Digging Deeper: Understanding Trajectories and Experiences of Shared Decision-Making for Primary Prevention ICD Implantation

Abstract

Background:

Shared decision-making using a decision aid is required for patients undergoing implantation of primary prevention implantable cardioverter-defibrillators (ICD). It is unknown how much this process has impacted patients’ experiences or choices. Effective shared decision-making requires an understanding of how patients make ICD decisions. A qualitative key informant study was chosen to capture the breadth of patients’ experiences making ICD decisions in the context of required shared decision-making.

Methods and Results:

We conducted in-depth interviews with 20 patients referred to electrophysiology clinics for the consideration of primary prevention ICD implantation. Purposeful sampling from a prior survey study evaluating mandated shared decision-making was based on patient characteristics and responses to the initial survey questions. Qualitative descriptive analysis of the interviews was performed using a multilevel coding strategy. Patients’ paths to an ICD decision often involved multiple visits with multiple clinicians. However, the decision aid was almost exclusively provided to the patient during electrophysiology clinic visits. Some patients used the numeric data in the decision aid to make an ICD decision based on the risk–benefit profile; others made decisions based on other data or based on trust in clinicians’ recommendations. Patients highlighted information related to living with the device as particularly important in helping them to make their ICD decisions. Some patients struggled with the emotional aspects of making an ICD decision.

Conclusions:

Patients’ ICD decision-making paths poses a challenge to episodic shared decision-making and may make tools such as decision aids perfunctory if used solely during the electrophysiology visit. Understanding patients’ ICD decision-making paths, especially in the context of encounters with primary cardiologists, can inform the implementation strategies of shared decision-making help to enhance its impact. Components of decision aids focusing on the experience of living with an ICD rather than probabilistic data may also be more impactful, although the nature of their impact will differ.

Lay Summary

• Patients’ decisions about primary prevention implantable cardioverter-defibrillators (ICD) occur in the context of multiple visits with multiple clinicians. Ongoing efforts to encourage shared decision-making need to align with actual decision-making paths and practices.

• When decision aids are generally given during electrophysiology consultations, patients may not heavily weight numerical, probabilistic data regarding risks and benefits in making an ICD decision.

• Heart failure clinicians or primary cardiologists are well-poised to play significant roles in facilitating shared decision-making about ICDs because of their longitudinal relationships and understanding of the trajectory of the patients’ illness.

Implantable cardioverter-defibrillators (ICD) are guideline recommended for primary prevention of sudden cardiac death in certain patients with heart failure with reduced ejection fraction (HFrEF).¹ Based on these guidelines, approximately 75,000 patients with HFrEF receive a primary prevention ICD in the United States annually.² However, patients with HFrEF have heterogenous values and preferences, which may not always align with ICD implantation. There is also evidence that patients’ understanding about the purpose of ICDs and risks of implantation are poor, resulting in downstream adverse outcomes.³ For instance, a small retrospective study showed that 52% of patients with active ICDs had a do not resuscitate order at the time of death, with some receiving shocks in the last 24 hours of life.⁴ This finding reflects either a failure to understand the purpose of the device in the context of the illness and goals of care or a lack of awareness of deactivation as an option. To address these issues and advance patient-centered care, the Centers for Medicare and Medicaid Services (CMS) now require that physicians conduct a shared decision-making interaction with patients referred for a primary prevention ICD.⁵ Shared decision-making provides an opportunity to improve patients’ engagement in ICD decision-making and enhance the extent to which their values align with the prescribed treatment plan.

