• End-of-life Decision Making Among Patients and Proxies

      Hinderer, Katherine Ann; Friedmann, Erika (2010)
      Background: End-of-life (EOL) decision making, including decisions to withdraw mechanical ventilation, can be extremely difficult for patients and their proxies. Advance directives provide proxies with guidance related to EOL decisions. Proxy accuracy ranges from 58% to 79% with the average around 68% (Shalowitz et al., 2006). Patients and proxies have varying levels of trust in the decisions proxies make. Proxy experience with EOL decisions may influence future decisions. Purpose: To explore patient and proxy decisions related to EOL care and treatment preferences in scenarios representing three distinct disease trajectories. Patient and proxy decisions were explored, and compared for concordance, as they related to the different disease trajectories and directives given by the patient in each of the scenarios. Methods: This was a secondary data analysis of the "From Contract to Covenant in Advance Care Planning" study (Fins et al., 2005). Participants included 59 patients and 103 inexperienced and experienced proxies. Hypotheses were tested with correlations, crosstabulations, t-tests, repeated measures ANOVAs, and generalized estimating equations. Results: Patients' mean age was 77.1+7 years, 91.5% were female, and 96.6% were Caucasian. Inexperienced proxies' mean age was 57.6 +16 years, 56.9% were male, and 100% were Caucasian. Experienced proxies' mean age was 60.1+12.5 years, 76.9% were female, and 96.2% were Caucasian. Patient/inexperienced proxy agreement was 48-94% and was highest in the poor prognosis stroke disease trajectory. Modification to directives improved concordance in the uncertain prognosis heart failure trajectory. Proxy agreement ranged from 70.9-95%. GEE analyses revealed proxies with lower education and no history of serious illness were least likely to remove patients from the ventilator. Proxy experience and valence (do everything/do nothing) nested within disease trajectory predicted removal (p = < .05); gender and age did not. Patient trust (33.44+6.8) in proxy decisions was higher than proxy trust (30.19+5.5) in their own decisions (p = < .05). Conclusion: This study supported the importance of directives in EOL decision making. In examining agreement and differences in proxy and patient decisions within different disease trajectories, healthcare providers can better understand patient conditions where further interventions are needed. These data suggest that further research related to patients, proxies, and EOL decisions is needed.