Development of a Content Valid Tool for Assessing End-of-Life Communication in Acute Care Settings

Introduction

Over 60% of patients that are admitted through the emergency department (ED) and die in-hospital present on a known dying trajectory, that is, their underlying medical conditions predicted significant health decline and death due to known medical conditions. ¹ The acute care physician (intensivists, emergency physicians, hospitalist-based internists) oftentimes must simultaneously stabilize patients while obtaining important information regarding medical diagnoses, current treatments, the patient's and family members' understanding of disease, and end-of-life goals of care. In most situations, the physician has little to no established relationship with the patient. The lack of a prior patient-physician relationship, combined with an acutely decompensating patient create significant time and communication challenges. These challenges must be overcome quickly so that emergent goals of care can be established. While there has been an increase in end-of-life curricula in undergraduate and graduate medical education, current guidelines do not adequately reflect the complexities of physician-family interactions that occur in actively dying patients. ² As a result, many times physicians do not accurately assess patient and family wishes about end-of-life care and often misinterpret the desire for life-sustaining measures. ³ This can lead to the initiation of futile medical treatment, increased family member stress, and failure to transition to more optimal hospice care settings.^4-7

This manuscript describes the development of a content valid instrument to measure the end-of-life communication skills of emergency medicine physicians. As such, it represents a crucial first step toward the construct validation of a comprehensive assessment tool geared toward measurement of end-of-life care provision in the acute care setting.

Methods

We conducted an extensive review of the literature to ascertain the knowledge, skills, and attitudes (KSAs) necessary for effective end-of-life family discussions. These KSAs were reviewed in light of the need for rapid assessment and decision-making characteristic of acute care settings, and were divided into five separate domains: (1) seek information, (2) assess life values, (3) educate family, (4) extend care in a consistent manner, and (5) respond to family questions and concerns. Known as SILVER, we posited that this acronym would help form a framework that could be used to guide teaching and assessment of effective end-of-life communication in acute care settings.^8,9

We further defined the five SILVER domains through a modified Delphi process that elicited examples of questions and comments representative of these domains. We identified a group of 18 participants with expertise in the field of palliative emergency medicine to participate in this process. Specifically, participants were either board certified emergency physicians with a hospice and palliative medicine subspecialty, or licensed emergency nurse practitioners with a publication record in communication issues relevant to palliative care. Both the New Jersey Medical School and Wayne State University institutional review boards approved this study.

The Delphi process involved the development of a web-based questionnaire, along with educational materials explaining the purpose and procedures of the study. The educational materials described the primary construct of the SILVER framework and provided supporting references taken from the communication and emergency medicine bodies of literature, e.g., the 4 Habits Approach to Effective Clinical Communication.^9,10 Participants were asked to provide anonymous examples of questions and comments that would be indicative of each SILVER domain. These comments were then aggregated into one document and an initial thematic analysis was independently conducted by each of the authors and later refined via consensus. This process identified one or more SILVER subdomains for each SILVER domain.

The document of aggregated, unaltered comments obtained from our content expert participants was then returned to the participants. Via an electronic survey, each participant was asked to identify the subdomain that he or she believed the comment was addressing, and simultaneously rate the relevance of the comment on a 0 (poor) to 10 (excellent) scale. We chose, a priori, to identify comments that were rated by at least half of the participants in a particular subdomain, and that also received a mean rating of at least 7. These comments became exemplars for use in describing the subdomains. The final communication assessment rubric (Figure 1) compiled the SILVER domains with appropriate subdomains as rating elements.

OPEN IN VIEWER

FIG. 1.

End-of-Life Assessment Tool for Emergency Medicine Practitioners

Results

Overall, our subject matter experts provided 104 unique questions and comments in the initial round of the Delphi process: 28 regarding “seeking information,” 41 regarding “addressing life values,” and 35 for “educating, extending care, and responding.” After independent and, later, consensus review of the comments, we identified four subdomains for each of the elements of “seeking information” and “addressing life values.” In contrast, the elements “educate,” “extend care,” and “respond” were each characterized by a single descriptor or subdomain. These subdomains, shown in Table 1, were rated for importance by each of the subject matter experts on a scale of 0 to 10 (“not at all” to “extremely”), with each receiving a mean score of greater than 7.

In round 2 of the Delphi process, the subject matter experts were asked to rate the relevance of each question or comment, and to apply the rating within one or more of the subdomains, yielding the results shown in Table 2. The results of this round demonstrate that a subject matter expert could reasonably be expected to identify comments and questions within each SILVER domain and subdomain. As described above, only questions and comments where at least half the experts rated it on average as a 7 or above were selected as exemplars to describe the subdomains.

