Psychosocial distress among spouses of persons with dementia before and after their partner's death.

BACKGROUND: Spouses of persons living with dementia may face heightened psychosocial distress in the years immediately before and after their partner's death. We compared the psychosocial needs of spouses of partners with dementia with spouses of partners with non-impaired cognition nearing and after the end of life, focusing on loneliness, depression, life satisfaction, and social isolation. METHODS: We used nationally representative Health and Retirement Study married couples data (2006-2018), restricting to spouses 50+ years old. We included 2098 spouses with data on loneliness and depressive symptoms 2 years before and after the partner's death. We additionally examined a subset of spouses (N = 1113) with available data on life satisfaction and social isolation 2 years before their partner's death. Cognitive status of partners was classified as non-impaired cognition, cognitive impairment not dementia (cognitive impairment), and dementia. We used multivariable logistic regression to determine: 1) the change in loneliness and depression for spouses pre- and post-partner's death, and 2) life satisfaction and social isolation 2 years before the partner's death. RESULTS: Spouses were on average 73 years old (SD: 10), 66% women, 7% Black, 7% Hispanic non-White, 24% married to persons with cognitive impairment, and 19% married to partners with dementia. Before their partner's death, spouses married to partners with dementia experienced more loneliness (non-impaired cognition: 8%, cognitive impairment: 16%, dementia: 21%, p-value = 0.002) and depressive symptoms (non-impaired cognition: 20%, cognitive impairment: 27%, dementia: 31%, p-value < 0.001), and after death a similar prevalence of loneliness and depression across cognitive status. Before their partner's death, spouses of partners with dementia reported less life satisfaction (non-impaired cognition: 74%, cognitive impairment: 68%; dementia: 64%, p-value = 0.02) but were not more socially isolated. CONCLUSION: Results emphasize a need for clinical and policy approaches to expand support for the psychosocial needs of spouses of partners with dementia in the years before their partner's death rather than only bereavement.

Variation in Hospice Aide Care by Residential Setting.

Background: Hospice care frequently includes hands-on care from hospice aides, but the need for hospice aide care may vary in residential settings (e.g., assisted livings and nursing homes). Objectives: The objective of this study is to compare hospice aide use and factors associated with use across residential settings. Design: This longitudinal cohort study used data from Medicare beneficiaries in the United States enrolled in the Medicare Current Beneficiary Survey (MCBS) who died between 2010 and 2019 and had hospice claims and available residential setting data in MCBS (n = 1,915). Analysis: Decedent hospice aide use was compared by residential settings; multivariable models controlling for sociodemographic, clinical/functional, and hospice characteristics examined factors associated with hospice aide care in different residential settings. Results: Hospice aide visits were least common in the community setting (64.4% vs. 76.6% vs. 72.6% with any hospice aide visits in community, assisted living, and nursing home, respectively, p = 0.001). In adjusted models, factors associated with hospice aide visits did not significantly differ by residential settings. Conclusions: Despite staff providing hands-on support in assisted livings and nursing homes, hospice aide visits were more common in residential as opposed to community settings, and factors associated with hospice aide visits were similar among settings. To maximize the potentially positive impact of hospice aides on overall care, additional work is needed to understand when hospice aides are used and how hospice aides collaborate with families and care teams. This will help to ensure that hospice care is appropriately tailored to individual care needs in all residential settings.

Individual Socioeconomic Factors Have a Greater Impact on End-of-Life Care Outcomes Than Regional Factors.

