Intern Emerg Med DOI 10.1007/s11739-016-1405-y

EM - ORIGINAL

Social worker assessment of bad news delivery by emergency medicine residents: a novel direct-observation milestone assessment Alice Ann Min1 • Karen Spear-Ellinwood2,3 • Melissa Berman4 • Peyton Nisson5 Suzanne Michelle Rhodes1,6,7

•

Received: 9 October 2015 / Accepted: 4 February 2016  SIMI 2016

Abstract The skill of delivering bad news is difficult to teach and evaluate. Residents may practice in simulated settings; however, this may not translate to confidence or competence during real experiences. We investigated the acceptability and feasibility of social workers as evaluators of residents’ delivery of bad news during patient encounters, and assessed the attitudes of both groups regarding this process. From August 2013 to June 2014, emergency medicine residents completed self-assessments after delivering bad news. Social workers completed evaluations after observing these conversations. The Assessment tools were designed by modifying the global Breaking Bad News Assessment Scale. Residents and social workers completed post-study surveys. 37 evaluations were received, 20 completed by social workers and 17 resident self& Alice Ann Min [email protected] 1

Department of Emergency Medicine, College of Medicine, The University of Arizona, P.O. Box 245057, Tucson, AZ 85724-5057, USA

2

Department of Obstetrics and Gynecology, College of Medicine, The University of Arizona, Tucson, AZ, USA

3

Faculty Instructional Development, Office of Medical Student Education, College of Medicine, The University of Arizona, Tucson, AZ, USA

4

Department of Clinical Resource Management, University of Arizona Medical Center, Tucson, AZ, USA

5

Mel and Enid Zuckerman College of Public Health, The University of Arizona, Tucson, AZ, USA

6

Division of Geriatrics, General Medicine, and Palliative Medicine, College of Medicine, The University of Arizona, Tucson, AZ, USA

7

Arizona Center on Aging, The University of Arizona, Tucson, AZ, USA

evaluations. Social workers reported discussing plans with residents prior to conversations 90 % of the time (18/20, 95 % CI 64.5, 97.8). Social workers who had previously observed the resident delivering bad news reported that the resident was more skilled on subsequent encounters 90 % of the time (95 % CI 42.2, 99). Both social workers and residents felt that prior training or experience was important. First-year residents valued advice from social workers less than advice from attending physicians, whereas more experienced residents perceived advice from social workers to be equivalent with that of attending physicians (40 versus 2.9 %, p = 0.002). Social worker assessment of residents’ abilities to deliver bad news is feasible and acceptable to both groups. This formalized self-assessment and evaluation process highlights the importance of social workers’ involvement in delivery of bad news, and the teaching of this skill. This method may also be used as direct-observation for resident milestone assessment. Keywords Communication
Bad news delivery
Resident education
Interprofessional team
Survey
Milestone assessment

Introduction The emergency department (ED) is often the setting for the delivery of bad news. This can come in many forms, and result in varying emotions and reactions. By convention, ‘‘any news that drastically and negatively alters the patient’s view of her or his future’’ is considered bad news [1], though death notification is often considered the most life changing for loved ones and the most challenging for physicians. The National Hospital Ambulatory Medical Care Survey in 2010 estimates there are 240,000 deaths in US EDs [2].

123

Intern Emerg Med

Surveys of survivors of sudden death in the ED report ED staff to be ‘‘cold, unsympathetic, and not reassuring,’’ or have negative experiences with death notification [3–6]. Many reports describe family members years later struggling or occupied with the manner in which they were given bad news by a health care provider, which raises concerns for a possible contribution to complicated grief [5–7]. Social workers have been a vital part of many published approaches to bad news delivery in the ED [6, 8–11]. The social worker is often tasked with identifying next of kin and making the initial request that family report to the hospital [8]. The role has been described as ‘‘a psychological bridge’’ to the rest of the medical team [9], an advocate to ensure individual questions and concerns are addressed, and to encourage and support the grief response, as well as to provide additional information and community resources [5, 8]. Social workers have also been identified as ‘‘experts’’ in communication, able to provide compassion, education, and emotional support to those in need—a key part of their education and training. The medical social work task of facilitating and verifying the understanding of information exchange in situations of death and dying between social work and the interdisciplinary team has long been recognized [12–16]. Social workers are a key component of the team delivering bad news, and assisting with the bereavement and future planning for loved ones, with several published protocols involving their expertise [8, 17]. In addition to development of protocols for delivery, medical education on the delivery of bad news using a variety of techniques has been published [18–26]. For the most part, ‘‘see one, do one,’’ remains the prevailing method in training early physicians [21]. Simulation training is one of the most commonly reported in the literature. One large concern with simulation training is that while research has suggested that students benefit from such training, when faced with real-life situations, they often feel unprepared [20]. A large, multi-site, randomized trial using simulation-based training during a four day workshop to teach the delivery of bad news fails to show improvements in the quality of communication [20]. The authors of that study describe several factors that might have influenced the outcome, including family and patients being unskilled evaluators, the validity of the questionnaire, and the length of time from assessment [20]. It is also possible that trainees are not continuing to receive direct assessment and feedback once the skill is moved from simulation to a real patient encounter [20]. Another challenge in medical education is the assessment of skill and provision of feedback. In terms of physician performance in delivering bad news, ‘‘[t]he final arbiter is the patient and patient outcomes’’ [27].

