Original ArticleA Decade of Changes in Family Caregivers' Preferences for Life-Sustaining Treatments for Terminally Ill Cancer Patients at End of Life in the Context of a Family-Oriented Society

ContextTemporal changes in different family caregiver cohorts' preferences for life-sustaining treatments (LSTs) at end of life (EOL) have not been examined nor have the concept of whether caregivers' LST preferences represent a homogeneous or heterogeneous construct. Furthermore, LST preferences are frequently assessed from multiple treatments, making clinical applications difficult/infeasible.ObjectivesTo identify parsimonious patterns and changes in the pattern of LST preferences for two independent cohorts of family caregivers for terminally ill Taiwanese cancer patients.MethodsPreferences for cardiopulmonary resuscitation, intensive care unit care, cardiac massage, intubation with mechanical ventilation, intravenous nutritional support, tube feeding, and dialysis were assessed among 1617 and 2056 family caregivers in 2003-2004 and 2011-2012, respectively. Patterns and changes in LST preferences were examined by multigroup latent class analysis.ResultsFive distinct classes were identified: uniformly preferring, uniformly rejecting, uniformly uncertain, and favoring nutritional support but rejecting or uncertain about other treatments. Class probability significantly decreased from 29.3% to 23.7% for the uniformly rejecting class, remained largely unchanged for the uniformly preferring (16.9%-18.6%), and favoring nutritional support but rejecting (37.1%-37.5%) or uncertain about other treatments (8.0%-10.4%) classes, but significantly increased from 7.0% to 11.5% for the uniformly uncertain class over time.ConclusionFamily caregivers' LST preferences for terminally ill cancer patients are a heterogeneous construct and shifted from uniformly rejecting all LSTs toward greater uncertainty. Surrogate EOL-care decision making may be facilitated by earlier and thorough assessments of caregivers' LST preferences and tailoring interventions to the unique needs of caregivers in each class identified in this study.