Background: End-of-life (EoL) care provided to Americans in urban and rural settings is distinct in terms of both available and delivered services. However, much less is known about which geographic, demographic, and health indicators are associated with disparities in EoL care and how individual versus regional characteristics influence quality of care (QoC). Objective: This study aimed to assess how regionality, rurality, and individual socioeconomic factors are associated with QoC in the last month of life (LML). Design: Nationally representative cross-sectional study using the proxy-completed LML questionnaire as part of the National Health and Aging Trends Study (NHATS). The data were linked at the zip code level to geographic and economic indicators. Settings/Subjects: A total of 2778 NHATS enrollees who died from 2012 to 2020. Measurements: Measurements included population density, socioeconomic indicators, health factors, and health outcomes. The primary independent variable was proxy-reported QoC during the LML (excellent vs. not excellent). Results: In our sample, 52.1% (n = 1447) reported not excellent care and 47.9% (n = 1331) reported excellent care. These populations varied in their demographic and socioeconomic characteristics. After accounting for survey weighting and design, decedents in the top (odds ratio [OR]: 1.58; 95% confidence interval [CI]: 1.08-2.32) income quartile had significantly greater odds of receiving excellent care than decedents in the bottom quartile. Decedents in zip codes with top quartile health outcome metrics had significantly greater odds of receiving excellent care (OR: 1.64; 95% CI: 1.17-2.29) than decedents in zip codes with bottom quartile health outcomes. County rurality index and county health factors were not correlated with QoC in the LML. Conclusions: High QoC at the EoL may be more associated with individual socioeconomic factors than regional indicators, including degrees of rurality. Clinicians should strive to recognize the interplay of individual characteristics and regional indicators to provide more personalized care.

Perspectives on Transfusions for Hospice Patients With Blood Cancers: A Survey of Hospice Providers.

CONTEXT: Patients with blood cancers have low rates of hospice use. While lack of transfusion access in hospice is posited to substantially contribute to these low rates, little is known about the perspectives of hospice providers regarding transfusion access in hospice. OBJECTIVES: To characterize hospice providers' perspectives regarding care for patients with blood cancers and transfusions in the hospice setting. METHODS: In 2022, we conducted a cross-sectional survey of a sample of hospices in the United States regarding their experience caring for patients with blood cancers, perceived barriers to hospice use, and interventions to increase enrollment. RESULTS: We received 113 completed surveys (response rate = 23.5%). Of the cohort, 2.7% reported that their agency always offers transfusions, 40.7% reported sometimes offering transfusions, and 54.9% reported never offering transfusions. In multivariable analyses, factors associated with offering transfusions included nonprofit ownership (OR 5.93, 95% CI, 2.2-15.2) and daily census >50 patients (OR 3.06, 95% CI, 1.19-7.87). Most respondents (76.6%) identified lack of transfusion access in hospice as a barrier to hospice enrollment for blood cancer patients. The top intervention considered as "very helpful" for increasing enrollment was additional reimbursement for transfusions (72.1%). CONCLUSION: In this national sample of hospices, access to palliative transfusions was severely limited and was considered a significant barrier to hospice use for blood cancer patients. Moreover, hospices felt increased reimbursement for transfusions would be an important intervention. These data suggest that hospice providers are supportive of increasing transfusion access and highlight the critical need for innovative hospice payment models to improve end-of-life care for patients with blood cancers.

Health Care Costs Associated With Hospice Use For People With Dementia In The US.

Policy makers in the US are increasingly concerned that greater use of the Medicare hospice benefit by people with dementia is driving up costs. Yet this perspective fails to incorporate potential cost savings associated with hospice. We estimated the association between hospice use by people with dementia and health care costs, using Medicare Current Beneficiary Survey data from the period 2002-19. For community-dwelling people with dementia, Medicare costs were lower for those who used hospice than for those who did not, whether hospice enrollment was in the last three days ($2,200) or last three months ($7,200) of life, primarily through lower inpatient care costs in the last days of life. In nursing homes, total and Medicare costs were lower for hospice users with dementia who enrolled within a month of death than for those who did not use hospice. Total costs for the entire last year of life for those who used any days of hospice in the last year compared with no hospice did not differ, although Medicare costs were higher and Medicaid costs lower for those in nursing homes. Medicare policies that reduce hospice access and incentivize hospice disenrollment may actually increase Medicare costs, given that hospice cost savings generally derive from a person's last days or weeks of life.

Hospice Quality, Race, and Disenrollment in Hospice Enrollees With Dementia.

