Improving Communication With Surrogate Decision-Makers: A Pilot Initiative

Important and difficult conversations often occur between physicians and patients' surrogate decision-makers. These are individuals, often family members or close friends, who are entrusted with medical decisions for adult patients who lack the capacity to make them. This is a necessity, as nearly half of hospitalized adults cannot make their own medical decisions, and this is particularly true for critically ill patients, for whom surrogates make 75% of medical decisions, including decisions about code status.^1,2  Poor communication between physicians and patient surrogates may exacerbate anxiety, guilt, and grief for surrogates, and may contribute to physician stress and burnout.² 

Empathic communication can reduce surrogate burden, improve medical decision-making and the quality of patient care, reduce the use of nonbeneficial life-sustaining treatments, and mitigate clinician burnout.³  Despite a 2011 systematic review advocating for surrogate support, medical training in communication skills continues to focus on the physician-patient relationship.¹  Formal training in how to effectively communicate with patients' surrogates and to promote patient-centered decision-making remains essentially nonexistent.¹ 

To improve this aspect of the curriculum for internal medicine interns, we developed a novel experiential learning workshop focused on communication with surrogates. We hypothesized that the interns would not demonstrate baseline proficiency in surrogate decision-making communication skills, and we anticipated that the educational intervention would result in a measureable improvement.

We designed an intensive 6-hour communication skills workshop for internal medicine residents at an urban academic medical center on a designated 2-week ambulatory rotation. The teaching focused on promoting best practices for working with surrogates who have to make difficult end-of-life and goals-of-care decisions. The format and content of the workshop were based on the prior work of collaborators at Memorial Sloan Kettering Cancer Center, the authors' research and clinical experience with surrogates and decision-making, and the authors' prior experience with teaching and assessing communication skills.^4–15  Colleagues in critical care, obstetrics and gynecology, pediatrics, and oncology provided additional expertise. To ensure the curriculum included current best practices for working with surrogates, developing teaching strategies, and evaluating communication skills, we completed a literature review using key terms.^5,16–23 

The workshop included a preworkshop communication skills assessment via an objective structured clinical examination (OSCE), a lecture and small group discussion, simulated encounters to practice new skills, and a postworkshop assessment via an OSCE, all completed on the same day. Figure 1 depicts key workshop elements.

The OSCEs consisted of encounters with actors portraying standardized surrogates of incapacitated patients (such as playing the adult child of a patient with dementia and cancer metastatic to the brain). The intern was instructed to assist the surrogate in determining how to proceed with the patient's care, and whether or not to pursue chemotherapy and/or radiation. A trained faculty facilitator present in the room observed the simulation, and evaluated the OSCE along with the standardized surrogate. Encounters were neither audiotaped nor videotaped. Actors had prior extensive experience in OSCEs, simulation, and providing feedback on communication skills. All actors and faculty facilitators completed a minimum of 2 hours of training, including instruction on surrogate decision-making, how to complete the OSCE assessment form, and how to provide feedback. To promote verisimilitude, clinical examination rooms were utilized for all OSCEs and practice encounters.^24–27 

Two OSCEs were created for assessment: participants were randomized, with half completing the first case as a preworkshop OSCE and the second as a postworkshop OSCE, and the remainder completing the OSCEs in reverse order. A crossover study design reduced bias from potential confounding factors present in the OSCE cases themselves.²⁸  The faculty-facilitated simulated practice encounters had unique cases. A total of 11 clinicians served as faculty facilitators, including physicians from internal medicine, anesthesiology, critical care, and psychiatry, and nurses and medical ethicists.