123

Unfortunately, barriers to direct patient or family feedback exist such as the burden of asking a grieving family member to assess the tragic news just received, as well as the lack of feasibility in training family members or loved ones on quality evaluations. The Accreditation Council for Graduate Medical Education (ACGME) has implemented a new milestones method of resident assessment. In many circumstances, direct observation of learners to assess competency is required. For emergency medicine, two sub-competencies incorporate the evaluation of communication and empathy [28]. The sub-competency of professional values asks that the resident ‘‘demonstrate compassion, integrity, respect, sensitivity, and responsiveness…consistently in common/ uncomplicated situations and with diverse populations’’ [28]. The sub-competency of patient-centered communication assesses interpersonal and communication skills, and explicitly includes the use of ‘‘flexible communication strategies…to resolve specific ED challenges, such as…delivering bad news’’ [28]. The ACGME encourages multi-source feedback to evaluate the residents [28]. Social worker assessment of the direct observation of these communication skills will be valuable to the global evaluation of the clinical performance of emergency medicine residents. The principle of deliberate practice, as described by Ericsson, offers a scenario for attaining expert performance by defining the goal, building motivation for improvement, applying reliable measures of performance, and offering feedback and opportunities for refinement of performance [29]. This framework has been used previously in medical education to improve communication skills among residents and fellows [30, 31]. Toward this end, we engaged the expertise of social workers, trained in communication and sensitive to patients’ and families’ needs, to evaluate residents’ delivery of bad news to actual patients and their families. Our objective was to investigate whether social workers could serve as effective evaluators and instructors of emergency medicine residents in delivering bad news.

Methods Study design and population This study was a feasibility and acceptability trial relying on survey data from social workers’ evaluations and residents’ self-assessment of residents’ skill in the delivery of bad news in the ED. The population studied in an academic medical center was 59 Emergency Medicine (EM) residents in a 3-year residency program as well as those in a 5-year combined EM/Pediatrics program, and seven

Intern Emerg Med

Emergency Department-based clinical social workers. The study was reviewed by the Institutional Review Board and classified as exempt. Study protocol A self-assessment and a social work assessment were designed by modifying the global Breaking Bad News Assessment Scale (mgBAS) [22]. The global BAS is a fiveitem scale scored using a Likert Scale of 1–5 (1 very good, 2 good, 3 medium, 4 poor, 5 very poor) that was developed and validated based upon the work of Miller et al. [22, 32]. The five themes identified by Schildman et al. were expanded with a brief explanation of the types of behaviors that might be expected from the work of Miller and Schildman [22, 32]. The instrument was reviewed by content experts, residents, and social workers to ensure the questions were clear and comprehensive, while avoiding too burdensome an assessment during clinical work. Revisions were made according to their recommendations, and reviewed repeatedly until no further recommendations were suggested. Social workers at our institution routinely are present during, and provide psychosocial support to physicians as well as patients and their families following physicians’ delivery of bad news in the ED. A core group of seven social workers were asked to participate. They were provided the assessment tool, and instructed on its use. All questions from the residents and social workers about the assessment tools were addressed prior to gathering data. The social workers used the tool to assess emergency medicine residents immediately following the delivery of bad news. Residents completed a self-assessment each time they delivered anything they considered to be bad news. Participants placed these evaluations in a sealed envelope in a locked box in the ED. Approximately once every 3 months, residents and social workers were reminded via email of the study, and to asked to complete the forms. Survey content and administration Given the lack of a prior study of this nature, the investigators designed the post-study survey to explore resident use of the self-assessments and social workers as evaluators of residents. Instruments asked respondents to indicate gender, age, training level, prior experience with delivery of bad news, and whether they had completed assessment forms. Questions were designed to reveal the acceptability, comfort level, reasons for non-participation, and any changes to practice resulting from the assessment. Items consisted of Likert scales, open-ended questions and free text responses. Following the development of the initial draft of the survey, content experts were asked to review for clarity and content. The questionnaire was pilot tested