Background: Racial and ethnic minoritized people with dementia (PWD) are at high risk of disenrollment from hospice, yet little is known about the relationship between hospice quality and racial disparities in disenrollment among PWD. Objective: To assess the association between race and disenrollment between and within hospice quality categories in PWD. Design/Setting/Subjects: Retrospective cohort study of 100% Medicare beneficiaries 65+ enrolled in hospice with a principal diagnosis of dementia, July 2012-December 2017. Race and ethnicity (White/Black/Hispanic/Asian and Pacific Islander [AAPI]) was assessed with the Research Triangle Institute (RTI) algorithm. Hospice quality was assessed with the publicly-available Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey item on overall hospice rating, including a category for hospices exempt from public reporting (unrated). Results: The sample included 673,102 PWD (mean age 86, 66% female, 85% White, 7.3% Black, 6.3% Hispanic, 1.6% AAPI) enrolled in 4371 hospices nationwide. Likelihood of disenrollment was higher in hospices in the lowest quartile of quality ratings (vs. highest quartile) for both White (adjusted odds ratio [AOR] 1.12 [95% confidence interval 1.06-1.19]) and minoritized PWD (AOR range 1.2-1.3) and was substantially higher in unrated hospices (AOR range 1.8-2.0). Within both low- and high-quality hospices, minoritized PWD were more likely to be disenrolled compared with White PWD (AOR range 1.18-1.45). Conclusions: Hospice quality predicts disenrollment, but does not fully explain disparities in disenrollment for minoritized PWD. Efforts to improve racial equity in hospice should focus both on increasing equity in access to high-quality hospices and improving care for racial minoritized PWD in all hospices.

Examining annual transitions in healthcare spending among U.S. medicare beneficiaries using multistate Markov models: Analysis of medicare current beneficiary survey data, 2003-2019.

Many studies have examined factors associated with individuals of high or low healthcare spending in a given year. However, few have studied how healthcare spending changes over multiple years and which factors are associated with the changes. In this study, we examined the dynamic patterns of healthcare spending over a three-year period, among a nationally representative cohort of Medicare beneficiaries in the U.S. and identified factors associated with these patterns. We extracted data for 30,729 participants from the national Medicare Current Beneficiary Survey (MCBS), for the period 2003-2019. Using multistate Markov (MSM) models, we estimated the probabilities of year-to-year transitions in healthcare spending categorized as three states (low (L), medium (M) and high (H)), or to the terminal state, death. The participants, 13,554 (44.1%), 13,715 (44.6%) and 3,460 (11.3%) were in the low, medium and high spending states at baseline, respectively. The majority of participants remained in the same spending category from one year to the next (L-to-L: 76.8%; M-to-M: 71.7%; H-to-H: 56.6 %). Transitions from the low to high spending state were significantly associated with older age (75-84, ≥85 years), residing in a long-term care facility, greater assistance with activities of daily living, enrollment in fee-for-service Medicare, not receiving a flu shot, and presence of specific medical conditions, including cancer, dementia, and heart disease. Using data from a large population-based longitudinal survey, we have demonstrated that MSM modelling is a flexible framework and useful tool for examining changes in healthcare spending over time.

End-of-Life Experiences Among "Kinless" Older Adults: A Nationwide Register-Based Study.

Background: The population of older adults who are unpartnered and childless (i.e., "kinless") is increasing across the globe, and may be at risk for lower quality end-of-life (EoL) experiences due to lack of family support, assistance, and advocacy. Yet, little research exists on the EoL experiences of "kinless" older adults. Objectives: To document associations between family structure (i.e., presence or absence of partner or child) and intensity of EoL experiences (i.e., visits to medicalized settings before death). Design: The study design is a cross-sectional population-based register study of the population of Denmark. Subjects: Participants include all adults age 60 years and older who died of natural causes in Denmark from 2009 to 2016 (n = 137,599 decedents). Results: "Kinless" older adults (reference = has partner, has child) were the least likely group to visit the hospital (two or more times; odds ratio [OR] = 0.74, confidence interval [CI] = 0.70-0.77), emergency department (one or more times; OR = 0.90, CI = 0.86-0.93), and intensive care unit (one or more times; OR = 0.71, CI = 0.67-0.75) before death. Conclusions: "Kinless" older adults in Denmark were less likely to experience medically intensive care at the EoL. Further research is needed to understand factors associated with this pattern to ensure that all individuals receive high quality EoL care regardless of their family structure and family tie availability.