The CMS shared decision-making mandate for ICDs focuses solely on the use of a patient decision aid.⁶ The increase in popularity of patient decision aids stems from data demonstrating that these tools enhance some components of shared decision-making.⁷ There are data demonstrating the potential positive impact of patient decision aids in the context of ICD decisions. However, there are multiple ways that decision aids may be integrated into a shared decision-making process, and a variety of different decision aids exist for ICDs. The only trial published to date involved the provision of multiple sources of information in advance of a visit with an electrophysiologist.⁸ Current practice in the context of the CMS mandate, however, likely involves providing the decision aid at the time of the encounter with the electrophysiologist, instead of beforehand, and there is mounting evidence that decision aid use alone is likely not a panacea for shared decision-making.⁹ In addition, different decision aids emphasize different types of information. This factor may explain why the CMS mandate has not demonstrated a major impact on shared decision-making metrics.¹⁰

The decision to implant a primary prevention ICD is complex. It necessitates balancing the patient-specific procedural risk with the benefit to patients based on their risk of arrhythmic death compared with nonarrhythmic death. In addition, patients’ values and preferences regarding end-of-life care and resuscitation—which may not be well-formed or clear—need to be integrated into the decision. The presence of a decision aid alone, which has been the focus of the CMS’s shared decision-making mandate, may not maximize the goals of shared decision-making.

A deeper understanding of patients’ ICD decision-making processes beyond decision aid use may help to ensure that shared decision-making does not become perfunctory. This process requires examining trajectories or paths of decisions, the roles different clinicians play, and how patients use information and arrive at decisions. This follow-up, qualitative key informant study was designed to provide these data.

Methods

Study Design and Population

We conducted a qualitative key informant study, which was nested within a previously published survey study comparing experiences of 101 patients who had an ICD implanted either before or after the CMS’s shared decision-making mandate.¹⁰ To deepen our understanding of why no major impact of the CMS’s shared decision-making mandate was observed, we used a mixed methods sequential explanatory design.¹¹ Key informant methodology involves selecting individuals with particular characteristics to probe the range of views among these individuals.¹² The primary objective of these in-depth, follow-up interviews was to explore drivers of patients’ decision-making about ICDs in greater depth and to identify potential refinements or interventions to improve the shared decision-making process. Participants for the key informant study were selected intentionally based on responses to the prior survey to represent varying levels of decisional conflict, health status, and demographics.

In general, 80%–90% of primary prevention ICD implantations at the participating institutions are performed as scheduled, outpatient procedures. During the postmandate period, the institutional practice for shared decision-making involved providing patients a paper copy of a decision aid developed at the University of Colorado⁶ during the preoperative electrophysiology visit (timing during the visit at clinicians’ discretion). The study was approved by the Emory Institutional Review Board. Confidentiality safeguards were affirmed, and completion of the survey was taken to indicate consent for participation. Verbal informed consent was obtained for all study participants, and participants were compensated for their time.

Sampling and Recruitment

Key informants were selected based on their responses to the prior survey and to represent key respondent types. After review of initial surveys by B.R. and E.A., participants were purposively sampled to ensure representation of different demographic characteristics, varying levels of decision conflict, health status, exposure to the decision aid, view on the usefulness of the decision aid, and feelings that they might choose differently about having an ICD implanted, if given the opportunity. This sampling strategy was used to capture the diversity of patients’ decision-making experiences and approaches. Exposure to the decision aid was determined by timing of the electrophysiology consult relative to the shared decision-making mandate enacted in February 2018. For patients who had seen an electrophysiologist after the CMS mandate, the initial survey asked their views on whether the decision aid was useful. Participants’ feelings that they would choose differently were determined by answers to the question on the initial survey, “If you had it to do over, would you decide to have an ICD implanted again?” A matrix of participants was developed to display these characteristics, and sampling continued until all cells in the matrix were represented and informational redundancy was achieved (Table 1).

Table 1.

