Surrogate Informed Consent: A Qualitative Analysis of Surrogate Decision Makers’ Perspectives

Abstract

Rationale: Clinical critical care research often hinges on surrogate informed consent, as patients commonly lack decision-making capacity because of their acute illness. The surrogate informed consent process has been identified as having flaws and needing improvement. A better understanding of surrogates’ perspectives is required to understand and address these shortcomings and thereby improve this process.

Objectives: To explore the perspectives of surrogate decision makers of critically ill, mechanically ventilated patients about being approached about having their loved one participate in hypothetical research studies.

Methods: We performed semistructured qualitative interviews of surrogate decision makers of critically ill, mechanically ventilated patients, exploring their decisional needs surrounding participation in research. These interviews were recorded and transcribed verbatim. A thematic analysis of transcripts was performed with an iterative group framework using a mixed inductive and deductive approach.

Results: A sample of 21 surrogate decision makers were interviewed. Thematic saturation was achieved by the consensus of the investigators. We identified trust as a unifying domain for the themes that emerged through the analytic process. Embedded within the domain of trust, two central themes became apparent: knowledge-based trust and context-based trust. Knowledge-based trust includes subthemes of logistics, accountability, and mutual respect, whereas context-based trust includes trust in the individual clinicians and trust in the hospital system.

Conclusions: Our findings highlight the nuanced layers of trust central to the surrogate informed consent process. To enhance the surrogate informed consent process for participation in critical care research studies, it is crucial that researchers recognize the inherent importance of trust to ensure an effective informed consent process.

As the global population ages and medically complex patients live longer, the demand for critical care services and research continues to increase (1). Recently, the coronavirus disease (COVID-19) pandemic has highlighted the need for high-quality randomized controlled trials to be designed and implemented expeditiously without sacrificing ethical standards (2). Research in the intensive care unit (ICU) is particularly challenging, as the majority of study subjects do not have decisional capacity because of their acute illness and the use of sedative agents (3, 4). Therefore, surrogate decision makers (SDMs) often provide informed consent for subjects to participate in research. When approached for potential research involvement, SDMs must make substituted judgments on behalf of an incapacitated patient to preserve their autonomy and make decisions in their best interest. The process of surrogate informed consent has been recognized as anxiety provoking for the surrogate and does not always accurately reflect the patient’s preferences (5, 6). In a study of critical illness survivors, nearly 80% of patients were in favor of surrogates making informed consent decisions on their behalf for research (7). Accordingly, improving the surrogate informed consent process has been identified as an area of focus to further advance critical care research (8).

Surrogate decision-making for ICU research consent has previously been organized into three stages: 1) being approached, 2) reflecting on participation, and 3) making a decision (9). Although informed consent always involves the disclosure of study details and rationale, surrogates often have a poor understanding of what the study entails or the potential risks (10). Yet assessment of potential risks is a critical juncture regarding potential study participation. After determining the amount of risk with a potential study intervention, an individual’s trust in healthcare professionals is crucial for potential study involvement (9). With regard to choosing medical or surgical treatments, patients often rely more on trust in their doctor than a scrupulous analysis of the information provided (11). The role of trust in surrogate informed consent for research participation has not been extensively explored. A better understanding of SDMs’ perspectives and thought processes regarding informed consent is required to improve the conduct of research and ensure an ethical informed consent process that reflects a patient’s known wishes (9).

The aim of our study was to determine central themes in decision-making around surrogate informed consent in critical care research to guide future methodology, standardization, and improvement of the surrogate informed consent process.

Methods

Participants and Recruitment

We recruited SDMs for mechanically ventilated patients lacking decisional capacity (informally deemed by their intubated and sedated status) admitted to the University of Colorado Hospital (Aurora, CO) medical ICU. Surrogates were approached within 96 hours of ICU admission. The SDM had to be 18 years or older and previously identified as the patient’s decision maker (defined as healthcare proxy or medical durable power of attorney) during the current hospitalization. Surrogate and patient dyads were identified after being screened for but not meeting specific inclusion criteria for ongoing critical care research studies. Surrogates provided informed consent and received a $25 VISA gift card upon completion. This study protocol was approved by the Colorado Multiple Institutional Review Board (CO-MIRB Protocol 12-0206).