with potential survey recipients to determine readability, and to identify confusing or leading questions. Items were then edited, moved, or clarified until no further recommendations were made for improvement. The final surveys consisted of 16 questions for residents (five free text) and 16 for social workers (eight free text). At the end of the academic year, all residents and social workers were emailed the final survey. A printed copy of the survey was also provided to residents not responding to the initial email. Additional periodic emails and personal reminders were issued. Data analysis Descriptive statistics including counts, percentages, mean and standard deviation or median and interquartile range for non-normal data were provided. A mean mgBAS score was calculated from each of five scores. Comparison of the mean mgBAS by groups (resident gender, patient age categories, patient gender, Caucasian versus other, Englishspeaking versus non-English speaking, and death notification versus other) for resident and social work evaluations was performed using Wilcoxon Rank-Sum or Kruskal– Wallis due to small sample size and non-normally distributed data. Paired assessments were compared with kappa using linear weighting. Demographic and educational features are compared using v2 for proportions (or Fischer’s Exact in the event of subsets with small counts). A p value of 0.05 is considered significant. We analyzed free-text responses from the evaluations and surveys using grounded theory [33]. Three reviewers (A.M., S.M.R., K.S.E.) independently read responses to each question to identify common themes. Each reviewer then re-read to identify additional themes, followed by a subsequent review to examine for differences in themes by category (gender, resident age, patient age, type of disclosure, prior training in breaking bad news, number of bad news deliveries in last year, and training year). The reviewers then met to discuss independent interpretations, identify disagreements, and come to a consensus.

Results Assessments The demographic characteristics of residents and patients are provided in Table 1. There were 37 evaluations total with 20 completed by 7 social workers and 17 self-evaluations by 10 residents. One social worker completed 9 assessments, one completed 4 assessments, another completed 3, and the remaining four social workers completed one evaluation each. Residents reported discussing a plan

123

Intern Emerg Med Table 1 Evaluations of bad news delivery: resident and patient demographic factors SW (N = 20)

Resident (N = 17)

1

4 (20)

2 (11.8)

2 3

7 (35) 7 (35)

5 (29.4) 6 (35.3)

4

0

0

5

2 (10)

4 (23.5)

Male

10 (50)

12 (70.6)

Female

10 (50)

5 (29.4)

Patient age, median (IQR)

45 (23.5–61.5)

55 (20–67)

Resident year, N (%)

Table 2 Mean global Bad News Assessment Scale by resident and patient features with non-parametric statistical comparison SW (N = 20)

Resident (N = 17)

1.3 (1, 2)

2.2 (1.8, 2.8)

Intern Junior

1.8 (4) 1.8 (7)

2.5 (2) 1.8 (5)

Senior

1 (9)

2.1 (10)

Male

1.4 (10)

2.2 (5)

Female

1.2 (10)

2 (12)

Median (IQR)

Resident sex, N (%)

Patient sex, N (%)

Resident year, median mgBAS (N)

Resident sex, median mgBAS (N)

Patient age, median mgBAS (N) 0–18 years

1.4 (4)

1.8 (4)

18–64 years

1.5 (12)

2.8 (8)

1 (4)

2.2 (5)

Male

1.6 (13)

2 (12)

Female

1.4 (5)

Missing

1 (2)

[65 years

Male

13 (72.2)

12 (70.6)

Female

5 (27.8)

5 (29.4)

Missing

2 (10)

–

White

9 (45)

5 (29.4)

Hispanic

9 (45)

7 (41.2)

Black or African American Missing

1 (5) 1 (5)

4 (23.5) 1 (5.9)

English

15 (75)

14 (82.4)

Spanish

1 (5)

1 (5.9)

English

Other

–

1 (5.9)

Spanish or other

Missing

4 (20)

1 (5.9)

Race/ethnicity, N (%)

Preferred language, N (%)

Patient sex, median mgBAS (N)

Race/ethnicity, median mgBAS (N) White Hispanic or African American

1 (9)* 1.9 (10)*

2 (5) 2.2 (11)

1 (1)

3.4 (1)

Missing

Notification type, N (%)

2.2 (5) –

Preferred language, median mgBAS (N)

Missing

1.2 (15)

2.0 (14) 

2 (1)

3.3 (2) 

1.5 (4)

3 (1)

Notification type, median mgBAS (N)

Death

6 (30)

9 (52.9)

Death

1.4 (6)

2.0 (9)

New life-limiting diagnosis

9 (45)

3 (17.7)

Other

1.3 (14)

2.2 (8)

Withdrawal of support

2 (10)

1 (5.9)

Post-cardiac arrest with ROSC

2 (10)

2 (11.8)

SW social work, N number, IQR interquartile range, mgBAS global Bad News Assessment Scale, ROSC return of spontaneous circulation

Other

1 (5)

2 (11.8)

* p = 0.03;

 

p = 0.02

SW social work, N number, IQR interquartile range, ROSC return of spontaneous circulation

Post assessment survey for the delivery prior to the conversation with an attending physician 82.4 % of the time (14/17; 95 % CI 53.7, 94.9), and social workers reported discussing a plan with the resident 90 % of the time (18/20, 95 % CI 64.5, 97.8). Social workers who had previously observed the resident delivering bad news (N = 10) reported that the resident demonstrated greater skill than in the prior encounter 90 % of the time (95 % CI 42.2, 99). There were six conversations in which residents and social workers evaluated the same event (12 actual evaluations). The social worker gave a higher (more favorable) score in all but one evaluation in which the resident rated the same total score. Agreement was measured with a kappa of 0.083 (95 % CI 0, 0.35). The total mean score for the mgBAS is provided in Table 2 by resident and patient features.

