–85) | 78 (73–84) | 80 (74–86) | 81 (75–87) | 80 (74–86) |
| Sex, n (%) | |||||
| Male | 20,105 (49.2) | 9,611 (47.8) | 5,206 (52.3) | 4,033 (50.1) | 1,255 (46.1) |
| Female | 20,728 (50.8) | 10,498 (52.2) | 4,746 (47.7) | 4,015 (49.9) | 1,469 (53.9) |
| Race/Ethnicity, n (%) | |||||
| Non-Hispanic White | 32,711 (80.1) | 16,749 (83.3) | 7,996 (80.3) | 5,997 (74.5) | 1,969 (72.3) |
| Non-Hispanic Black | 4,300 (10.6) | 1,741 (8.7) | 1,053 (10.6) | 1,109 (13.8) | 397 (14.6) |
| Non-Hispanic API* | 1,767 (4.3) | 686 (3.4) | 404 (4.1) | 487 (6.1) | 190 (7.0) |
| Hispanic | 2,055 (5.0) | 933 (4.6) | 499 (5.0) | 455 (5.6) | 168 (6.1) |
| Marital Status, n (%) | |||||
| Married | 15,306 (37.5) | 7,704 (38.3) | 3,659 (36.8) | 2,951 (36.6) | 992 (36.4) |
| Widowed | 14,452 (35.4) | 6,891 (34.3) | 3,638 (36.6) | 2,969 (36.9) | 954 (35.0) |
| Divorced/Separated | 5,161 (12.6) | 2,605 (13.0) | 1,268 (12.7) | 972 (12.1) | 316 (11.6) |
| Never Married | 5,099 (12.5) | 2,543 (12.6) | 1,178 (11.8) | 996 (12.4) | 382 (14.0) |
| Unknown | 815 (2.0) | 366 (1.8) | 209 (2.1) | 160 (2.0) | 80 (3.0) |
| Dual Eligible Status, n (%) | |||||
| No | 28,435 (69.6) | 15,332 (76.2) | 6,746 (67.8) | 4,798 (59.6) | 1,559 (57.2) |
| Yes | 12,398 (30.4) | 4,777 (23.8) | 3,206 (32.2) | 3,250 (40.4) | 1,165 (42.8) |
| Cancer Type, n (%) | |||||
| Lung | 17,510 (42.9) | 8,733 (43.4) | 4,340 (43.6) | 3,329 (41.4) | 1,108 (40.7) |
| Breast | 5,128 (12.6) | 2,490 (12.4) | 1,160 (11.7) | 1,093 (13.6) | 385 (14.1) |
| Colorectal | 7,152 (17.5) | 3,432 (17.1) | 1,783 (17.9) | 1,415 (17.6) | 522 (19.2) |
| Pancreatic | 3,807 (9.3) | 2,089 (10.4) | 842 (8.5) | 628 (7.8) | 248 (9.1) |
| Prostate | 7,236 (17.7) | 3,365 (16.7) | 1,827 (18.3) | 1,583 (19.6) | 461 (16.9) |
| Length of Care in NF, n (%) | |||||
| Short-term (≤ 100 days) | 33,779 (82.7) | 17,497 (87.0) | 8,205 (82.4) | 6,017 (74.8) | 2,060 (75.6) |
| Long-term (> 100 days) | 7,054 (17.3) | 2,612 (13.0) | 1,747 (17.6) | 2,031 (25.2) | 664 (24.4) |
| Year of Death, n (%) | |||||
| 2013 | 7,633 (18.7) | 3,451 (17.2) | 1,921 (19.3) | 1,651 (20.5) | 610 (22.4) |
| 2014 | 8,273 (20.3) | 3,948 (19.6) | 2,054 (20.7) | 1,687 (21.0) | 584 (21.4) |
| 2015 | 8,296 (20.3) | 4,184 (20.8) | 1,952 (19.6) | 1,602 (19.9) | 558 (20.5) |
| 2016 | 8,431 (20.6) | 4,328 (21.5) | 2,001 (20.1) | 1,600 (19.9) | 502 (18.4) |
| 2017 | 8,200 (20.1) | 4,198 (20.9) | 2,024 (20.3) | 1,508 (18.7) | 470 (17.3) |
| Elixhauser Comorbidity Count, n (%) | |||||
| ≤4 | 10,009 (24.5) | 5,000 (24.9) | 2,355 (23.7) | 1,984 (24.7) | 670 (24.6) |
| ≥5 | 30,824 (75.5) | 15,109 (75.1) | 7,597 (76.3) | 6,064 (75.3) | 2,054 (75.4) |
| Mean (SD) | 6.4 (2.6) | 6.4 (2.6) | 6.5 (2.7) | 6.4 (2.7) | 6.4 (2.6) |
| Median (Q1-Q3) | 6 (5–8) | 6 (5–8) | 6 (5–8) | 6 (5–8) | 6 (5–8) |
Abbreviations: NF = Nursing Facility; FFS = fee-for-service, COG-I = cognitive impairment; API = Asian or Pacific Islander