Data Collection

We collected patient demographic information via chart abstraction and surrogate demographic information via questionnaires. Two 14-item surveys were administered to SDMs, the Hospital Anxiety and Depression Scale (HADS)—a widely used and validated instrument that assesses symptoms of anxiety and depression—as well as the Perceived Stress Scale (PSS)—a psychological instrument that evaluates feelings and thoughts of stress in the preceding month to quantify levels of stress (12 –14). Qualitative interviews were performed in the fall of 2014 with SDMs of patients with acute respiratory failure requiring mechanical ventilation. The interview questions were developed iteratively by the investigators (C.N., M.M., and D.D.M.) on the basis of an inductive framework. The interview questions were open-ended and aimed to identify essential components required to provide or decline informed consent for critical care research (see appendix in the online supplement). These interviews were conducted in a hospital meeting room and were audio recorded and transcribed verbatim. The third author, C.N. (a social worker with over 20 years of qualitative interviewing experience) conducted the interviews.

Qualitative Analysis

A thematic analysis of transcripts was performed with an iterative group framework using a mixed inductive and deductive approach (15). Thematic saturation was achieved by consensus of the investigators to develop a thematic framework (16). Initially, 10% of the transcripts were coded individually by group members. Our team had considerable coding overlap, allowing us to easily develop an initial codebook, which was created synergistically based on the preliminary findings and their applicability to the research question. The research team met on three separate occasions to discuss emerging codes, including modification, inclusion, and exclusion of additional codes. Codes were then applied to the remaining transcripts by a master’s level qualitative research expert (E.B.) and adjusted appropriately. The final codes were then presented to, and cross-referenced by, the research group to ensure validity. Codes were subsequently organized based on shared subject matter into themes and subthemes inductively. We used ATLAS.ti version 8.4 software (ATLAS.ti) for coding and analysis. Some of the results of this study have been previously reported in the form of abstracts (17, 18).

Results

We approached 34 SDMs for this study, and 21 (62%) agreed to be interviewed. The average age of SDMs was 57 years, 71% of whom were women, and relation to the patients included 33% spouses, 33% children, 19% siblings, and 15% parents. SDMs had, on average, known the patient for 43 years, and 38% of surrogates were either healthcare professionals or had a family member with medical training. The SDM average HADS scores for the anxiety and depression domains were 9 ± 5 and 7 ± 4, respectively, whereas the average PSS score was 34 ± 4. Eight surrogates (38%) had discussed research previously with patients, predominantly in the context of past research participation. Additional surrogate details are listed in Table 1. Of the 13 SDMs who declined consent, 5 were not interested in participating, there was a language barrier with 3, 2 were unwilling to make time for the interview, 1 had “too much going on,” 1 was transitioning to comfort care on the day of eligibility, and 1 would not clarify their rationale for declining. The average age of patients was 62 years, 35% were women, and 80% were white. The admitting diagnosis was most commonly respiratory failure, followed by sepsis, and 100% were mechanically ventilated. The median ICU length of stay was 12 days (interquartile range [IQR], 5–17 d). Additional patient details are listed in Table 2. Surrogates were interviewed in person, and the interviews ranged from 31 to 80 minutes in duration, with an average of 48 minutes. Surrogates were interviewed on a median of the fifth hospital day (IQR, 2–7 hospital d).

Table 1.

Characteristics of surrogate decision makers

Surrogates, n	21*
Age, yr, mean, SD	57 ± 11
Sex, female, n (%)	15 (71)
White race, n (%)	17 (81)
Relationship to patient, n (%)
Spouse	7 (33)
Daughter/son	7 (33)
Parent	3 (15)
Sibling	4 (19)
Highest level of education, n (%)
High school	4 (19)
College	7 (33)
College graduate	7 (33)
Graduate school	3 (15)
Years known patient, yr, mean, SD	44 ± 11†
Healthcare professional or family member with medical training, n (%)	8 (38)†
Previously discussed research with patient, n (%)	8 (38)
HADS score, mean, SD‡	16 ± 9
Anxiety domain, mean, SD	9 ± 5
Depression domain, mean, SD	7 ± 4
PSS score, mean, SD§	34 ± 4

Definition of abbreviations: HADS = Hospital Anxiety and Depression Scale; PSS = Perceived Stress Scale; SD = standard deviation.