123

The response rate for social workers was 100 % (responses, Table 3). Response rate for residents was 83 % (responses, Table 4). Responses are further broken down by resident characteristics in Table 5 for key survey questions. Some residents frequently did not complete an evaluation (41.7 %), stating this was because they simply forgot to do the self-assessment. Qualitative Free text comments were provided by social workers 86 % of the time (6/7) for the survey and 85 % (17/20) for the evaluations of residents. Residents included comments in 72 % (21/29) of surveys and 76 % (13/17) of their self-

Intern Emerg Med Table 3 Social worker survey responses post assessment year Gender, N (%)

Table 4 Resident survey responses following year of social work evaluation (N = 49)

0 (0)

Survey online, N (%)

7 (100)

Age, median (IQR)

29 (28–31)

Years in position, mean (SD)

5.36 (2.93)

Female sex, N (%)

21 (42.9)

Years of experience, mean (SD)

15.7 (4.11)

PGY level, N (%)

Male Female

1

15 (30.6)

7 (100)

2

15 (30.6)

0 (0)

3

18 (36.7)

Do you feel this evaluation interfered with your work duties? N (%) No Yes

29 (59.2)

Were you more or less likely to give feedback to the residents after this evaluation process? N (%) Extremely unlikely

0 (0)

Unlikely Neutral

0 (0) 1 (14.29)

Likely

6 (85.71)

Extremely likely

0 (0)

Did you feel the residents were receptive to your feedback? N (%) Very unreceptive

0 (0)

Unreceptive

0 (0)

Neutral

1 (14.29)

Receptive

4 (57.14)

Very receptive

2 (28.57)

Did you feel comfortable evaluating the residents? N (%) Very uncomfortable

0 (0)

Uncomfortable

0 (0)

Neutral

0 (0)

Comfortable

3 (42.86)

Very comfortable

4 (57.14)

N number, SD standard deviation

evaluations. Three themes emerged in the qualitative analysis of social worker and resident perceptions of the process of delivering bad news, identifying (1) what is important to delivering bad news effectively; (2) what is challenging about delivering bad news; and (3) what is helpful in the current training process (Fig. 1). Social workers and residents felt that preparation and some type of training or previous experience was important before having a difficult conversation. During the conversation, it was helpful to have another professional there to help with the process, and to communicate in a direct yet gentle manner. Feedback after the conversation occurred was also felt to be important. A common challenge identified when breaking bad news was managing emotion, displayed by the patient and family as well as those feelings experienced by the residents. Resident self-evaluators were more likely to rate themselves poorly (greater than the mean of 2.2) when they commented upon a challenging situation, such as a strong emotional response from family, familial anger (66 versus 36.4 %), or having to deliver the news by phone. Social workers’ suggestions addressed similar themes—that

4 Estimate the number of times you have led a disclosure of bad news discussion in the last year median (IQR)

1 (2.0) 4 (3–10)

Did you receive training as a medical student in delivery of bad news? Yes

38 (77.6)

No

11 (25.5)

Did you complete a self-assessment (the green forms) evaluating your delivery of bad news in the ED this year? Yes

13 (26.5)

No

36 (73.5)

Did the self-assessment lead you to discuss the case with the social worker after the delivery of bad news? (N = 47) Never

3 (6.4)

Rarely

1 (2.1)

Sometimes

7 (14.9)

Often

5 (10.6)

Always Not applicable

5 (10.6) 26 (55.3)

How influential do you think you think the assessments by the social workers were in your approach to the delivery of bad news? Not at all influential

2 (4.2)

Slightly influential Somewhat influential

4 (8.3) 3 (6.3)

Very influential Extremely influential Not applicable

14 (29.2) 5 (10.4) 20 (14.7)

Do you see social workers as the appropriate medical professional to evaluate your communication skills? N (%) Yes

45 (95.7)

No

2 (4.3)

Do you value the advice of a social worker more than an attending physician with respect to delivery of bad news? N (%) Much less

1 (2.0)

Somewhat less

6 (12.2)

About the same

22 (44.9)

Somewhat better

14 (28.6)

Much better

6 (12.2)

Do you talk to anyone to help prepare before you deliver the bad news? N (%) Attending

37 (75.5)

Social worker

35 (71.4)

Other resident

7 (14.3)

N number, IQR interquartile range, mgBAS global Bad News Assessment Scale, ROSC return of spontaneous circulation

123

Intern Emerg Med Table 5 Resident survey responses by sex, intern status, prior training, and level of reported experience Survey item

Sex M

PGY F

Training

1

2–5

Value advice of SW about the same, somewhat or much better than attending

82 (23/28)

90 (19/21)