Perceptions of Palliative Care Program Viability During the Pandemic: Qualitative Results From a National Survey.

Background: Palliative care programs have played a significant role during the COVID-19 pandemic. However, the financial impact of the pandemic and operational challenges for palliative care programs have raised concerns for their future viability. Objectives: To explore palliative care program leaders' perceptions of the future viability of their programs in the context of the pandemic and inform future educational and program development. Methods: Surveys were sent to 1430 specialist palliative care program leaders, identified through the Center to Advance Palliative Care's contact lists, via email in May 2020 and January 2021. Leaders were asked why they were or were not concerned about the viability of their palliative care programs. Qualitative content analysis was applied to determine themes. Results: We received 440 responses. Most programs served hospital settings and were geographically located across all US regions. We identified four themes: 1) The importance of being valued by organizational leadership and peers, 2) The importance of adequate and supported palliative care staff, 3) The pandemic validated and accelerated the need for palliative care, and 4) The pandemic perpetuated organizational financial concerns. Conclusion: Findings provide insights about palliative care program viability from the perspective of program leaders during a global pandemic. Technical assistance to support palliative care teams and their relationships with stakeholders, methods to measure the impact of peer support, efforts to educate administrators about the value of palliative care, and efforts to reduce burnout are needed to sustain palliative care programs into the future.

Care Disruptions and End-Of-Life Care Experiences Among Home-Based Primary Care Patients During the COVID-19 Pandemic in New York City: A Retrospective Chart Review.

Background: Research on deaths during COVID-19 has largely focused on hospitals and nursing homes. Less is known about medically complex patients receiving care in the community. We examined care disruptions and end-of-life experiences of homebound patients receiving home-based primary care (HBPC) in New York City during the initial 2020 COVID-19 surge. Methods: We conducted a retrospective chart review of patients enrolled in Mount Sinai Visiting Doctors who died between March 1-June 30, 2020. We collected patient sociodemographic and clinical data and analyzed care disruptions and end-of-life experiences using clinical notes, informed by thematic and narrative analysis. Results: Among 1300 homebound patients, 112 (9%) died during the study period. Patients who died were more likely to be older, non-Hispanic white, and have dementia than those who survived. Thirty percent of decedents had confirmed or probable COVID-19. Fifty-eight (52%) were referred to hospice and 50 enrolled. Seventy-three percent died at home. We identified multiple intersecting disruptions in family caregiving, paid caregiving, medical supplies and services, and hospice care, as well as hospital avoidance, complicating EOL experiences. The HBPC team responded by providing clinical, logistical and emotional support to patients and families. Conclusion: Despite substantial care disruptions, the majority of patients in our study died at home with support from their HBPC team as the practice worked to manage care disruptions. Our findings suggest HBPC's multi-disciplinary, team-based model may be uniquely suited to meet the needs of the most medically and socially vulnerable older adults at end of life during public health emergencies.

The devil's in the details: Variation in estimates of late-life activity limitations across national cohort studies.