Matrix of Participant Characteristics

Race (Gender)	Exposure to Decision Aid	Decision Conflict Score	Was the Decision Aid Helpful?	Health Status	Choose Similarly If You Could Do It Over?
White (F)	No	8	N/A	Fair	No
White (f)	No	28	N/A	Good	Yes
Black (M)	Yes	2	Yes	Good	Yes
White (M)	Yes	0	Yes	Very good	Yes
White (M)	Yes	0	Yes	Good	Yes
White (M)	Yes	0	Yes	Very good	No
White (F)	No	0	N/A	Very good	Yes
Black (F)	Yes	6	Yes	Fair	Yes
White (M)	Yes	12	N/A	Fair	Yes
Black (F)	Yes	4	Yes	Good	Yes
White (M)	Yes	0	Yes	Fair	No
Black (F)	Yes	0	Yes	Poor	Yes
Black (M)	Yes	0	Does not remember	Poor	Yes
Black (F)	No	14	N/A	Very good	Yes
Black (M)	Yes	10	Yes	Poor	Yes
White (M)	Yes	4	Does not remember	Poor	Yes
Black (M)	Yes	0	Yes	Good	Yes
Black (F)	Yes	20	Does not remember	Good’	No
Black (F)	Yes	N/A	No	N/A	N/A
Black (M)	Yes	N/A	Yes	N/A	N/A

N/A, not available.

Data Collection

In depth, semistructured interviews were conducted using an interview guide developed by the research team. The interview guide was designed by a multidisciplinary team with prior experience conducting qualitative studies and expertise in cardiology, shared decision-making, and bioethics to explore answers to survey questions in greater depth and to provide context for how participants made decisions regarding ICD implantation. The interview guide incorporated both open-ended questions and closed-ended questions with interactive probing of responses. Specific domains include patients’ ICD decision-making process, the path to make an ICD decision, the use of data in their decision, and areas where clinicians could better support their decision. Interviews were conducted by a trained qualitative interviewer (B.R.) via telephone from January 2021 through April 2021 and generally lasted between 20 and 30 minutes. All interviews were audio recorded and transcribed verbatim. Sampling continued until thematic saturation was achieved.

Qualitative Analysis

Interviews were transcribed verbatim, corrected for errors, and imported into MAXQDA (VERBI GmbH, Berlin, Germany) for analysis. Interviews were analyzed using a multilevel template coding strategy.¹³ Transcribed interviews were reviewed by B.R. and E.A. and corrected for typographical errors. An initial codebook was developed by the research team based on domains of the interview guide designed to explore drivers of ICD decision-making and then refined inductively as themes emerged during transcript review (constant comparison). Interviewers were coded by 2 separate coders (E.A. and C.B.), and discrepancies in coding were adjudicated by consensus by N.D.

Results

Population

Twenty interviews were completed. Six individuals did not respond to attempted contact by telephone (response rate of 77%). At the time of the interview, 2 participants had chosen not to have a primary prevention ICD implanted, and 18 had an active ICD. The sample was balanced in terms of race, sex, and health status (Table 2). Fourteen patients were aged more than 65 years. The indications for ICD implantation were predominantly HFrEF with a left ventricular ejection fraction of less than 35% (18/20). One participant had hypertrophic cardiomyopathy and another Brugada syndrome, both with high-risk features meeting the criteria for primary prevention ICD implantation. Three participants reported having experienced an ICD shock since implantation, and the remaining 15 with ICDs did not report any shocks. Five participants had their ICD implanted before CMS’s shared decision-making mandate; 13 were implanted after the mandate and, hence, received a decision aid.

Table 2.

Participants’ Demographic Data

	Overall (n = 20), n (%)
Age, years
<65	6 (30)
≥65	14 (70)
Sex
Female	9 (45)
Male	11 (55)
Race
Black or African American	11 (55)
White/Caucasian	9 (45)
Education*
Some high school	0 (0)
Graduated from high school	5 (25)
Some college	3 (15)
Graduated from college	4 (20)
Some graduate school	1 (5)
Graduated from a graduate school	6 (30)
Not reported	1 (5)
In general, would you say your health is†
Excellent	0 (0)
Very good	4 (20)
Good	6 (30)
Fair	4 (20)
Poor	4 (20)
Left ventricular ejection fraction
≤35%	18 (90)
>35%	2 (10)
Etiology of cardiomyopathy
Ischemic	6 (33)
Nonischemic	12 (66)
Has your ICD ever shocked you?
Yes	3 (15)
No	15 (75)
N/A (bever received ICD)	2 (10)

ICD, implantable cardioverter-defibrillator.