60 (9/15)*

Social workers are appropriate medical professional to evaluate communication skills

93 (25/27)

100 (20/20)

86 (12/14)

SW assessments somewhat, very or extremely influential in approach to BBN

75 (15/20)

88 (7/8)

50 (5/10) 

97 (33/34)*

100 (33/33)

94 (17/18) 

N

Experience Y

0–4

C5

90 (10/11)

84 (32/38)

77 (20/26)

96 (22/23)

100 (10/10)

95 (35/37)

96 (25/26)

95 (20/21)

100 (6/6)

73 (66/22)

64 (9/14)

93 (13/14)

Spoke with SW to prepare

71 (20/28)

71 (15/21)

53 (8/15)

79 (27/34)

82 (9/11)

68 (26/38)

65 (17/26)

78 (18/23)

Spoke with attending to prepare

68 (19/28)

86 (18/21)

87 (13/15)

71 (24/34)

82 (9/11)

74 (28/38)

73 (19/26)

78 (18/23)

Spoke with resident to prepare

4 (1/27)à

29 (6/21)à

20 (3/15)

12 (4/34)

27 (3/11)

11 (4/38)

15 (4/26)

13 (3/23)

M male, F female, PGY year of residency training, N no, Y yes, SW social work, BBN breaking bad news * Fisher’s p = 0.002   à

Fisher’s p = 0.013 Fisher’s p = 0.019

delivering bad news is emotional, uncomfortable and filled with variables, not all of which can be anticipated. In addition, social workers proposed what would be helpful to residents in learning this skill, ideas that aligned well with what is important before, during, and after delivering bad news. Social workers suggested ongoing training and periodic evaluations, which corresponded well to the components of training that resonated with residents (case-based practice, being accompanied and evaluated by health care professionals with expertise in communication skills, receiving constructive feedback on performance). These also aligned with Ericsson’s framework for deliberate practice and the teaching of communication skills. Common recommendations made by social workers were to be direct, but compassionate with families and to avoid giving additional information to ‘‘soften the blow.’’ Analysis segmented resident participants by those who received prior training in delivering bad news while in medical school, as well as by year in residency (first year residents and all others). Comparing resident respondents who had not received prior training as medical students (N = 11) to those who had some training (N = 38), 4 of those with no prior training cited the need for additional focus on the topic in the residency curriculum and simulation cases. In contrast, 7 of the 38 with prior training commented on valuing the ‘‘real thing,’’ learning by delivering news to real patients, as opposed to simulation. First year residents reported valuing the advice of a social worker much less or somewhat less than the advice of an attending physician, whereas more experienced residents perceived advice from social workers to be on par with that of attending physicians (40 versus 2.9 %, p = 0.002).

123

Discussion Assessment of resident performance in delivering bad news by social workers paired with resident self-assessments was feasible and acceptable to emergency medicine residents and social workers. Our approach to assessing the skills of our residents is focused on the expertise of the clinical social workers in the ED. The evaluation process by the social workers of the residents provides opportunities for formative feedback recognized by a majority of participants in this study as influential to the development of this essential communication skill. Our study, unlike the majority of published reports on teaching bad news delivery in medical education [22, 23, 25, 26, 31, 34–36], is focused on actual patient encounters as they occurred in real-time. Whether a simulation-based practice discussion or the ‘‘real thing,’’ most residents report that this skill can really be honed only by repeating the experience, that is, with practice. At this time in medical education, many medical schools have included didactics on communication skills and delivery of bad news. Among our residents, 77 % had some exposure as a medical student in addition to the formal lecture they received as emergency medicine residents. It is our hope that this assessment represents a possible ‘‘missing link’’ between simulation and classroom learning, and improves real-world communication skills for this difficult task. The idea of deliberate practice to achieve expert performance has been popularized by Ericsson, and has been tested with respect to communication by Wouda et al. [29, 37]. It is our hope that social work assessment and selfassessment with the mgBAS set a clear goal of quality

Intern Emerg Med

2. What’s Challenging 1. What’s important RESIDENTS and SOCIAL WORKERS agree preparaon is important Preparaon through training and real experience Assessing emotion or needs of the patient or family before or during the encounter

RESIDENTS and SOCIAL WORKERS agree… Delivering bad news involves a “significant amount of emotion”, making it uncomfortable to deliver the news, appropriately address family or patient reactions, or control one’s emotions Physicians lack specialized training in this type of communication Other factors affect delivery, such as: o Physician work load o Variable (unpredictable) experience o Limited opportunies for pracce

Relates back to

4. What would be helpful SOCIAL WORKERS suggest Residents should receive guidance on how to be direct, while demonstrating compassion Residents should focus on: o Facts families need to know o HOW bad news is delivered Residents should participate in: o Ongoing training o Regular evaluations with notice of evaluation o Regular pre- and postconversations between the resident and social workerobserver/evaluator (include attendings, when available) o Periodic discussion groups for residents to address emergent issues regarding delivering bad news