Discussion

The objective of this study was to create a content valid physician assessment tool for end-of-life care in an emergency care setting. The authors used five domains within the SILVER mnemonic to structure a three-phase modified Delphi process that (1) further characterized and defined the SILVER domains, and (2) elicited examples of questions and comments representative of these domains. The result is an assessment rubric with content validity to measure physicians' communication in end-of-life discussions regarding acutely decompensating patients. While the experts participating in this study represented emergency medicine practitioners with substantive palliative care expertise, we believe the skill set targeted would be applicable to any practitioner treating an unfamiliar, critically ill patient with significant comorbidities.

Several guidelines exist to assist physicians caring for dying patients.^2,8,12 For example, Knops and Lamba outline mnemonics described in the literature and put forth their own communication tool (ASCEND). ¹³ While these tools contain useful, evidence-supported techniques for communicating with patients and families, they often make several assumptions that limit their usability in ED-based settings.^2,8,12 First, there is a basic assumption that significant time can be devoted to the end-of-life discussion. While this may be true in clinic and inpatient settings, it is often not the case in the ED. Frequently the decision whether or not to treat a patient aggressively and utilize life-sustaining treatment is required within minutes of a patient's arrival. Recommendations that physicians “plan” and “rehearse” the disclosure of bad news are not feasible. ⁸

A second assumption made by most guidelines is that the physician is the most important source of information to guide patient care. While a number of reports in the medical literature address the importance of family participation in end-of-life conversations, guidelines still focus on the delivery of bad news. This is impractical in the ED, as there is usually an emergent need to obtain information from a patient's family. This shift in focus from delivering information to obtaining critical data requires advanced communication skills and interpersonal approaches that are not adequately captured within current guidelines.

Finally, most guidelines assume the patient is an active participant. While the importance of family and companion presence is noted in the literature, the goal of end-of-life discussions is geared toward the patient making an informed decision, with or without family input. ¹⁴ In contrast, ED physicians must engage family in a decision-making process for which the family members often feel ill-prepared. ED physicians must assess the family members' knowledge of the patient's prognosis and end-of-life choices. Since the majority of patients with chronic illnesses do not discuss terminal diagnoses and life-sustaining treatments with their primary physicians, this information is often unknown.^4,15–17 As a result, the ED physician is challenged to assist families in making an informed decision at a time of heightened stress.

The SILVER mnemonic with its associated subdomains offers a focused assessment tool grounded in observable, measurable communication. As such, it can provide a foundation for training that focuses on practical knowledge and skills that effectively translate to the ED or other acute care settings. Each domain and subdomain identified can be linked to previously defined clinical skill sets defined in the literature. For example, Seeking Information likely contains similar behaviors described in the SPIKES “perception” domain, focused on asking the patient what she or he knows. ⁸ However, unlike SPIKES, the SILVER approach acknowledges the limited preparatory time available to set up an interview, and considers the likelihood that information will be provided by a surrogate rather than the patient. The anchoring of each domain to clear behaviors (Table 2) will help guide training for learner assessments. Additionally, the clearly defined subdomains will offer substantive structure for observation-based assessments while allowing the flexibility needed to capture the variable performance present in advanced learners. ¹⁸

As an educational tool, the SILVER mnemonic provides focused guidelines that a practitioner at any stage or specialty of training could use when treating an unfamiliar, acutely decompensating patient with need for an end-of-life care plan. The discreet domains provide the opportunity for modular training focused on specific portions of the SILVER methodology. Additionally, the expert-derived statements and behaviors offer concrete examples that can be adopted by learners while the subdomains provide clear yet flexible guidelines that allow practitioners to tailor their behaviors for particular situations.

SILVER does not attempt to capture more affect-driven skills described elsewhere, e.g., empathy.^8,19 As in ASCEND, an assumption is made that the practitioner possesses a baseline level of sensitivity and empathy. ¹³ These are important skill sets that have been shown to impact patient care and establishment of patient care goals in critically ill patients. However, measures of basic communication skills have been described in the health care education literature. ^{20 21 22} What remains unclear is the appropriate “baseline” effort given these skills during critical information exchange and decision making.

This study represents a first step toward developing end-of-life care assessments that can easily translate to the emergency (acute) care setting. The goal of this work is to lay a foundation for curricular development and assessment design that will allow the further testing and experimental validation of the SILVER approach. A well-studied, valid measure of provider performance in end-of-life care discussions focused on the ED setting will allow researchers to further probe the value and impact ED-based palliative care has on patient outcomes and patient resource utilization.