As seen in Supplementary Table S1 nearly two thirds (63.6%) of all patients received any type of aggressive EOL care, (62.6% among cognitively intact; 74.2% among those with severe COG-I). Receipt of any aggressive EOL care was the result of >1 ED visits ( 22.1% among cognitively intact; 27.0% among those with severe COG-I); and in-hospital death (35.7% among cognitively intact; 55.9% among those with severe COG-I). Receipt of cancer-directed treatment was highest among cognitively intact patients (21.2%) and lowest among severe COG-I patients (15.8%).
As seen in Figure 1, patients with severe COG-I had 86% higher odds than those who were cognitively intact to receive any type of aggressive EOL care (aOR:1.86 (95% confidence interval: 1.70–2.04)). The aORs associated with severe COG-I were highest for >1 ED visits (aOR:1.38 (1.26–1.52)) and for in-hospital death (aOR:2.29 (2.11–2.49)). Conversely, the aORs associated with severe COG-I were lowest for cancer-directed treatment (aOR:0.76 (0.68–0.85)) and for late-entry into hospice (aOR:0.71 (0.62–0.81)).
Figure 1:
Unadjusted and Adjusted Odds Ratios for Markers of Aggressive End-of-Life Care by Levels of Cognitive Impairment
The following are observations from our multivariable analysis, which examined correlates of aggressive EOL care markers, after adjusting for all variables listed in Table 1 and holding COG-I severity constant (Supplementary Table S2). First, compared with patients ages 66–74 at death, those age 85+ had significantly lower odds of receiving most types of aggressive EOL care; age at death was not associated with late entry into hospice and in-hospital death. Those with dual eligibility status had lower odds of receiving aggressive EOL care than those without. The only exception was in-hospital death, which did not reach statistical significance. Second, compared to non-Hispanic White patients, non-Hispanic Black patients had higher odds of receiving any aggressive EOL care. However, non-Hispanic Black patients had lower odds than non-Hispanic White patients of receiving cancer-directed treatment (aOR:0.90 (0.82–0.98)). In addition, minority patients had lower odds than non-Hispanic White patients of entering hospice in the last 3 days of life. Compared with lung cancer patients, those with other types of cancer had significantly lower odds of receiving aggressive EOL care; and compared with patients who had ≤4 comorbid conditions, those with higher comorbidity burden had higher odds of receiving any aggressive EOL care, particularly for hospital admission (aOR:2.93 (2.69–3.19)) and ICU admission (aOR:2.72 (2.54–2.92)).
In our sensitivity analysis (Table 2), we observed marked differences in the odds of receiving aggressive EOL care between patients with short- and long-term NFS with the aOR associated with severe COG-I (compared to cognitively intact) being much higher among those with short- than with long-term stays (aOR:2.22 (1.99–2.48) and aOR:1.15 (0.96–1.39), respectively). While some variations were observed between patients with short- vs. long-term stays among aggressive care indicators, this difference was mostly driven by in-hospital deaths, which was higher among those with short- than long-term stays (aOR: 2.58 (2.35–2.84), and aOR:1.40 (1.17–1.67), respectively).
Table 2.