BACKGROUND: Assessing activity limitations is central to aging research. However, assessments of activity limitations vary, and this may have implications for the populations identified. We aim to compare measures of activities of daily living (ADLs) and their resulting prevalence and mortality across three nationally-representative cohort studies: the National Health and Aging Trends Study (NHATS), the Health and Retirement Survey (HRS), and the Medicare Current Beneficiary Survey (MCBS). METHODS: We compared the phrasing and context of questions around help and difficulty with six self-care activities: eating, bathing, toileting, dressing, walking inside, and transferring. We then compared the prevalence and 1-year mortality for difficulty and help with eating and dressing. RESULTS: NHATS, HRS, and MCBS varied widely in phrasing and framing of questions around activity limitations, impacting the proportion of the population found to experience difficulty or receive help. For example, in NHATS 12.4% [95% confidence interval (CI) 11.5%-13.4%] of the cohort received help with dressing, while in HRS this figure was 6.4% [95% CI 5.7%-7.2%] and MCBS 5.3% [95% CI 4.7%-5.8%]. When combined with variation in sampling frame and survey approach of each survey, such differences resulted in large variation in estimates of the older population of older adults with ADL disability. CONCLUSIONS: In order to take late-life activity limitations seriously, we must clearly define the measures we use. Further, researchers and clinicians seeking to understand the experience of older adults with activity limitations should be careful to interpret findings in light of the framing of the question asked.

Functional and clinical needs of older hospice enrollees with coexisting dementia.

BACKGROUND: The Medicare Hospice Benefit increasingly serves people dying with dementia. We sought to understand characteristics, hospice use patterns, and last-month-of-life care quality ratings among hospice enrollees with dementia coexisting with another terminal illness as compared to enrollees with a principal hospice diagnosis of dementia, and enrollees with no dementia. METHODS: We conducted a pooled cross-sectional study among decedent Medicare beneficiaries age 70+ using longitudinal data from the National Health and Aging Trends Study (NHATS) (last interview before death; after-death proxy interview) linked to Medicare hospice claims (2011-2017). We used unadjusted and adjusted regression analyses to compare characteristics of hospice enrollees with coexisting dementia to two groups: (1) enrollees with a principal dementia diagnosis, and (2) enrollees with no dementia. RESULTS: Among 1105 decedent hospice enrollees age 70+, 40% had coexisting dementia, 16% had a principal diagnosis of dementia, and 44% had no dementia. In adjusted analyses, enrollees with coexisting dementia had high rates of needing help with 3-6 activities of daily living, similar to enrollees with principal dementia (62% vs. 67%). Enrollees with coexisting dementia had high clinical needs, similar to those with no dementia, for example, 63% versus 61% had bothersome pain. Care quality was worse for enrollees with coexisting dementia versus principal dementia (e.g., 61% vs. 79% had anxiety/sadness managed) and similar to those with no dementia. Enrollees with coexisting dementia had similar hospice use patterns as those with principal diagnoses and higher rates of problematic use patterns compared to those with no dementia (e.g., 16% vs. 10% live disenrollment, p = 0.004). CONCLUSIONS: People with coexisting dementia have functional needs comparable to enrollees with principal diagnoses of dementia, and clinical needs comparable to enrollees with no dementia. Changes to hospice care models and policy may be needed to ensure appropriate dementia care.

Disruptions in Home Hospice Care due to the COVID-19 Pandemic.

Background: There is limited evidence regarding the challenges of providing hospice care to those dying at home during the COVID-19 pandemic. Objective: To describe the challenges of home hospice care and the specific types of disruptions in care processes experienced by patients and families. Design: Qualitative study of the electronic medical record notes of a large New York City (NYC) home-based primary care program. Setting/Subjects: Subjects were 58 patients referred to hospice who died during the initial NYC COVID-19 surge from March to June 2020. Results: We identified six domains of disruptions in home hospice care: delayed hospice enrollment, inability to conduct home visits, lack of needed supplies, communication failures, strained caregivers, and limitations of telehealth. Conclusions: This study provides a critical first analysis of disruptions in home hospice care that can feasibly be addressed and must be prioritized by hospices throughout the ongoing pandemic and in advance of future emergencies.

Association Between Hospice Enrollment and Total Health Care Costs for Insurers and Families, 2002-2018.

