Background
The intersection of language barriers and end-of-life care represents a critical yet understudied dimension of healthcare equity. In the United States, individuals with preferred language other than English (PLOE) face well-documented disparities in healthcare access, quality, and outcomes across numerous clinical domains. For patients with dementia, whose cognitive impairment progressively diminishes their capacity for independent decision-making, effective communication becomes not merely advantageous but essential for ensuring care aligns with patient values and preferences.
Dementia, affecting an estimated 6.7 million Americans aged 65 and older, represents one of the leading causes of death in the United States. The progressive nature of cognitive decline means that by the disease’s advanced stages, most patients require surrogate decision-makers and rely heavily on healthcare providers to facilitate meaningful discussions about care goals, treatment preferences, and end-of-life wishes. When language barriers exist in these conversations, the risk of miscommunication, misunderstanding, and ultimately goal-discordant care increases substantially.
Despite the clinical importance of this issue, limited empirical data exist characterizing how language preference influences the quality and intensity of end-of-life care received by patients with dementia. The study by Pollack and colleagues addresses this gap by examining a large retrospective cohort of decedents with dementia, comparing end-of-life care characteristics between those with PLOE and those preferring English.
Study Design and Methods
The investigators conducted a retrospective cohort study using data from the University of Washington Medicine health system, spanning the period from 2011 to 2021. The study population comprised 7,777 decedents with a documented diagnosis of dementia. Participants were stratified into two groups based on language preference: those with preferred language other than English (PLOE, n=725, 9%) and those preferring English (n=7,052, 91%).
The primary exposure of interest was language preference, while the study examined multiple outcome measures related to end-of-life care quality and healthcare utilization. These included documentation of a goals-of-care discussion occurring within the six months preceding death, presence of advance care planning (ACP) documents in the medical record, hospital-based healthcare utilization during the last 30 days of life (emergency department visits, hospital admissions, and intensive care unit admissions), palliative care consultation, and in-hospital death.
The research team employed multivariable regression models adjusted for potential confounders including age at death, sex, race/ethnicity, education level, marital status, comorbidity burden, dementia type, and duration of dementia diagnosis. This comprehensive adjustment strategy strengthens the validity of observed associations by controlling for demographic and clinical factors that might independently influence end-of-life care patterns.
Key Findings
Demographic Characteristics
The median age at death was 84 years across the cohort, with approximately half of patients female. Notably, the proportion of patients with PLOE (9%) is consistent with demographic trends in Washington State, which has experienced substantial immigration and linguistic diversification over recent decades.
Goals-of-Care Discussions
Perhaps unexpectedly, patients with PLOE were more likely to have documented goals-of-care discussions compared to those preferring English. Among the PLOE group, 62% had such documentation in the six months preceding death, compared to only 45% of the English-preferring group. The adjusted risk difference was 0.08 (95% CI: 0.04-0.13), indicating a statistically significant 8 percentage point increase in documentation frequency among patients with language preferences other than English.
This finding suggests that healthcare providers may recognize language barriers as requiring additional communication efforts, potentially triggering more frequent and earlier goals-of-care conversations when working with interpreters or bilingual staff. Alternatively, this could reflect family or caregiver advocacy driving increased documentation requirements for patients whose care decisions involve more complex communication pathways.
Healthcare Utilization in the Final Month of Life
Despite comparable rates of documented goals-of-care discussions, patients with PLOE experienced significantly higher hospital-based healthcare utilization during the last 30 days of life. The adjusted analyses revealed substantial differences across multiple utilization metrics:
Emergency Department Visits: Patients with PLOE had a 9 percentage point higher likelihood of ED visits in the final month (adjusted risk difference 0.09; 95% CI: 0.04-0.14).
Hospital Admissions: The adjusted risk difference for hospital admission was 0.06 (95% CI: 0.01-0.11), representing a 6 percentage point increase among PLOE patients.
Intensive Care Unit Admissions: ICU admissions were 6 percentage points higher in the PLOE group (adjusted risk difference 0.06; 95% CI: 0.02-0.09).
In-Hospital Death: Perhaps most striking, patients with PLOE were 11 percentage points more likely to die in the hospital (adjusted risk difference 0.11; 95% CI: 0.07-0.14).
These findings paint a consistent picture: despite more frequent documented goals-of-care discussions, patients with language barriers experienced substantially more intensive medical intervention at the end of life, culminating in a significantly higher likelihood of dying in an acute care setting rather than in a hospice, home, or other community-based setting.