^*Missing values: 1.

^†Missing values: 2.

Key Domains

Patients’ Path to an ICD Decision: Multiple Visits With Multiple Clinicians.

Patients’ ICD implantation decisions rarely occurred during a single encounter. Eighteen of the 20 patients described a decision-making path that occurred over multiple visits, starting with learning about an ICD and its purpose and ending with a decision (Table 3).

Table 3.

Patients’ Path to ICD Decisions: Multiple Visits With Multiple Clinicians

Multiple visits

Subject 105: “I think it was about three to six months, I had a couple of visits with [general cardiologist] in that time. He told me about it more than once. The first couple of times he told me about it, like I said, I didn’t like being cut open and I didn’t know how—I didn’t know if it would really help or not… finally, after I just sat down, looked it up, did some research and tried to think about it logically.”

Multiple clinicians

Subject 71: “Initially, it was something that was brought up by my cardiologist. He gave me a brochure about the ICD… [and] we would talk about it from time to time when I would have my regular checkups. Then, he offered to make an appointment with one of the cardiologists who was an electrophysiologist to talk more about how it worked and what the procedure would be for implanting the defibrillator… Then…circling back and talking with [cardiologist name] I think there were a couple of factors that came to making the decision…”

ICD, implantable cardioverter-defibrillator.

In addition to deciding over an extended period of time, patients generally reported receiving input about the ICD from multiple clinicians and other sources of support. Often, the patients’ general cardiologist or heart failure specialist would first introduce the idea of an ICD and then refer the patient to an electrophysiologist, who provided additional procedural details. Notably, some patients had multiple visits with their cardiologist and a visit with the electrophysiologist before they made a decision regarding ICD implantation. The encounters with nonelectrophysiologists were often critical to decision-making (Table 3). Along with involving multiple clinicians, patients may involve friends or family members in their decision-making process. As stated by 1 patient “having my wife involved and discussing it with my children and other trusted friends, I think that’s important as well to get as many voices as possible as you trust.”

Electrophysiologists played a heterogenous role in these patients’ ICD decisions. For some patients, the substantive portion of the decision about whether to get an ICD implanted occurred before seeing the electrophysiologist, during discussions with their primary cardiologist. These patients had functionally made their ICD decision before the electrophysiologist consultation, who primarily clarified procedural details. However, other patients reported only cursory discussions about the ICD before meeting the electrophysiologist. In these cases, the decision-making process that occurred with the electrophysiologist was substantive, and patients described the electrophysiologist’s guidance as crucial. For 1 patient, it was only at the electrophysiology visit, after extensive discussion with his heart failure clinician, that the patient thought to ask, “Is this at all optional, or is this just something I just really need to do?” Although some stated that a general cardiologist had provided some written materials, the ICD decision aid was exclusively provided by electrophysiologists (Table 4).

Table 4.

ICD Decision Made Either Before or After Clinic Visit With the Electrophysiologist

Before	Interviewer: I see. Had you made up your mind that you were gonna go through with it before you saw the surgeon, or did you make up your mind after you saw the surgeon? Subject 73: I made up my mind before… When I talked to my heart doctor, I had already made up in my mind.
	Subject 100: …My heart failure doctor said I just needed to have it. I didn’t want to, but now at that point I was just scared enough I just thought I’d just let the doctors make the decision. I didn’t really pushback or anything. I hoped that I would improve and wouldn’t need it. He said it was the time to get it… They sent me over to the implant doctor, and I really didn’t question him. He said you’re here to get a defibrillator, let’s talk about it. He gave me the—well, his people gave me the paperwork to read to get oriented to what it would do and not do and some details what it would be like… Honestly, it wasn’t until I was actually leaving, and I just thought, I don’t know, it just kinda hit me to ask him. I said, “Is this at all optional, or is this just somethin’ I just really need to do?” He was very nice. He woulda talked to me more if I’d had questions I feel sure ‘cause he was an excellent doctor and very nice person. He said it’s the standard of care. My heart failure doctor had been so emphatic about it that I just accepted it at that point. Then they did it, and so here I am.
After	Subject 116: Yeah. After I talked to him and after I actually met with the electrophysiologist. That’s when I definitely made the decision and he explained to me more about what it was and the effect it was gonna have on me and the benefit, the pros, and the cons, he went over both of those. Gave me a little history behind the ICD itself, and the benefit and the negative about it.

