A prospective, single-blinded, controlled, and randomized trial to find if physician's use of the Portuguese Advance Directives, in palliative care, as a communication tool between patients and caregivers, improves the agreement and reliability between patients and their health surrogates, in their decisions about end-of-life care.

This trial was a prospective, single-blinded, controlled, and randomized study analyzing if Advance Directives' discussion between patients and caregivers, promoted by a palliative care physician, improves concordance between them, concerning to patients' last wishes, therefore improving surrogates' accuracy when caregivers decide and act on behalf of the patients. This trial was conducted in a central hospital of Portugal's north inland, serving an estimated population of 465.000 habitants. Patients referenced to all three units of the palliative care service from September 2018 to September 2019, and caregivers, were sequentially approached and recruited on the first consult.

All participants gave written consent to enroll in the trial. This study occurred in two phases, within a 1-month interval. At baseline (phase 1) participants separately fulfilled the Advance Directives' document considering patients' preferences for end-of-life care. Caregivers also fulfill a similar document with their own end-of-life preferences.

Dyads were then randomly assigned to two different groups - The Advance Directives and the Control groups.

Simple individual randomization was previously informatically achieved and the randomization sequence was disclosed by a sealed envelope only after trial enrollment of the dyad, to ensure proper investigator concealment. Both patients and caregivers were blinded to the assigned group until the end of the trial. In both groups (Advance Directives and Control), dyads were engaged in a conference meeting, in the same room, with the same average duration, with the same investigator.

In the Advance Directives group, the palliative care physician promoted an open discussion between patients and caregivers, about patients' answers to the Advance Directives document. In the Control group, the palliative care physician underwent a conference with both patients and caregivers to evaluate patients' clinical symptoms.

At phase 2, one month after the first interview, caregivers were asked to fill in another Advance Directives document, as patients´ surrogates.

Each participant had the choice to drop out at any moment of the trial, and the pair was excluded from data analysis.

Portuguese official Advance Directives was the central instrument used in this trial. Participants were asked to choose 1 to 3 different scenarios, to apply the following 12 questions regarding end-of-life preferences. Each question in each scenario was informatically registered as a Yes or No item, respectively corresponding to items selected or not by all participants. For each participant, the investigators expected 36 possible answers for concordance analysis.

Categorical variables were described by absolute and relative frequencies. Age was described by the mean and standard deviation (mean ± SD), as its distribution was not deviated from normality in each group, according to visual analysis of histograms and confirmed with the Shapiro-Wilk's test of normality.

Agreement between patients and caregivers was assessed (in each question) with the Overall Proportions of Agreement (Overall PA) and Proportions of Specific Agreement (for Yes and No), with respective 95% confidence intervals. Reliability was accessed with Cohen's kappa (k).

Descriptive data analysis was performed using SPSS® Statistics. Proportions of agreement and Cohen's κ with respective confidence intervals were computed using packages "obs. agree" and "psych" from R software, v 3.4.0.

Inclusion Criteria:

• Adult patients with 18 years or more
• Patients referenced to the palliative care service with a chronic, progressive, and incurable disease
• Patients with the ability to comprehend, write and speak the Portuguese language
• Patients with an absence of major cognitive disorders (Portuguese validated Mini-Mental State Test score superior to 22, 25, or 27, according to patients' scholarly)
• Patients that accepted to participate in the trial
• Patients who were capable of nominate a caregiver as their surrogate in decision making
• Adult caregivers with 18 years or more
• Caregivers who were nominated by patients as their surrogate decision-maker
• Caregivers with the ability to comprehend, write and speak the Portuguese language
• Caregivers that accepted to participate in the trial

Exclusion Criteria:

• Patients or caregivers that refused to participate
• Patients with cognitive impairment
• Patients with incapacity to dialogue
• Patients who were too sick to cooperate
• Patients without a caregiver

All participants were blinded to the allocated intervention, from the beginning to the end of the trial. Investigator only had access to the allocation group after participants enrollment on the trial, sociodemographic data collected, and phase 1 Advance Directives fulfilled.

Preliminary data analysis was made without knowledge of the participants' allocated group