Importance: Use of hospice has been demonstrated to be cost saving to the Medicare program and yet the extent to which hospice saves money across all payers, including whether it shifts costs to families, is unknown. Objective: To estimate the association between hospice use and total health care costs including family out-of-pocket health care spending. Design Setting and Participants: This retrospective cohort study of health care spending in the last 6 months of life used data from the nationally representative Medicare Current Beneficiary Survey (MCBS) between the years 2002 and 2018. Participants were MCBS participants who resided in the community and died between 2002 and 2018. Exposures: Covariate balancing propensity scores were used to compare participants who used hospice (n = 2113) and those who did not (n = 3351), stratified by duration of hospice use. Main Outcomes and Measures: Total health care expenditures were measured across payers (family out-of-pocket, Medicare, Medicare Advantage, Medicaid, private insurance, private health maintenance organizations, Veteran's Administration, and other) and by expenditure type (inpatient care, outpatient care, medical visits, skilled nursing, home health, hospice, durable medical equipment, and prescription drugs). Results: The study population included 5464 decedents (mean age 78.7 years; 48% female) and 38% enrolled with hospice. Total health care expenditures were lower for those who used hospice compared with propensity score weighted non-hospice control participants for the last 3 days of life ($2813 lower; 95% CI, $2396-$3230); last week of life ($6806 lower; 95% CI, $6261-$7350); last 2 weeks of life ($8785 lower; 95% CI, $7971-$9600); last month of life ($11 747 lower; 95% CI, $10 072-$13 422); and last 3 months of life ($10 908 lower; 95% CI, $7283-$14 533). Family out-of-pocket expenditures were lower for hospice enrollees in the last 3 days of life ($71; 95% CI, $43-$100); last week of life ($216; 95% CI, $175-$256); last 2 weeks of life ($265; 95% CI, $149-$382); and last month of life ($670; 95% CI, $530-$811) compared with those who did not use hospice. Health care savings were associated with reductions in inpatient care. Conclusions and Relevance: In this population-based cohort study of community-dwelling Medicare beneficiaries, hospice enrollment was associated with lower total health care costs for the last 3 days to 3 months of life. Importantly, we found no evidence of cost shifting from Medicare to families related to hospice enrollment. The magnitude of lower out-of-pocket spending to families who enrolled with hospice is meaningful to many Americans, particularly those with lower socioeconomic status.

Hospice Improves Care Quality For Older Adults With Dementia In Their Last Month Of Life.

The Medicare hospice benefit was originally designed around a cancer disease paradigm but increasingly serves people living with dementia. At this time, almost half of all older adults receiving hospice care have dementia. Yet there is minimal evidence as to whether hospice benefits people living with dementia outside of nursing facilities. We asked whether and how the perceived quality of last-month-of-life care differed between people with and without dementia and whether hospice use among people living with dementia was associated with perceived quality of care compared with the quality of care for those who did not use hospice. We used nationally representative data from the National Health and Aging Trends Study and Medicare claims from the period 2011-17 to examine the impact of hospice enrollment on proxy perceptions of last-month-of-life care quality. Proxies of people living with dementia enrolled in hospice compared with proxies of those not enrolled more often reported care to be excellent (predicted probability: 52 percent versus 41 percent), more often reported having anxiety or sadness managed (67 percent versus 46 percent), and less often reported changes in care settings in the last three days of life (10 percent versus 25 percent). There were no differences in the impact of hospice on proxy ratings of care for people with and without dementia. Policy makers should consider these benefits when weighing changes to hospice policy and regulations that may affect people living with dementia.

A national study of disenrollment from hospice among people with dementia.