ICD, implantable cardioverter-defibrillator.

Drivers of Patients’ ICD Decisions: Typically Not the Numbers.

Some patients engaged with the numeric, probabilistic data regarding risks and benefits presented in the decision aid to make decisions about device implantation. For instance, 1 patient reported the decision aid enhanced his understanding that the risk reduction with ICD therapy was not 100%. This patient stated his doctor said, “you need to have this…Then I saw the paperwork, and I was a little bit surprised… 6%…decrease in risk of death over the course of five years… that’s not really that big of a change.” This interaction later revealed that the patient was surprised at how high the baseline risk of death from heart failure can be, and, as a result, the patient agreed to ICD therapy because “it’s nice to have some improvement [in mortality].”

Another patient relied on quantitative data not in the decision aid: the list price of the device. This patient found the price of ICD implantation online and surmised that “Medicare would not have approved it if I really didn’t need [it] …because it’s an expensive procedure.” Here the patient equated price with value. “We’re offered a life-saving device at no cost…. Why not?”

For most patients, numeric data did not play a role in their ICD decisions; for these individuals, trust in their clinician’s recommendation drove their decision about implantation. These patients often had limited participation in the decision and trusted that their clinicians would recommend what was best for them. Patients reported multiple reasons to trust their clinician’s recommendation. First, patients cited a longitudinal nature of the relationship with their clinician (typically a general cardiologist or heart failure specialist) as an important factor. Second, patients described the clinician’s communication skills as a factor that engendered trust. Patients appreciated clinicians who took the time to go “step by step to explain to me exactly what was gonna happen,” and those who would “listen to you about you telling them what about your body and how you feel.” Of note, this type of trust was only peripherally connected to the substance of an ICD implantation decision. Third, patients trusted their clinician’s recommendation because they felt the clinician had expertise owing to the clinicians’ years of experience, position at the institution, or the prestige of the institution at which they practice.

Importantly, some patients made their decision based on a frank misunderstanding of the purpose of ICD therapy. For instance, 1 patient who was exposed to the decision aid reported that “When they put in the defibrillator, I’ve gained that little speed, my heart sped up a little bit. It was pumping a lot better, and I could breathe. Yeah, without it, I don’t think I would have made it.” This patient did not have any pacing requirement, nor did he have cardiac resynchronization therapy. His decision to undergo ICD implantation was simply made with an incorrect understanding of the function of the device.

The Role of Decision Aids in Patients’ ICD Decisions.

Patients’ reported use of the decision aid was heterogenous. Among patients who had received the decision aid, many reported it was rarely used during the encounter, although some referenced the document afterward. Patients who reported that the decision aid was used by the clinician during the encounter felt it was helpful because the clinician “actually went through it with me, where I could see it, to show you diagrams and stuff on how everything’s done.” Some patients wished they had the decision aid before meeting with the electrophysiologist. One patient stated, “because looking at the data and the graph was the most important thing…to decide, if I coulda had that earlier… I really would’ve liked to look over that.” Some patients reported that they “brought it home… and reviewed it at my leisure,” whereas others reported “I looked at some of it… I looked at the pictures and all the other stuff, and stuff like that.” As noted in 2 of these quotations, patients’ recollection of the decision aid information was sometimes more that it gave them details about the device itself rather than the numeric information regarding risk or benefit that it involves.