3. What’s helpful RESIDENTS and SOCIAL WORKERS agree… BEFORE Encounter Training through simulaon and casebased pracce (e.g., role play)

DURING Encounter A health care professional with experse this kind of communicaon (social worker) in the room or able to “step in”, as needed, when residents deliver bad news

AFTER Encounter Construcve (“direct” but “gentle”) feedback on simulated or real performance by health care professional (closer to paent’s perspecve)

Fig. 1 Alignment of residents’ and social workers’ perspectives on delivering bad news

communication, provide opportunity and motivation for immediate reflection and feedback on the event, as well as plans for future improvements in order to create the setting of deliberate practice in the skill of delivering bad news in a way that represents the least interference and burden on patients, family, clinical staff and residents. Resident participation in self-assessments as well as the frequency of meetings with social workers indicates they are motivated learners of this necessary skill [38]. They regard the process as feasible, and are amenable to consulting with social workers in planning how to deliver bad news. In fact,

residents seek out social workers and attendings for advice to prepare for delivery of bad news 71 and 76 % of the time, respectively. The introduction of this process, training and periodic reminders to participate highlighted the need for focused, repetitive practice of this skill. The emphasis on assessment and evaluation and the use of reliable measurements offered residents opportunities for formative feedback and social worker evaluators a consistent method of evaluation and factual foundation for providing feedback. Discussing a plan for the delivery of bad news prior to the actual

123

Intern Emerg Med

patient and family interaction is a practice that is encouraged by our ED. It is unknown if this study has an effect on these rates. It does appear to increase the likelihood that social workers will give feedback (85.7 % were ‘‘likely’’ when asked if they were more or less likely to give feedback to the residents after this evaluation process). Residents surveyed after the year of evaluations state that they are discussing the communication event after the fact sometimes, often or always 81 % of the time and that the evaluation was ‘‘somewhat influential,’’ ‘‘very influential,’’ or ‘‘extremely influential’’ for 79 % of responding residents. The social workers feel that residents whom they had evaluated previously in the delivery of bad news are improving when evaluating subsequent encounters 90 % of the time. This supports the idea that this communication skill may improve with experience and feedback. Social workers agree that they possess the requisite skill to evaluate residents, and report that evaluating residents did not amount to an undue burden on their work duties. In fact, they value the opportunity to help teach this important skill to new emergency medicine physicians. Social workers report being more likely to give feedback to the residents following the institution of the evaluation system. The majority felt the residents are receptive or very receptive, and all feel comfortable or very comfortable in evaluating the residents. Nearly all the residents report that social workers are the appropriate health care professionals to assess these communication skills. This perception increases with years in residency. Post-graduate year level appeared to be more important than the number of times a resident led a bad news discussion. This indicates that longevity in residency rather than number of experiences affects perceptions of social worker expertise with this specific skill. We also find that prior training in breaking bad news is associated with valuing the advice of social workers more than that of an attending physician. The value placed on feedback from an attending, however, is dependent upon the specific attending physician. Each faculty member has an individual style for delivering bad news, and residents may not want to emulate behaviors with which they do not identify. Self-assessment of performance has clear limitations as well as strengths. The use of self-assessment has two approaches: one for competency and one for personal development [39]. In our study, the self-assessments completed by residents were not used as indicators of competency. It has been shown in previous works that the accuracy of self-assessments is limited [39]. Inflated selfconfidence and scoring of effort rather than competency or achievement are just a couple of the issues with selfassessment [39]. It has also been shown that more experienced students tend to become more self-critical over time [39]. Notably in our data, in the six conversations in which

123

residents and social workers evaluated the same event, the social worker gave a higher (more favorable) score in all but one evaluation in which the resident rated the same total score. Rather than a measure of competency, we intended to use the self-assessments as tools to encourage self-reflection and as stepping-stones to conversations about their experience. It has been shown that honest reflections on residents own medical practice provides an element of learning that cannot be achieved through standard didactics or passive review of data [40]. Residents want to close gaps in knowledge, especially if the gaps are self-discovered [40]. In many circumstances, the new Accreditation Council for Graduate Medical Education (ACGME) milestones method of assessment require direct observation of learners to assess competency. The sub-competencies of Professional Values and Patient-Centered Communication are aligned with the evaluation goals of our assessments [28]. The social workers’ assessments of their direct observation of these communication skills would significantly contribute to the global evaluation of the clinical performance of emergency medicine residents used by residency program leadership.