Advance Care Planning and Palliative Care Consultation
Interestingly, the study found no significant differences between groups in either the frequency of advance care planning documentation or palliative care consultation rates. This null finding is noteworthy because it suggests that while documentation of conversations may differ by language preference, the formal structures intended to translate preferences into care plans did not vary systematically between groups.
Interpretation and Implications
The paradoxical finding of more frequent goals-of-care documentation alongside higher hospital utilization raises important questions about the quality and effectiveness of communication for patients with language barriers. Several potential explanations merit consideration.
First, documented conversations may not always translate to effective mutual understanding. Even with interpreter assistance, the complexity of discussing prognosis, treatment trade-offs, and end-of-life preferences requires nuanced communication that may be compromised by language barriers. A conversation that appears documented may not achieve the shared understanding necessary to genuinely align care with patient values.
Second, cultural factors beyond language may influence preferences regarding aggressive medical intervention. Families from certain cultural backgrounds may strongly prefer comprehensive treatment efforts, or conversely, may feel uncomfortable expressing wishes that seem to suggest acceptance of death. Without culturally informed communication strategies, documented discussions may fail to capture authentic patient or family preferences.
Third, the ED and hospital may serve critical safety-net functions for patients with language barriers who face challenges accessing outpatient care, home health services, or community-based hospice programs. When crises occur and communication barriers exist, emergency and hospital-based care may represent the most accessible and immediate care options, regardless of previously documented preferences.
Fourth, surrogate decision-makers for patients with PLOE may face compounded challenges: navigating their loved one’s dementia-related communication impairments while simultaneously managing language barriers with healthcare providers. These layered challenges could lead to default decisions favoring aggressive treatment, particularly when clear guidance from the patient is unavailable.
Expert Commentary and Study Limitations
While this study provides valuable insights into language-related disparities in dementia end-of-life care, several limitations should be acknowledged. The single-health-system design limits generalizability to other geographic regions or healthcare contexts with different patient populations or practice patterns. The reliance on documented language preference may not capture the full complexity of actual language proficiency or interpreter needs during clinical encounters.
The retrospective design precludes causal inference about the relationship between language preference and care outcomes. Observed associations could reflect unmeasured confounders such as family structure, social support, health literacy, or access to community-based services that correlate with both language preference and care-seeking patterns.
The study’s focus on documentation as a proxy for communication quality has inherent limitations. Documentation may reflect billing requirements, medicolegal considerations, or variable clinician documentation practices rather than genuine communication quality or shared decision-making processes.
Despite these limitations, the study’s strengths—including its large sample size, rigorous adjustment for confounders, and examination of multiple outcome dimensions—make a meaningful contribution to understanding how language barriers influence the dying experience for patients with dementia.
Conclusion
This retrospective cohort study reveals a troubling paradox in end-of-life care for patients with dementia and language preferences other than English: despite more frequent documentation of goals-of-care discussions, these patients experience substantially higher hospital-based healthcare utilization in their final month of life, including more ED visits, hospital and ICU admissions, and in-hospital deaths.
These findings suggest that simply increasing the frequency of documented conversations is insufficient to improve end-of-life care for patients with language barriers. Rather, efforts must focus on enhancing the quality, cultural appropriateness, and actual comprehension achieved in these communications. Healthcare systems should invest in trained medical interpreters, culturally sensitive communication training for providers, and systematic approaches to ensure that documented preferences genuinely reflect informed, shared decisions.
Furthermore, recognizing the safety-net function of emergency and hospital-based care for populations facing access barriers, interventions should target these settings specifically. ED-based palliative care integration, multilingual care navigation programs, and community-health system partnerships may help bridge gaps for patients with PLOE and their families.
Future research should investigate whether the higher hospital utilization observed in this study represents goal-concordant or goal-discordant care, examine mechanisms underlying the gap between documented discussions and care outcomes, and evaluate interventions designed to improve end-of-life communication quality for patients with language barriers.
As the United States continues to diversify linguistically, addressing these disparities becomes increasingly urgent. Ensuring that all patients with dementia—regardless of language—can achieve peaceful, dignified deaths aligned with their values and preferences represents both a clinical imperative and a moral obligation.
Funding and Trial Registration
Study data derived from the University of Washington Medicine health system electronic medical record. Specific funding sources were not detailed in the available abstract. No clinical trial registration applicable to this retrospective cohort study.
References
1. Pollack LR, Downey L, Engelberg RA, Sibley J, Ko LK, Domoto-Reilly K, Brumback LC, Chen AT, Sharma RK. Language Preference Is Associated with Goals-of-Care Communication and End-of-Life Care in Dementia. Journal of General Internal Medicine. 2026. PMID: 41912943.