Patient Experiences Making an ICD Decision.

Almost all patients denied feeling pressured into their decision. When asking patients to reflect on whether they felt they could say no to ICD implantation, 19 of the 20 patients reported they did not feel pressured into their decision. The 1 patient who reported feeling pressured reported a scenario where the decision to implant an ICD was made as an inpatient while the patient was acutely ill. She felt she did not have sufficient time to understand the ramifications of her decision.

Some patients recommended improvements to the decision aid, and others requested support for the emotional struggles associated with their decision about ICD implantation. First, patients desired specific information addressing aspects of living with an ICD, such as aesthetic concerns and the sensation of an ICD shock, which is not highlighted in the paper decision aid. Some patients were surprised by the size, location, and bulkiness of the device. Second, patients recommended strategies for clinicians to improve communication during the encounter. Generally, these patients requested clinicians recognize “when you’re talkin’ about your heart it’s very scary.” Some patients’ reported ICD decisions may have been made without much thought because they were “scared enough [to]…just let the doctors make the decision.” Some patients described an emotional component to their decision focused on the anxieties about ICD implantation that patients noted were related to the procedure such as fear of “being put to sleep and not waking up” or fear of postprocedural recovery. Another patient struggled with the emotional significance of making decisions related to mortality. This patient “felt tortured about it” because he struggled with balancing his fear of procedures with his risk of cardiac arrest if he deferred ICD implantation.

Discussion

This qualitative, follow-up study provides deeper insight to patients’ decision-making processes and paths regarding ICD implantation, particularly in the context of the CMS’s recent mandate stipulating the use of a decision aid. These data reveal that ICD implantation decisions happen over time, with the involvement of multiple clinicians, and many of these decisions are made before a patient ever sees an electrophysiologist. These data also illustrate the important role of information other than probabilistic risk–benefit information for patients’ decisions. In both of these respects, these findings have important implications related to the current CMS decision aid requirement.

The most striking findings in this study were the extent to which decisions are made across time and the important role that nonelectrophysiologists play in shared decision-making regarding ICDs. Specifically, many patients have substantive discussions about ICD implantation long before any interaction with an electrophysiologist provides them with a decision aid. For these patients, a decision aid provided late in the process could provide confirmatory information or serve as a reference document moving forward, but it seems unlikely to impact actual decisions. Providing decision aids upstream from the electrophysiology consultation within the context of discussions by heart failure specialists, primary cardiologists, or primary care clinicians who may share strong therapeutic relationships with patients may improve shared decision-making. However, widening the number of physicians involved in the shared decision-making process, and the time period over which that process occurs, may pose challenges for documenting the shared decision-making interaction and the provision of the decision aid for the purpose of meeting CMS coverage requirements. Ultimately, maximizing any efforts to promote shared decision-making will require an implementation strategy that matches the clinical context in which ICD decisions occur.

These data also highlight the fact that patients make ICD decisions based on factors other than what is often highlighted in decision aids. Decision aids prominently feature probabilistic data related to benefits and risks. Yet, many patients rely on other types of information to make ICD decisions, like how living with an ICD will impact their day-to-day life. This may be reflective of the narrow use of decision aids during the electrophysiology consultation. Another explanation for patients de-emphasizing probabilistic information may be related to health numeracy.¹⁴ For instance, patients with lower health numeracy may value experiential information about the feeling of shocks or aesthetic information about the device after implantation as more valuable and even more reliable compared with probabilistic numerical data.¹⁵ These patients may strongly rely on the recommendation of family, friends, or their clinician, especially if it comes from a clinician with whom patients have an existing relationship. Conversely, even patients with high health numeracy may not heavily value probabilistic data related to risks and benefits of ICDs. Although decision aids mitigate numeracy issues by using info-graphics, some patients, either with low or high numeracy, may not strongly rely on survival data to make ICD decisions.

