Assessment of Surrogates’ Knowledge of Patients’ Treatment Goals and Confidence in Their Ability to Make Surrogate Treatment Decisions

Abstract

This study uses telephone interviews of patients and their surrogates to evaluate surrogates’ knowledge of patients’ treatment goals and confidence in their ability to make patients’ treatment decisions.

At patients’ end of life, surrogates are frequently called upon to make health care decisions for patients who lose capacity. Numerous studies have demonstrated that these surrogates lack knowledge of patients’ preferences,¹ but patients cite the belief that their loved ones already know their wishes as a reason for not engaging in advance care planning (ACP).² Little is known about how prepared surrogates feel to make decisions on behalf of loved ones. Similar to patients, if surrogates already believe that they know their loved ones’ wishes, they may not see the need to engage in ACP. Through telephone interviews, we examined surrogates’ confidence in their knowledge of patients’ treatment goals and their actual knowledge of these goals.

Methods

Participants were randomly selected from a list of community-living Veterans, 55 years or older, receiving primary care within the VA Connecticut Healthcare System (patients) with oversampling of women and minorities. Patients were asked to identify the person they would choose to make medical decisions on their behalf if they were unable (surrogates).³ The study protocol was approved by the Human Subjects Subcommittee of the VA Connecticut Healthcare System. Participants provided verbal consent and received $20.

Patients and surrogates completed separate closed-ended interviews conducted by telephone between April 2014 and February 2016. To assess goals, patients were asked to rate 3 health states that could result from treatment of serious illness (Box) as acceptable or unacceptable. Surrogates were asked to rate as they thought the patient would. Surrogate knowledge was assessed as agreement between surrogate and patient ratings for each state individually and for all 3. Surrogate confidence was assessed with a single item (Box). Confidence level responses ranged from 1 (not at all confident) to 5 (extremely confident).

Box. Patient’s Health States and Surrogates’ Confidence Measures.

Severe Physical Disability

• Being bedbound and requiring assistance with bathing, dressing, grooming, and toileting

Severe Cognitive Disability

• Being unable to recognize family members

Severe Pain

• Daily pain feeling like a broken bone or appendicitis

Confidence Measures

• Consider a situation in which your loved one could not speak for himself or herself, and the doctors asked you to make a medical decision on his or her behalf. Imagine that the decision involved a treatment that could result in one of the states we just talked about. How confident are you that you know and accept your loved one’s views on these states well enough to make a decision that represents his or her views?

Agreement between patient and surrogate ratings was examined using the κ statistic. The association between surrogate confidence and knowledge was analyzed using the χ² statistic, with confidence categorized as extremely confident vs less than extremely confident, based on the distribution of the responses.

Results

The 349 patient participants had a mean (SD) age of 66 (9) years; 68% were men (n = 237); and 36% were nonwhite (n = 126). Of the 349 surrogates 78% were women (n = 272); 66% were white (n = 231); and 52% were the patient’s spouse or partner (n = 182). While 75% of surrogates rated themselves extremely confident (n = 261), only 21% of surrogates (n = 72) knew the patients’ ratings for all health states: 23% (n = 60) among those extremely confident vs 14% (n = 12) among those less confident (P = .07). Agreement between surrogate and patient ratings for each health state ranged from 54% (κ = 0.08) to 59% (κ = 0.15) and did not differ according to confidence level (Figure).

Figure. Proportion of Surrogates Agreeing With Patient Health State Rating.

The dark blue columns represent the proportion of extremely confident surrogates who agreed with the patient on each health state rating. The light blue columns represent the proportion of less than extremely confident surrogates who agreed with the patient on each health state rating.

Discussion

Surrogates’ confidence in their ability to make treatment decisions based on knowledge of patients’ ratings of the acceptability of health states resulting from treatment of serious illness far exceeded and was not associated with their actual knowledge of these ratings. This study adds to a small evidence base consistently demonstrating high confidence among surrogates in their ability to make end-of-life decisions.⁴ In a study comparing surrogate confidence with knowledge, 79% of surrogates rated themselves as confident or very confident, but only 35% knew patients’ goals.⁴

Surrogates may be overestimating their confidence in advance of facing an actual decision. Substantial proportions of surrogate decision makers who have made end-of-life decisions experience burden, expressing stress, guilt, and doubts about having made the right decision.⁵ An intervention involving facilitated discussions between patients and families regarding the patient’s goals was shown to reduce symptoms of posttraumatic stress, depression, and anxiety among family members of patients who died.⁶ However, these conversations can be difficult, and they may be less likely to occur with surrogates who are highly confident and think the conversation unnecessary. Confidence requires consideration as a barrier to ACP, with surrogates requiring greater understanding of what they need to know to lessen the burdens of surrogate decision making.