General Medicine: Open Access

ISSN - 2327-5146


Surrogate End-Of-Life Decisions: Projection or Simulation

Safa Hammami, Faduma A Farah and Muhammad M Hammami

Background: Substituted judgment at the end-of-life is often needed to extend autonomy of decisionally incompetent patients who don’t have advance directive. Its moral appeal centers on surrogates’ ability to simulate patients’ choices as opposed to projecting their personal choices. The accuracy of the standard substituted judgement model has been questioned, and an integrative patient’s life-story narrative version has been suggested. The accuracy of the latter cannot be evaluated using the common methodology of life-saving intervention and binary-decision scenarios. We explored the accuracy of the life-story narrative version of substituted judgment using forced-ranking and Q-methodology.

Methods: A volunteer sample of seven family pairs (3 husband-wife, 1 parent-child, 3 sib-sib) participated in a pilot study. All 14 respondents rank-ordered 47 opinion statements on end-of-life issues using a 9-category symmetrical distribution according to their personal preferences (personal sort), and 11 of the 14 according to predicted preferences of individuals in their pair (surrogate sort). Accuracy of within-pair prediction was explored by comparing: 1) correlation of individual statements’ ranking scores between surrogate sorts and their paired-personal sorts (simulation) to correlation between surrogate sorts and their self-personal sorts (projection), and 2) percentage of surrogate sorts that co-loaded with their paired-personal sorts to percentage of surrogate sorts that co-loaded with their self-personal sorts, using factor Q-methodology.

Results: Mean (SD) age was 31.9 (8.9) years; 50% were females; 43% reported average religiosity, 65% ≥ very good health, 64% ≥ very good life quality, and 100% ≥ high school education. Mean (95% confidence interval) simulation and projection correlations were 0.222 (0.136 to 0.305) and 0.976 (0.971 to 0.980), respectively, with a mean difference of 0.754 (0.628 to 0.880, p < 0.001). With 5-factor extraction, 45% of the 11 surrogate sorts co-loaded with their paired-personal sorts and 82% with their self-personal sorts. With 3-factor extraction, the corresponding percentages were 45% and 91%.

Conclusions: Q-methodology may be useful in exploring the accuracy of substituted judgment that is based on patients’ life-story narratives. Surrogate end-of-life decisions by family members are more likely to project personal choices and life-story narrative than to simulate patient’s choices and life-story narrative.