^*One patient had identified two surrogates.

^†Missing data on two participants.

^‡Scores on the HADS range from 0 to 42 (21 points for each anxiety and depression domains), with higher scores indicating worse symptoms.

^§PSS scores range from 0 to 56, with higher scores associated with greater vulnerability to stressful life-event–elicited depressive symptoms.

Table 2.

Characteristics of patients

Patients, n	20*
Age, yr, mean, SD	62 ± 15
Sex, female, n (%)	7 (35)
White race, n (%)	16 (80)
Admission diagnosis†, n (%)
Respiratory failure	12 (60)
Sepsis	10 (50)
Pneumonia	5 (25)
Pancreatitis	1 (5)
Other	7 (35)
Disposition location, n (%)
Died	8 (40)
LTACH	5 (25)
SNF	4 (20)
Home	3 (15)
Mechanical ventilation, n (%)	20 (100)
APACHE II score, mean (SD)	23 (7)
ICU length of stay, d, median (IQR)	12 (5–17)

Definition of abbreviations: APACHE = Acute Physiology and Chronic Health Evaluation; ICU = intensive care unit; IQR = interquartile range; LTACH = long-term acute-care hospital; SD = standard deviation; SNF = skilled nursing facility.

^*One patient had identified two surrogates.

^†Some patients had >1 admitting diagnosis.

We identified trust as a unifying domain for the themes that emerged through the analytic process. Embedded within the domain of trust, two central themes became apparent: knowledge-based trust and context-based trust. Knowledge-based trust includes the subthemes of logistics, accountability, and mutual respect, whereas context-based trust includes trust in the individual clinicians and trust in the hospital system. Outside of the domain of trust, surrogates also identified karma and altruism as additional motivators to potential research participation. A medical background was another minor theme that we observed among some of the surrogates.

Theme 1: Knowledge-based Trust

Logistics

As a foundation for potential study enrollment, SDMs expressed the importance of understanding the logistical details of the study. They desired to be completely informed of the nuances of the study as well as potential drawbacks to enrollment:

“I’d want to know exactly what it was and what research you would be doing and how you’d be going about doing it before I would want her to be a part of it.” #1, Daughter.

SDMs desired complete transparency, including preliminary findings and previous outcomes:

“I think that I would want to know how many other people have done this. What was the outcome of those people? If I understand research […] there’s first round trials and there’s second round trials and the list goes on and on, so where are we at in that?” #2, Daughter.

Accountability

In conjunction with being apprised of the logistics of a potential research study, surrogates also believed that the research team should be accountable for ensuring the legitimacy of the research itself as well as minimizing the risks of harm:

“That would pique my interest a little bit because it’s […] funded by somebody that is, you know, a larger entity trying to do something to benefit the patients.” #3, Son and Daughter.

Surrogates identified that study involvement could lead to serious consequences that may cause personal guilt:

“Well, it’s kind of hard [be]cause you could [go] both ways with that. You could say yeah, I’d be willing to try anything to save my mom’s life, but then again, would you? Because if you do, you may very well have made the decision that killed her.” #4, Daughter.

One surrogate wanted to know who was held responsible for adverse events related to potential study interventions or if the study conclusion was overall harmful:

“What happens if you have ill effects from a medication study or you’re on a ventilator and they add too much, you know, oxygen or something like that, then what will the study do? Will they be responsible?” #1, Husband.

Mutual respect

Furthermore, SDMs needed the study coordinators informing them to be respectful, personable, and authentic to build rapport and trust:

“I’m going to feel way less guilty that my choice could be bad if I feel like I’m a person to you rather than a number. I think that sometimes with studies […] somebody may feel like they’re just a number. I don’t want to feel like I’m a number.” #2, Daughter.

Surrogates wanted honest and unfiltered information even if it meant there were potential bleak outcomes:

“I don’t want a bunch of bull. You know, if you do this your survival rate here is 10%. I want to know that.” #1, Father.