Fundamentally, the tradeoff in a decision about ICD implantation involves balancing patients’ desired aggressiveness in their medical care, quantity of life, and quality of life. Patients with ICDs functionally chose to live longer and to avoid a sudden death. However, this benefit comes at the risk of potential for ICD implant complications and for experiencing painful ICD shocks, both appropriate and inappropriate shocks. Shared decision-making provides a framework to engage patients in a discussion about aggressiveness and willingness to tolerate future complications. For some patients, aggressive escalations of medical care may result in a lower quality of life. However, few patients in our cohort seemed to truly address this tradeoff when making an ICD decision. The fact that some patients reported discomfort with discussing mortality is, in some sense, encouraging, because it means that they did engage in thinking about the mortality risks associated with heart failure. And it is known that communicating prognosis can be complex and uncomfortable in the context of heart failure; this continues to be a challenge when discussing heart failure treatment options.^16,17 Although discussing mortality may be difficult for patients, ICD decisions really are predicated upon patients’ understanding of the nature of their illness and the nature of the device. This is just another way in which patients’ general cardiologists or heart failure specialists may play an important role in facilitating shared decision-making interactions for ICD implantation. Patients’ primary cardiologists are well-suited to discuss their prognosis and the trajectory of their illness, which go hand in hand with eliciting values regarding mortality and quality of life. However, compared with electrophysiologists, general cardiologists and heart failure specialists may be less well-equipped to discuss the procedural risks and technical aspects of living with an ICD, both crucial elements for decision-making.

These data raise interesting challenges regarding policy-mandated shared decision-making. Although the CMS mandate of shared decision-making for primary prevention ICDs formally recognizes the need to integrate patients’ values and preferences into decisions for a guideline-recommended therapy, the mandate only stipulates the use of a decision aid. This tool may not be sufficient to substantively impact decisions, because the impact of a decision aid likely depends on the type of aid used, its implementation strategy, and the timing of its use. If this requirement is going to make ICD shared decision-making more robust at a population level, there needs to be a better understanding of how ICD decisions are actually made and implementation strategies that align with those realities. Most concretely, there needs to be recognition of the fact that many patients make ICD decisions over time, with multiple clinicians, and with different types of information.

This study has several limitations. First, the sample size is small. This sample size is appropriate for a qualitative key informant study, which was designed to provide a deeper understanding of how patients make decisions about primary prevention ICD implantation. However, these data do not provide information on prevalence of particular views. Second, there were no Latinx/Hispanic participants in this study, although there was good representation of White and Black patients. Third, the majority of patients in this study choose to have an ICD implanted. Only 2 patients were interviewed who did not have an ICD implanted. It is important to study the impact of shared decision-making processes for people who decide not to have an ICD implanted as well, although these patients can be difficult to identify because they do not see an electrophysiologist. Fourth, because this study was done on average 3 years after patients made their choice regarding ICD implantation, there is some potential for recall bias. Fifth, the decision aid used in this study was specifically cited in the CMS mandate, but other decision aids for ICDs may support shared decision-making in different ways. For instance, a video form of a decision aid for ICDs that was provided to patients before their clinic visits was found to be beneficial by patients.⁸ The presentation and timing of decision aid delivery may also impact shared decision-making.

Conclusions

These data reveal multiple ways in which substantive shared decision-making for ICD implantation must go beyond simple provision of a decision aid. Shared decision-making needs to be integrated effectively into the real patterns of patients’ care discussions, which often involve primary cardiologists and heart failure specialists. Effective shared decision-making also requires patients to understand their disease prognosis and consider their values about aggressiveness of care and quantity versus quality of life. Finally, it is important for patients to understand the experience of living with an ICD and not just probabilistic data about risks and benefits. Ultimately, the effectiveness of policy-mandated shared decision-making will depend on the implementation strategies that recognize these realities. Further studies evaluating different strategies will be important to ensure that this CMS mandate does not simply incentivize perfunctory interactions.