Stratified sensitivity analysis by short- and long-term care patients.
| Short-Term Care | Long-Term Care | ||||
|---|---|---|---|---|---|
| EOL Indicator | Impairment Level vs Cognitively Intact | Univariable Model OR (95% CI) |
Multivariable Model aOR (95% CI) |
Univariable Model OR (95% CI) |
Multivariable Model aOR (95% CI) |
| Any EOL | Mild | 1.08 (1.02–1.14) | 1.10 (1.04–1.16) | 0.86 (0.77–0.98) | 0.83 (0.73–0.94) |
| Moderate | 1.12 (1.05–1.19) | 1.17 (1.09–1.24) | 0.86 (0.76–0.97) | 0.85 (0.75–0.96) | |
| Severe | 2.09 (1.88–2.33) | 2.22 (1.99–2.48) | 1.15 (0.97–1.38) | 1.15 (0.96–1.39) | |
| Any Cancer-Directed Treatment | Mild | 0.86 (0.81–0.92) | 0.88 (0.82–0.94) | 0.73 (0.60–0.90) | 0.73 (0.59–0.89) |
| Moderate | 0.67 (0.62–0.72) | 0.70 (0.65–0.76) | 0.68 (0.56–0.82) | 0.71 (0.58–0.87) | |
| Severe | 0.80 (0.71–0.89) | 0.82 (0.73–0.93) | 0.51 (0.36–0.70) | 0.53 (0.38–0.74) | |
| > 1 Hospital Admission | Mild | 1.11 (1.04–1.2) | 1.13 (1.05–1.22) | 0.89 (0.71–1.10) | 0.87 (0.7–1.09) |
| Moderate | 1.02 (0.94–1.11) | 1.05 (0.96–1.14) | 0.84 (0.68–1.04) | 0.87 (0.69–1.08) | |
| Severe | 1.14 (1.01–1.29) | 1.17 (1.03–1.32) | 1.11 (0.83–1.48) | 1.15 (0.85–1.55) | |
| >1 ED Visit | Mild | 1.17 (1.10–1.24) | 1.18 (1.11–1.26) | 0.93 (0.79–1.10) | 0.91 (0.77–1.08) |
| Moderate | 1.11 (1.03–1.19) | 1.13 (1.06–1.21) | 0.83 (0.7–0.98) | 0.85 (0.71–1.00) | |
| Severe | 1.43 (1.29–1.58) | 1.47 (1.33–1.63) | 1.15 (0.91–1.43) | 1.17 (0.93–1.48) | |
| Any ICU Admission | Mild | 0.88 (0.82–0.94) | 0.88 (0.82–0.94) | 0.92 (0.79–1.08) | 0.89 (0.75–1.04) |
| Moderate | 0.84 (0.79–0.91) | 0.85 (0.79–0.91) | 0.83 (0.71–0.97) | 0.81 (0.69–0.95) | |
| Severe | 1.03 (0.93–1.15) | 1.03 (0.92–1.15) | 0.99 (0.79–1.23) | 0.96 (0.76–1.21) | |
| Hospice Entry in Last 3 Days of Life | Mild | 0.91 (0.84–0.98) | 0.94 (0.87–1.01) | 0.77 (0.63–0.94) | 0.77 (0.62–0.94) |
| Moderate | 0.82 (0.75–0.9) | 0.87 (0.8–0.95) | 0.70 (0.57–0.85) | 0.74 (0.60–0.91) | |
| Severe | 0.68 (0.59–0.79) | 0.73 (0.63–0.84) | 0.59 (0.43–0.81) | 0.65 (0.47–0.89) | |
| In-Hospital Death | Mild | 1.17 (1.11–1.24) | 1.17 (1.11–1.24) | 0.99 (0.87–1.12) | 0.95 (0.84–1.08) |
| Moderate | 1.35 (1.27–1.43) | 1.35 (1.27–1.43) | 0.99 (0.88–1.12) | 0.95 (0.84–1.08) | |
| Severe | 2.56 (2.34–2.81) | 2.58 (2.35–2.84) | 1.46 (1.23–1.73) | 1.40 (1.17–1.67) | |
Discussion:
To our knowledge, this is the first study to examine EOL care patterns in nursing facility residents dying with metastatic cancer across gradients of COG-I.