Limitations This was a feasibility study of an educational intervention to increase deliberate practice as well as formal assessment of the delivery of bad news in the ED. There were a limited number of self-assessments and social worker evaluations submitted over the course of the study despite multiple email reminders, which may have introduced selection bias. Additionally, due to the poor response rate, interpretation of data based on any single post-graduate year (PGY) class is limited. There were 2 self-assessments from the PGY-1 class, 5 from the PGY-2 class, 6 from the PGY3 class, 0 from the PGY-4 class, and 4 from the PGY-5 class. Some residents stated they forgot the study was occurring, or simply forgot to complete the forms, and suggest a more automated process to increase participation (e.g., part of the death packet or delivered by research assistants in response to a death). The residents also selfselected what they considered to be a delivery of bad news (i.e. death notification, serious diagnosis), which may have further limited response rate. Lastly, when attending physicians assisted or did most of the notification, social workers and residents did not complete the evaluations. Attending physicians did not complete evaluations of bad news encounters during the study period. There were only six occasions with paired resident self-assessments and social worker evaluations, further limiting direct comparison.

Intern Emerg Med

The assessment tool, the modified mgBAS, was developed with simulated patients and a variety of medical professionals, and has been validated with medical students using self-assessments as well as standardized patients and practitioners [22, 32]. It has not been validated in its current form, nor for emergency medicine residents or social workers as evaluators. This intervention was performed at a single site, which may limit generalizability. EDs without access to social workers skilled in and committed to assisting with patients and families who have experienced loss are unlikely to be able to replicate our study. We are fortunate to have a team of experienced social workers (average of 16 years) who have been in our ED for an average of five years. Social workers with less experience may not feel as comfortable in evaluating and providing feedback for residents. Finally, patient and family reported-outcomes were not evaluated. Instead, experienced ED social workers evaluated the interaction. Agreement between SW and patient/family perceptions of quality communication have not been established empirically, and imposing upon families to engage in such research may not always be indicated or appropriate. Our hope was that social workers would be able to apply their expertise in communication and patient advocacy to empathize with the patient and family and provide a consistent, timely and objective evaluation.

Conclusion Continued training in direct and compassionate communication, regular pre- and post-conversation sessions with social workers and attending physicians, allowing opportunity to reflect on issues that arise, and regular evaluations are potential methods for initiating deliberate practice in order achieve expert-level skill in communication of bad news. This direct-observation assessment aligns with the evaluation of sub-competencies defined in the ACGME milestone assessments. Our study demonstrates that residents value the advice and expertise of social workers in delivering bad news in the ED, sometimes more than that of their supervising faculty. The skill of delivering bad news is a necessary and important aspect of emergency medicine training. We should attempt to use all educational tools at our disposal. An integrative approach with emergency medicine faculty and social workers is a promising method of teaching this crucial skill. Compliance with ethical standards Conflict of interest of interest.

The authors declare that they have no conflict

Statement of human and animal rights The study was reviewed by the Institutional Review Board and classified as exempt. Informed consent

None.

References 1. Buckman R (1984) Breaking bad news: why is it still so difficult? Br Med J (Clin Res Ed) 288(6430):1597 2. National Hospital Ambulatory Medical Care Survey (2015) 2010 Emergency Department Summary Tables. http://www.cdc.gov/ nchs/data/ahcd/nhamcs_emergency/2010_ed_web_tables.pdf. Accessed 30 Jan 2015 3. Parrish GA, Holdren KS, Skiendzielewski JJ, Lumpkin OA (1987) Emergency department experience with sudden death: a survey of survivors. Ann Emerg Med 16(7):792–796 4. Merlevede E, Spooren D, Henderick H et al (2004) Perceptions, needs and mourning reactions of bereaved relatives confronted with a sudden unexpected death. Resuscitation 61(3):341–348 5. Wisten A, Zingmark K (2007) Supportive needs of parents confronted with sudden cardiac death: a qualitative study. Resuscitation 74(1):68–74 6. Iserson KV, Iserson K (1999) Grave words: notifying survivors about sudden, unexpected deaths. Galen Press, Tucson 7. Dubin WR, Sarnoff JR (1986) Sudden unexpected death: intervention with the survivors. Ann Emerg Med 15(1):54–57 8. Adamowski K, Dickinson G, Weitzman B, Roessler C, CarterSnell C (1993) Sudden unexpected death in the emergency department: caring for the survivors. CMAJ 149(10):1445–1451 9. Takayesu JK, Hutson HR (2004) Communicating life-threatening diagnoses to patients in the emergency department. Ann Emerg Med 43(6):749–755 10. Walters DT, Tupin JP (1991) Family grief in the emergency department. Emerg Med Clin North Am 9(1):189–206 11. McLauchlan CAJ (1990) ABC of major trauma: handling distressed relatives and breaking bad-news. Br Med J 301(6761):1145–1149 12. Deja K (2006) Social workers breaking bad news: the essential role of an interdisciplinary team when communicating prognosis. J Palliat Med 9(3):807–809 13. Cagle JG, Kovacs PJ (2009) Education: a complex and empowering social work intervention at the end of life. Health Soc Work 34(1):17–27 14. Lord B, Pockett R (1998) Perceptions of social work intervention with bereaved clients: some implications for hospital social work practice. Soc Work Health Care 27(1):51–66 15. Educational Policy and Accreditation Standards Educational Policy and Accreditation Standards for Baccalaureate and Master’s Social Work Programs Developed by the Council on Social Work Education (CSWE) Commission on Educational Policy and the CSWE Commission on Accreditation; Educational Policy approved by the CSWE Board of Directors on March 20, 2015; Accreditation Standards approved by the CSWE Commission on Accreditation on June 11, 2015. Copyright  2015 Council on Social Work Education 16. Kaul R (2001) Coordinating the death notification process: the roles of the emergency room social worker and physician following a sudden death. Brief Treat Crisis Interv 1:101–114 17. VonBloch L (1996) Breaking the bad news when sudden death occurs. Soc Work Health Care 23(4):91–97 18. Gisondi MA, Lu DW, Yen M et al (2010) Adaptation of EPECEM curriculum in a residency with asynchronous learning. West J Emerg Med 11(5):491–499