SDMs shared the potential for increased trust if researchers went out of their way to not gloss over the potential downsides of study enrollment:

“I don’t want a surprise [be]cause I just think the more you lay it out just plain and simple, it’s going to be a lot less stress for everybody.” #1, Wife.

Theme 2: Context-based Trust

Trust in physicians

SDMs’ trust in the clinicians taking care of their loved ones oriented many to be open to the idea of potential research enrollment:

“If the doctor was recommending or asking you to participate in the study, I think that would be one of the things that would push you toward making that decision.” #2, Husband.

On the other hand, some expressed concern that by having physicians consent surrogates for research, there is a possibility that it would be more difficult to decline enrollment in a study:

“If you have the doctor do it [obtain consent] then that’s maybe not a good thing because they might feel not necessarily coerced, but pressure to, you know, do what the doctor wanted.” #3, Son and Daughter.

Trust in the hospital system

Some surrogates shared a near reflexive willingness to consider research participation based on the positive reputation of the hospital where they were being treated and their past experience:

“We’re on the same team and I trust the community here because I brought my dad here, and for them to approach me and say, you know, we think your dad could benefit from participating in this research. It kind of makes it an easy answer.” #1, Son.

Minor Themes: Research Karma, Altruism, and Medical Background

Although several interviewees expressed a belief that research was intended to benefit the individual patient, known as a therapeutic misconception, others identified different motivating factors for participation (19). Some interviewees recognized that the patient they represented as surrogates had already benefited in their life from past research:

“We have benefited from it [research]. When she was 12 years old and had cancer, and all this really got started you think of the technology in the last 46 years. We’ve benefited from a lot of it because we wouldn’t be here if it wasn’t for that.” #2, Father.

Some expressed their interest in future involvement being motivated by the potential for individual gain, such that volunteering for research could result in positive karma for them. Furthermore, there was a very strong signal of altruism, such that engaging in research would ultimately help others more than the individual participant:

“I think my dad would want to participate with that because he’s benefiting now from people prior to him that underwent the same process, that got the health care industry to where we are today. If we can help, you know, move health care forward for tomorrow or the next day, he would be for that.” #5, Daughter.

Roughly a third of surrogates described the importance of their personal or family medical background when approaching the decision to participate in research, predominately with a positive outlook secondary to this experience:

“My dad was a biomedical engineer so all of our lives we’ve been around hospitals and learned that this is the latest, greatest, coolest thing and this is what it does for you and it saves peoples’ lives […] we’re all in the mindset of it’s good to try new things.” #4, Daughter.

Discussion

Our study of SDMs of critically ill patients illustrates the central role of trust, in its many different forms, as imperative to decision-making for potential research enrollment in ICU settings. Prior studies have outlined the importance of clearly informing surrogates as to the risks and benefits of potential study enrollment, whereas our analysis emphasized the underlying theme of trust weaved within all aspects of information delivery (4, 20 –22). The surrogates sampled in our study stressed a desire for complete transparency in all stages of the consent process and how this forthrightness would engender trust in both the individuals obtaining consent as well as the research process itself. To establish trust, SDMs needed accountability, transparency, and mutual respect throughout this process to consider involving their loved ones in an ICU research study.

The process of informed consent in medicine has been developed to rely on accountability and transparency more than trust. However, no matter how explicit or specific informed consent may be, the additional element of trust is always required. There are many factors that increase or decrease trust in the process of informed consent as well as in the people conducting research, and it is important to recognize and understand these reasons. In the setting of informed consent, “well-judged” trust has been contrasted to “immature” or “blind” trust, with the latter described as “a form of deference that could make sense only on the simplistic assumption that others have goodwill toward us” (23). Our qualitative findings of the complexities of trust within the surrogate informed consent process align with this concept of “well-judged” trust (23).

Hutchison and colleagues (24) studied trust in the ICU between surrogates and clinicians at an academic medical center and found that trust in the institution or medical profession was the starting point for the development of trust. Beyond this, there were five dimensions (technical competence, communication, honesty, benevolence, and interpersonal skills) of trust contributing to the degree that trust was established by the surrogate during an ICU admission (24). This research was not focused on surrogate informed consent, but some of these paradigms can likely be translated into this realm.