Contrary to our hypothesis, our findings showed that those with severe COG-I had the highest odds of receiving any aggressive EOL care—largely because they experienced in-hospital death. Our findings of severe COG-I patients having the highest odds of >1 ED visit and >1 hospitalization in the last 30 days of life support findings of others who have noted an increasing trend of ED, hospital (and ICU) utilization of patients with NFS with poor cognitive status.18 In particular, we note the marked differences in the use of aggressive EOL care between patients with short- and long-term NFS, driven mostly by in-hospital deaths. Financial (e.g. incentives for nursing facilities to send a patient to the hospital) and system issues (e.g. DNR orders not accompanying patients sent to EDs) have been hypothesized to contribute to the increased use of ED and hospital services for patients with NFS who suffer severe COG-I.18–20 In order to more fully elucidate these findings, a mixed methods study is needed to gain a better understanding of the circumstances surrounding these decisions for ED use and/or to hospitalize. Aggressive EOL care imposes burden without providing trade-offs in health recovery and quality of life and yields even more diminished return in patients with NFS with high multimorbidity and severe COG-I.19,20 Our definition of “aggressive EOL care” is based on established measures that are available in administrative databases and are largely driven by resource use. We acknowledge, however, that EOL care needs vary greatly by the patient’s or family’s perceived needs for specific types or settings of care based on their cultural, religious, or spiritual beliefs – among others – and that EOL care decisions should be based on care goal alignment between the patient and their family on the one hand, and the care team on the other.
Our study has several strengths, including a large study population, and a sizable subgroup of patients with severe COG-I, yielding robust estimates, even as we examined multiple outcomes. We also note the following limitations: First, given our study population (patients with NFS dying with metastatic cancer and with complete information to derive COG-I status), study findings may not be generalizable to their community-dwelling counterparts, especially given the structural and organizational factors contributing to decision making regarding EOL care.21 Second, we were unable to account for individual and/or family preferences regarding EOL care or advance directives, or any relevant communications between the family and the nursing facilities. Third, a dose-response relationship was observed between COG-I and study outcomes only when analyzing late entry into hospice and in-hospital death. It appears that the CFS may be more adequate in capturing the associations of interest in the extremes of the cognitive scale than along a gradient, as intended in this study.
In conclusion, further research is needed to elucidate factors related to EOL treatment decisions for both short- and long-term NFS patients.
Supplementary Material
Key Points.
Among older metastatic cancer patients with nursing facility stays (NFS), those with severe COG-I had the highest odds of receiving aggressive EOL care--compared to their cognitively intact counterparts—largely because they experienced higher rates of in-hospital death. However, we observed marked differences in the odds of receiving aggressive EOL care between patients with short- and long-term NFS, and this difference was mostly driven by in-hospital deaths, which was higher among those with short- than long-term stays.
We were unable to account for individual’s or family’s EOL treatment preferences nor whether there was any evidence of advance directives or EOL care discussions. Such information is needed in order to fully understand why those least able to advocate for their wishes (severe COG-I) had the highest receipt of aggressive EOL care.
Why does this paper matter?
These findings support the need for further investigations of older patients with cancer and nursing facility stays to describe the individual, family, and system level factors associated with aggressive EOL care. This information is needed to place these findings in a fuller context which will lead to the development and testing of interventions aimed at reducing unwanted aggressive EOL care.
Acknowledgments:
The funder had no role in the conceptualization or design of the study; acquisition, analysis or interpretation of the data; or in the drafting or revision of the manuscript.
