Patient–Family Concordance in End-of-Life Decision-Making Preferences Among Terminally Ill Older Adults: A Cross-Sectional Study

Technical Summary: Patient–Family Concordance in End-of-Life Decision-Making Preferences Among Terminally Ill Older Adults

Problem Statement
Family members frequently act as surrogate decision-makers for terminally ill older adults, particularly in cultural contexts like China where family-centered medical decision-making is prevalent. However, a critical gap exists in understanding whether family caregivers accurately perceive and represent the specific end-of-life (EOL) preferences of the patients they care for. Limited concordance between patients and their families can lead to treatment decisions that deviate from the patient's actual wishes, potentially resulting in unwanted aggressive interventions, unmet care needs, and a reduced quality of dying. While previous research has noted discrepancies in EOL preferences, there is a lack of specific data regarding the concordance levels between terminally ill older patients and their family members in mainland China, a region with a rapidly aging population and distinct cultural dynamics regarding death and decision-making.

Methodology
This study employed a cross-sectional survey design conducted between October 2022 and November 2023 at a tertiary teaching hospital in Sichuan Province, China.

• Participants: The study recruited 505 dyads consisting of hospitalized terminally ill older patients (aged ≥60, with a life expectancy ≤12 months as determined by the "Surprise Question" approach) and their family members involved in medical decision-making. Convenience sampling was used.
• Exclusion Criteria: Patients with unstable vital signs, severe acute illness precluding participation, cognitive impairment, or diagnosed mental disorders were excluded. Family members with similar cognitive or mental health conditions were also excluded.
• Data Collection: Researchers utilized a demographic questionnaire and a structured survey adapted from previous concordance studies (e.g., Tang et al.). The survey assessed three primary domains:
  1. Preferences for Life-Sustaining Treatment: Seven specific interventions (CPR, tracheotomy, mechanical ventilation, blood transfusion, hemodialysis, nasogastric feeding, pharmacological resuscitation) were presented with binary "yes/no" response options.
  2. Preferred Place of Death: Options included hospital, home, or other.
  3. Preferred End-of-Life Decision-Maker: Options included the patient, family member, healthcare provider, or other.
• Analysis: Data were analyzed using IBM SPSS Statistics version 22.0. Concordance between patient and family responses was evaluated using Cohen's kappa statistics. Kappa values were interpreted as follows: <0.40 (poor), 0.40–0.75 (moderate), and ≥0.75 (good).

Key Results
The study revealed substantial discrepancies between patients and their family members across all measured domains, with concordance generally ranging from poor to moderate.

• Life-Sustaining Treatment Preferences:

  • Kappa coefficients ranged from 0.308 to 0.566, indicating poor to moderate agreement.
  • Highest Agreement: Tracheotomy preferences showed the highest concordance (Kappa = 0.566; agreement rate 79.4%).
  • Lowest Agreement: Intravenous blood transfusion preferences showed the lowest concordance (Kappa = 0.308; agreement rate 66.7%).
  • Pharmacological resuscitation also demonstrated poor concordance (Kappa = 0.322).
  • Notably, 43.2% of dyads agreed on preferring CPR, while 51.5% agreed on preferring blood transfusions.

• Preferred Place of Death:

  • Moderate concordance was observed (Kappa = 0.504; agreement rate 72.7%).
  • Among concordant dyads, the hospital was the most frequently preferred place of death (42.6%), followed by home (28.3%). This contrasts with some previous studies in other Asian regions where home was the primary preference.

• Preferred Decision-Maker:

  • Moderate concordance was found (Kappa = 0.468; agreement rate 71.1%).
  • A significant finding was the low preference for patient autonomy: only 22.2% of concordant dyads identified the patient as the preferred final decision-maker.
  • Conversely, 45.5% of concordant dyads identified a family member as the preferred decision-maker. Only 3.2% preferred healthcare providers.

Significance and Claims
The authors claim that this study provides critical evidence that family members do not consistently or accurately represent the EOL wishes of terminally ill older adults, particularly within family-centered cultural contexts. The findings highlight that:

1. Discrepancies are Common: Substantial gaps exist between patient preferences and family perceptions regarding treatment intensity and decision-making roles.
2. Cultural Context Matters: The low rate of patients identifying themselves as the preferred decision-maker reflects the prevailing family-oriented approach to medical decision-making in China, where individual autonomy is often secondary to collective family responsibility.
3. Communication Gaps: The variation in concordance across different treatment types (e.g., higher for invasive procedures like tracheotomy, lower for routine measures like blood transfusion) suggests that families may engage in explicit discussions about severe interventions but fail to communicate preferences for supportive care measures.

Clinical Implications
The paper concludes that healthcare professionals should not assume family members can accurately represent patient preferences. The authors advocate for:

• Routine assessment of patient–family concordance to identify discrepancies early.
• Initiating structured advance care planning and shared decision-making discussions earlier in the disease trajectory.
• Facilitating open communication among patients, families, and healthcare teams to align care with patient values while respecting cultural sensitivities.
• Improving the accessibility of home-based palliative care to better support preferences for dying at home, which may be hindered by current service limitations in mainland China.

The study acknowledges limitations, including its single-center design, cross-sectional nature (capturing only a single time point), and the potential divergence between stated preferences and actual decisions made during critical illness. However, it asserts that these findings underscore the complexity of EOL decision-making for older adults and the urgent need for interventions that improve patient-centered care through better communication and planning.