123

Intern Emerg Med 19. Han PK, Keranen LB, Lescisin DA, Arnold RM (2005) The palliative care clinical evaluation exercise (CEX): an experiencebased intervention for teaching end-of-life communication skills. Acad Med 80(7):669–676 20. Curtis JR, Back AL, Ford DW et al (2013) Effect of communication skills training for residents and nurse practitioners on quality of communication with patients with serious illness: a randomized trial. JAMA 310(21):2271–2281 21. Ptacek J, Eberhardt TL (1996) Breaking bad news: a review of the literature. JAMA 276(6):496–502 22. Schildmann J, Kupfer S, Burchardi N, Vollmann J (2012) Teaching and evaluating breaking bad news: a pre-post evaluation study of a teaching intervention for medical students and a comparative analysis of different measurement instruments and raters. Patient Educ Couns 86(2):210–219 23. Rosenbaum ME, Ferguson KJ, Lobas JG (2004) Teaching medical students and residents skills for delivering bad news: a review of strategies. Acad Med 79(2):107–117 24. Wakefield A, Cocksedge S, Boggis C (2006) Breaking bad news: qualitative evaluation of an interprofessional learning opportunity. Med Teach 28(1):53–58 25. Lienard A, Merckaert I, Libert Y et al (2010) Is it possible to improve residents breaking bad news skills? A randomised study assessing the efficacy of a communication skills training program. Br J Cancer 103(2):171–177 26. Bowyer MW, Hanson JL, Pimentel EA et al (2010) Teaching breaking bad news using mixed reality simulation. J Surg Res 159(1):462–467 27. Chi J, Verghese A (2013) Improving communication with patients learning by doing. JAMA 310(21):2257–2258 28. Education ACfGM (2014) The emergency medicine milestone project 29. Ericsson KA (2008) Deliberate practice and acquisition of expert performance: a general overview. Acad Emerg Med 15(11):988–994 30. Gelfman LP, Lindenberger E, Fernandez H et al (2014) The effectiveness of the Geritalk communication skills course: a real-

123

31.

32.

33. 34.

35.

36.

37.

38.

39. 40.

time assessment of skill acquisition and deliberate practice. J Pain Symptom Manage 48(4):738–744.e731-736 Szmuilowicz E, Neely KJ, Sharma RK, Cohen ER, McGaghie WC, Wayne DB (2012) Improving residents’ code status discussion skills: a randomized trial. J Palliat Med 15(7):768–774 Miller SJ, Hope T, Talbot DC (1999) The development of a structured rating schedule (the BAS) to assess skills in breaking bad news. Br J Cancer 80(5–6):792–800 Charmaz K (2014) Constructing grounded theory. Sage, New York Amiel GE, Ungar L, Alperin M, Baharier Z, Cohen R, Reis S (2006) Ability of primary care physician’s to break bad news: a performance based assessment of an educational intervention. Patient Educ Couns 60(1):10–15 Dikici MF, Yaris F, Cubukcu M (2009) Teaching medical students how to break bad news: a Turkish experience. J Cancer Educ 24(4):246–248 Vail L, Sandhu H, Fisher J, Cooke H, Dale J, Barnett M (2011) Hospital consultants breaking bad news with simulated patients: an analysis of communication using the Roter Interaction Analysis System. Patient Educ Couns 83(2):185–194 Wouda JC, van de Wiel HB (2012) The communication competency of medical students, residents and consultants. Patient Educ Couns 86(1):57–62 McGaghie WC, Issenberg SB, Cohen ER, Barsuk JH, Wayne DB (2011) Does simulation-based medical education with deliberate practice yield better results than traditional clinical education? A meta-analytic comparative review of the evidence. Acad Med 86(6):706–711 Evans AW, McKenna C, Oliver M (2002) Self-assessent in medical practice. J R Soc Med 95:511–513 Hildebrand D, Trowbridge E, Roach MA, Sullivan AG, Broman AT, Vogelman B (2009) Resident self-assessment and self-reflection: university of Washington-Madison’s five-year study. J Gen Intern Med 24(3):361–365