Funding source:
National Cancer Institute, Supplement to P30 CA043703
Footnotes
Conflicts of interest: None
Supplementary Materials for Results from Additional Analyses
References:
- 1.Smith BD, Smith GL, Hurria A, Hortobagyi GN, Buchholz TA. Future of cancer incidence in the United States: burdens upon an aging, changing nation. J Clin Oncol. 2009;27(17):2758–2765. [DOI] [PubMed] [Google Scholar]
- 2.Collaborators GBDDF. Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019. Lancet Public Health. 2022;7(2):e105–e125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Givens JL, Selby K, Goldfeld KS, Mitchell SL. Hospital transfers of nursing home residents with advanced dementia. J Am Geriatr Soc. 2012;60(5):905–909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mitchell SL, Teno JM, Kiely DK, et al. The clinical course of advanced dementia. N Engl J Med. 2009;361(16):1529–1538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Koroukian SM, Schiltz NK, Warner DF, Stange KC, Smyth KA. Increasing Burden of Complex Multimorbidity Across Gradients of Cognitive Impairment. Am J Alzheimers Dis Other Demen. 2017:1533317517726388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bradley CJ, Clement JP, Lin C. Absence of cancer diagnosis and treatment in elderly Medicaid-insured nursing home residents. J Natl Cancer Inst. 2008;100(1):21–31. [DOI] [PubMed] [Google Scholar]
- 7.Buchanan RJ, Barkley J, Wang S, Kim M. Analyses of nursing home residents with cancer at admission. Cancer Nurs. 2005;28(5):406–414. [DOI] [PubMed] [Google Scholar]
- 8.Heck JE, Albert SM, Franco R, Gorin SS. Patterns of dementia diagnosis in surveillance, epidemiology, and end results breast cancer survivors who use chemotherapy. J Am Geriatr Soc. 2008;56(9):1687–1692. [DOI] [PubMed] [Google Scholar]
- 9.Gorin SS, Heck JE, Albert S, Hershman D. Treatment for breast cancer in patients with Alzheimer’s disease. J Am Geriatr Soc. 2005;53(11):1897–1904. [DOI] [PubMed] [Google Scholar]
- 10.Harrison KL, Ritchie CS, Patel K, et al. Care Settings and Clinical Characteristics of Older Adults with Moderately Severe Dementia. J Am Geriatr Soc. 2019;67(9):1907–1912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Thomas KS, Boyd E, Mariotto AB, Penn DC, Barrett MJ, Warren JL. New Opportunities for Cancer Health Services Research: Linking the SEER-Medicare Data to the Nursing Home Minimum Data Set. Med Care. 2018;56(12):e90–e96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project. Elixhauser Comorbidity Software Refined for ICD-10-CM. https://www.hcup-us.ahrq.gov/toolssoftware/comorbidityicd10/comorbidity_icd10.jsp.
- 13.Thomas KS, Dosa D, Wysocki A, Mor V. The Minimum Data Set 3.0 Cognitive Function Scale. Med Care. 2017;55(9):e68–e72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Taylor K, Diaz A, Nuliyalu U, Ibrahim A, Nathan H. Association of Dual Medicare and Medicaid Eligibility With Outcomes and Spending for Cancer Surgery in High-Quality Hospitals. JAMA Surg. 2022;157(4):e217586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gozalo PL, Miller SC, Intrator O, Barber JP, Mor V. Hospice effect on government expenditures among nursing home residents. Health Serv Res. 2008;43(1 Pt 1):134–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Miller SC, Intrator O, Gozalo P, Roy J, Barber J, Mor V. Government expenditures at the end of life for short- and long-stay nursing home residents: differences by hospice enrollment status. J Am Geriatr Soc. 2004;52(8):1284–1292. [DOI] [PubMed] [Google Scholar]
- 17.RTI International. MDS 3.0 Quality Measures User’s Manual. Effective April 1, 2016. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-instruments/NursingHomeQualityInits/Downloads/MDS-30-QM-Users-Manual-V10.pdf. Accessed Accessed March 21, 2023.
- 18.Fulton AT, Gozalo P, Mitchell SL, Mor V, Teno JM. Intensive care utilization among nursing home residents with advanced cognitive and severe functional impairment. J Palliat Med. 2014;17(3):313–317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Reinhardt JP. What Predicts Hospice Use in the Nursing Home? J Soc Work End Life Palliat Care. 2020;16(3):286–293. [DOI] [PubMed] [Google Scholar]
- 20.Unroe KT, Caterino JM, Stump TE, et al. Long-Stay Nursing Facility Resident Transfers: Who Gets Admitted to the Hospital? J Am Geriatr Soc. 2020;68(9):2082–2089. [DOI] [PubMed] [Google Scholar]
- 21.Sutherland N, St Amant O, Dupuis S, Kontos P, Wiersma E, Brennan M. Just and inclusive end-of-life decision-making for long-term care home residents with dementia: a qualitative study protocol. BMC Palliat Care. 2022;21(1):202. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.