Modifiable Risk Factors of Prolonged Grief Disorder, Posttraumatic Stress Disorder, Depression, and Anxiety Symptom States After ICU Loss

BACKGROUND: After loss, prolonged grief disorder (PGD), posttraumatic stress disorder (PTSD), depression, and anxiety often co-occur, intensifying distress. Correlates of these patterns remain understudied, as most work omits anxiety, uses cross-sectional designs, or focuses on fixed factors. Modifiable ICU end-of-life care factors are largely overlooked despite global calls for improvement. OBJECTIVES: To examine factors associated with states of PGD, PTSD, depression, and anxiety, emphasizing modifiable ICU end-of-life care over 2 bereavement years. DESIGN: Prospective cohort study. SETTING: Two Taiwanese medical ICUs. SUBJECTS: Two hundred ninety-one family members. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Multinomial logistic regression examined factors associated with four symptom states: resilient (reference), subthreshold PGD-depression, PGD-dominant, and co-occurring PGD, PTSD, depression, and anxiety states. Factors spanned intrapersonal, interpersonal (perceived social support), and contextual (bereavement-related and death-circumstance) domains. Higher-burden states were associated with lower education, financial hardship, mixed perceived social support, spousal loss, death of a female-cancer patient, shorter time since death, receipt of palliative care, and suboptimal end-of-life indicators, including absence of social worker involvement or a do-not-resuscitate order, family presence at death, and less than high perceived quality of dying and death (QODD). Do-not-resuscitate orders and social worker involvement were linked to lower odds of the co-occurring PGD, PTSD, depression, and anxiety state, while palliative care and family presence at death were associated with increased odds. Compared with high QODD, moderate QODD was associated with all higher-burden states, while poor-to-uncertain or worst QODD was associated with the subthreshold PGD-depression and co-occurring PGD, PTSD, depression, and anxiety states. CONCLUSIONS: Higher-burden symptom states reflect both unmodifiable vulnerabilities and potentially modifiable peri- and post-loss factors, including interpersonal and death-related circumstances. These findings suggest that aspects of end-of-life care-such as QODD, family support (especially when encouraging presence at patient death), timely do-not-resuscitate orders, social worker involvement, and early palliative care referral-may be relevant targets for future interventions to mitigate concurrent emotional distress in bereaved families.