By synthesizing past work above with our findings, we have established a framework (Figure 1) to conceptualize trust in the surrogate informed consent process. The foundation for trust in the surrogate informed consent process begins with context-based trust: trust in the hospital system and trust in physicians. Without trust in both of these, SDMs are unlikely to proceed forward with potential study enrollment. Some surrogates may have “blind” or “immature” trust at this point and accept to participate without becoming fully informed (23). This misses the true purpose of informed consent, and we have labeled it as “cursory informed consent” in our framework. There are no consistent measures to prevent this from happening, although some have proposed methods to assess or test surrogate comprehension during informed consent (25). If a surrogate has context-based trust, then they can move on to develop knowledge-based trust to thereby make a thoughtful decision about potential study participation. As discussed above, our findings showed the importance of logistics, transparency, accountability, and mutual respect in this phase of trust development. Trust also depends on communication and interpersonal skills. Ultimately, these knowledge-based factors are assessed and help a surrogate come to an ethically sound informed consent decision using substituted judgment.

Figure 1.

Trust in the surrogate informed consent process.

The population of individuals we interviewed were under considerable psychological distress because of the ongoing illnesses of their loved ones. This was demonstrated within the interviews as well as by the participants’ moderately elevated scores on two validated tools, HADS and PSS, with average scores of 16 and 34, respectively (13, 14). This ongoing stress has been acknowledged as one of the common challenges with surrogate informed consent (26). For instance, in a multicenter survey study of 284 family members of ICU patients, a third of all participants and nearly half of those who shared in decision-making were found to have substantial symptoms of post-traumatic stress (27). We suspect the underlying psychological distress innate to having a critically ill family member likely leads to feelings of loss of control and helplessness, highlighting the importance of trust in the surrogate informed consent process.

On the basis of our findings as well as prior similar research, the process of surrogate informed consent warrants further investigation and improvements. Unfortunately, there continues to be an ongoing overemphasis on the written documents of informed consent, whereas the actual process of surrogate informed consent continues to be much less clearly defined and operationalized (28). A recent study evaluated behavioral nudges embedded within a preconsent survey for a sham ICU trial compared with standard recruitment methods. This study showed no significant difference in study enrollment. This study is compelling, as it explicitly tested a unique intervention targeted at the conceptual biases and heuristics of decision-making involved with surrogate informed consent (29). We believe the surrogate informed consent process lends itself to a shared decision-making approach defined as “a collaborative communication process between clinicians and [surrogates] that integrates the best evidence available with the patients’ values and preferences, to promote high-quality health care decisions” (30). The development of decision aids to complement this process is one potential approach to enhance the surrogate informed consent process but would require additional thought and oversight to minimize bias (30). Lastly, there should also be further development of processes to obtain informed consent from ICU patients who are mechanically ventilated yet retain their decisional capacity (26).

Strengths and Limitations

The strengths of our study include rigorous qualitative methods, including a multidisciplinary team-based process to triangulate the findings and avoid preconceived biases from overly influencing the presentation of the results. In addition, these surrogates were interviewed during the time that their loved one was in the ICU, and their perspectives were salient to the ongoing stressful and anxiety-provoking situation akin to when one would be approached for potential study consent. We adhered to the consolidated criteria for reporting qualitative research to provide methodological transparency. We did not perform a member check of our analysis (see online supplement) (31).

Our study has several important limitations. First, as informed consent was required to participate in this study, there is selection bias because we did not capture the valuable perspectives of those who decline research participation generally. Second, our sampled population was disproportionately female, white, and from a single institution, limiting the transferability of these findings to other populations or settings. Lastly, the interview guide was initially written with a dual purpose, both to explore surrogate decision-making as well as to inform and later create a video to help explain informed consent in the ICU. This latter and more direct intention may have led to some probing or elicitation from the interviewer to ensure participants discussed specific aspects of the interview guide.

Conclusions

Our findings highlight the nuanced layers of trust central to the surrogate informed consent process in critical care research. To enhance the surrogate informed consent process, it is necessary but not sufficient that researchers recognize the importance of trust. Further emphasis must be placed on shared decision-making to ensure surrogate informed consent is a high-value and patient-centered process in addition to a fundamentally ethical–legal process (30).