Only about one-third of US adults have completed an advance directive, despite the evidence that having one substantially improves the alignment between what patients want at end of life and what they receive. Silveira and colleagues’ 2010 NEJM analysis of 3,746 decedents in the Health and Retirement Study found that, among incapacitated patients who had a living will requesting limited care, 83.2% received care consistent with that preference; the study measured concordance between documented wishes and care, not regret. No published survey has documented significant regret among people who completed an advance directive — the literature treats under-completion, not over-completion, as the problem. The Detering 2010 BMJ randomised trial of advance care planning in 309 elderly hospital inpatients found that, among families of the patients who died, 30% in the no-planning control group showed clinically significant depression versus 0% in the planning group, with similar gaps for anxiety (19% vs 0%) and high PTSD risk (15% vs 0%).
What makes the deferral decision costly is that it often becomes permanent. The roughly two-thirds of US adults without any advance directive (36.7% completion in the 2017 Health Affairs review) did not all consciously decide to defer — many simply never got around to it. The legal infrastructure for advance directives in the US involves 50 different state forms with varying requirements for witnesses, notarization, and scope; completing an AD in one state and dying in another creates enforcement uncertainty. The evidence on when advance directives are consulted and followed is also imperfect: documents that exist in filing cabinets but not in electronic health records have limited practical effect. The completion decision and the accessibility decision are not the same.
The Detering RCT’s setting — elderly hospital inpatients in Australia, median age in the mid-80s — is different from middle-age preventive ACP completion. The benefit of completing an AD at 45 rather than 79 likely operates through a different mechanism: it forces a conversation about values and preferences that becomes more difficult when illness has already begun, it reduces the burden on surrogates who must guess at preferences without guidance, and it avoids the scenario where capacity is lost before the conversation can be had. No study to date has documented significant regret among people who completed an advance directive — the consistent finding is that too few people complete one, not that completers wish they hadn’t. The main cost of acting early is administrative; the main cost of waiting is the non-trivial probability that waiting becomes permanent.