BACKGROUND: People with dementia (PWD) are at high risk for hospice disenrollment, yet little is known about patterns of disenrollment among the growing number of hospice enrollees with dementia. DESIGN: Retrospective, observational cohort study of 100% Medicare beneficiaries with dementia aged 65 and older enrolled in the Medicare Hospice Benefit between July 2012 and December 2017. Outcome measures included hospice-initiated disenrollment for patients whose rate of decline ceased to meet the Medicare hospice eligibility guideline of "expected death within 6 months" (extended prognosis) and patient-initiated disenrollment (revocation). Hospice, regional, and patient risk factors and variation were assessed with multilevel mixed-effects logistic regression models. RESULTS: Among 867,695 hospice enrollees with dementia, 70,945 (8.2%) were disenrolled due to extended prognosis and 43,133 (5.0%) revoked within 1-year of their index admission. There was substantial variation in hospice provider disenrollment due to extended prognosis (10th-90th percentile 4.5%-14.6%, adjusted median odds ratio (MOR) 1.86, 95% confidence interval (CI) 1.82, 1.91) and revocation (10th-90th percentile 2.5%-10.1%, MOR 2.09, 95% CI 2.03, 2.14). Among hospital referral regions (HRR), there was more variation in revocation (10th-90th percentile 3.5%-7.6%, MOR 1.4, 95% CI 1.34, 1.47) than extended prognosis (10th-90th percentile 7.0%-9.5%, MOR 1.23, 95% CI 1.18, 1.27), with much higher revocation rates noted in HRRs located in the Southeast and Southern California. A number of patient and hospice characteristics were associated with higher odds of both types of disenrollment (younger age, female sex, minoritized race and ethnicity, Medicaid dual eligibility, Medicare Part C enrollment), while some were associated with revocation only (more comorbidities, newer, smaller, and for-profit hospices). CONCLUSIONS: In this nationally representative study of hospice enrollees with dementia, hospice disenrollment varied by type of hospice, geographic region, and patient characteristics including age, sex, race, and ethnicity. These findings raise important questions about whether and how the Medicare Hospice Benefit could be adapted to reduce disparities and better support PWD.

Characteristics Associated With Disparities Among Older Adults in Coronavirus Disease 2019 Outcomes in an Academic Health Care System.

BACKGROUND: An improved understanding of the coronavirus disease 2019 (COVID-19) pandemic is needed to identify predictors of outcomes among older adults with COVID-19. OBJECTIVE: The objective of this study was to examine patient and health system factors predictive of in-hospital mortality, intensive care unit (ICU) admission, and readmission among patients with COVID-19. DESIGN, SETTING, AND PARTICIPANTS: A cohort study of patients aged 18 years and older with COVID-19 discharged from 5 New York hospitals within the Mount Sinai Health System (March 1, 2020-June 30, 2020). MEASURES: Patient-level characteristics (age, sex, race/ethnicity, comorbidities/serious illness, transfer from skilled nursing facility, severe acute respiratory syndrome coronavirus 2 viral load, Sequential Organ Failure Assessment score, treatments); hospital characteristics. OUTCOMES: All-cause in-hospital mortality; ICU admission; 30-day readmission. RESULTS: Among 7556 subjects, mean age 61.1 (62.0) years; 1556 (20.6%) died, 949 (12.6%) had an ICU admission, and 227 (9.1%) had a 30-day readmission. Increased age [aged 55-64: odds ratio (OR), 3.28; 95% confidence interval (CI), 2.41-4.46; aged 65-74: OR, 4.67; 95% CI, 3.43-6.35; aged 75-84: OR, 10.73; 95% CI, 7.77-14.81; aged 85 y and older: OR, 20.57; 95% CI, 14.46-29.25] and comorbidities (OR, 1.11; 95% CI, 1.16, 2.13) were independent risk factors for in-hospital mortality. Yet older adults (aged 55-64 y: OR, 0.56; 95% CI, 0.40-0.77; aged 65-74: OR, 0.46; 95% CI, 0.33-0.65; aged 75-84: OR, 0.27; 95% CI, 0.18-0.40; aged above 85 y: OR, 0.21; 95% CI, 0.13-0.34) and those with Medicaid (OR, 0.74; 95% CI, 0.56-0.99) were less likely to be admitted to the ICU. Race/ethnicity, crowding, population density, and health system census were not associated with study outcomes. CONCLUSIONS: Increased age was the single greatest independent risk factor for mortality. Comorbidities and serious illness were independently associated with mortality. Understanding these risk factors can guide medical decision-making for older adults with COVID-19. Older adults and those admitted from a skilled nursing facility were half as likely to be admitted to the ICU. This finding requires further investigation to understand how age and treatment preferences factored into resource allocation.