Reviewed by Dr. Chad Guenther
Enguidanos, S, and Ailshire, J. J Pain Symptom Manage 2017 Jan;53(1):49-56
There is widespread speculation about whether the timing of completion of advance directives (AD) is related to the aggressiveness of care (AC). This article presents the first empiric data on the relationship of timing of AD and AC as well as the relationship of certain patient characteristics with timing of AD or AC. The authors use data from the Health and Retirement Study, a longitudinal survey of U.S. adults age 51 and older queried every 2 years until death followed by an exit interview with a proxy (typically family members). Patients included in this study expired between 1997 and 2013. AD completed within the last 3 months of life were defined as “late.” AC was defined as unconditionally giving all possible care to prolong life. Bivariate analysis was performed comparing health status and sociodemographic data by AC and timing of AD. Multivariate logistic regression analysis was performed comparing factors associated with late AD and AC.
Out of 9164 respondents with an exit interview, 45% completed an AD before death. The median time for AD completion was 41 months before death. Overall, 5.35% of patients with AD preferred AC, mainly younger respondents (age 50-74), ethnic minorities (defined as Hispanic or black), those with less education, and those with greater overall monetary assets. There was no association between specific health condition and preference for AC. Late completion of AD was associated with younger age (age 50-74), ethnic minorities, lowest quartile of monetary assets, and underlying cancer or lung disease. Overall, preference for AC was relatively low with AD completed 4 months to 10+ years before death, but rose to 12% with AD completed 2-3 months before death and then dropped to 3.9% for AD completed in the last month of life. Factors associated with preference for AC were lower education and minority ethnicity. Patients whose death was expected were less likely to choose AC.
Prior to this study, there was limited data to help understand trends in patient decision-making regarding end-of-life care and advanced planning. The authors hypothesized that both early and late completion of AD would be associated with preference for AC. While later completion of AD was associated with preference for AC, early completion of AD was not. Given that only 5.35% of all patients who complete AD prefer AC, the sharp increase in preference for AC at the end of life deserves discussion. Since ethnic minorities were more likely to choose AC and also have late AD, it is possible that some of the increase in AC seen later is due to the effect of greater representation of this subpopulation. Also, it is possible that early AD include more respondents who understand that the standard of care is aggressive and want to make their preferences known to limit end-of-life care. Most importantly, there is a possible association with increased AC with late AD in the setting of hurried goals of care discussions prior to emergent procedures. With less time for the patient and their family to think as well as less time for an in-depth discussion with the care provider, a patient’s decision may be less likely to reflect their true preferences and values. Thus, goals of care discussions may be more likely to represent a patient’s values when performed earlier in their care.