Faculty facilitators and standardized surrogates evaluated participants' OSCE performances using the expanded Gap Kalamazoo Consensus Statement Assessment Form (provided as online supplemental material). This instrument was chosen due to substantial evidence documenting its validity in similar populations.^12–15  Based on the Kalamazoo Consensus Statement, this tool uses an anchored 5-point Likert scale to assess 9 aspects of communication: building a relationship; opening the discussion; gathering information; understanding the patient's perspective; sharing information; reaching agreement on problems and plans; providing closure; demonstrating empathy; and communicating accurate information. Scores of 3 (good) or higher are defined as an appropriate level; below 3 is suboptimal.^12–15 

Three components were added to adapt the tool for this study: a surrogate decision-making skill (ability to identify the patient's preferences via the surrogate) and 2 summative categories (overall therapeutic value of the interaction and overall communication skills). The modified tool had 12 items. The faculty facilitator and standardized surrogate collectively completed 1 assessment form per participant, after reaching agreement through an iterative negotiation process. Participants completed a course evaluation form, which assessed their perceptions of the workshop and their resulting confidence for communicating with surrogate decision-makers.

We compared average preworkshop and postworkshop OSCE scores to test our hypotheses that residents would not demonstrate baseline proficiency in skills relevant to surrogate decision-making, and that an educational intervention would improve skills; we also made a comparison of average preworkshop and postworkshop scores for each of the assessment tool's 12 items. We used Wilcoxon signed rank tests to compare mean scores.

The Weill Cornell Medicine Institutional Review Board approved this study.

Linear regression analysis measured the association between self-reports of any previous instruction in best practices for surrogate decision-making and OSCE performance. Cronbach's α determined the internal consistency of the instrument. Data analysis was completed using SPSS Statistics version 20.0.0 (IBM Corp, Armonk, New York). All tests of statistical significance were set at an α of .05.

From April 2016 through June 2016, 33 of 44 internal medicine residents (75%) participated in the workshop in 4 different sessions. There were roughly equal numbers of male and female participants. All participants had previously watched a goals-of-care discussion, and almost all (97%, 32 of 33) previously contributed to a goals-of-care discussion and “engaged a surrogate in end-of-life decision-making.” Only 36% (12 of 33) self-reported previously receiving instruction in best practices for surrogate decision-making.

Average preworkshop OSCE scores (M = 3.3, SD = 0.9) were significantly lower than postworkshop scores (M = 4.3; SD = 0.8; Z = 4.193; P < .001; effect size r = 0.52), indicating a significant improvement in communication skills. Results were similar for analyses when comparing average scores for the 9-item original Gap Kalamazoo Consensus Statement Assessment Form versus average scores for the modified 12-item scale. Similar to both the Kalamazoo Consensus Statement and the Gap Kalamazoo Consensus Statement Assessment Form, preworkshop (α = .94) and postworkshop (α = .90) assessment scales had excellent internal consistency.^12–15 

A subgroup analysis comparing mean scores for each skill for preworkshop and postworkshop OSCEs are represented in figure 2. Average baseline scores were above 3 (good) for most skills, except “understands the patient's and family's perspective,” which had an average score of 2.7. The largest number of participants demonstrated suboptimal scores on the preworkshop OSCE for “understands patient's and family's perspective” (42%, 14 of 33) and “reaches agreement” (39%, 13 of 33), which decreased after the workshop. As shown in figure 2, mean postworkshop scores significantly increased for each skill (P < .05), with the largest increases in “understands the patient's and family's perspective” (+1.4) and “demonstrates empathy” (+1.2).

Prior to participating in the workshop, 36% (12 of 33) of participants self-reported that they had received previous instruction in best practices for surrogate decision-making. Linear regression demonstrates that OSCE performance was not affected by self-reporting of previous instruction in best practices for surrogate decision-making (B = 0.18(0.24); t = 7.29; P < .47). The majority of participants (91%, 30 of 33) self-reported feeling more confident after the workshop; 6% (2 of 33) felt the same/more confident, and 3% (1 of 33) felt the same confidence.

Our findings show that a significant portion of residents have room for improvement in communication skills relevant to critical surrogate decision-making, including demonstrating empathy and understanding the patient's and family's perspective. The workshop showed promise in improving these and other important skills.

Other institutions have similarly reported positive outcomes using OSCEs and simulation to evaluate and improve trainees' abilities to navigate difficult conversations. Most of these were conducted in pediatrics. This is one of the few studies that highlight skills unique to working with surrogates for adult patients.^5,18–22  Many surrogates experience stress, anxiety, and even subsequent posttraumatic stress disorder; thus, the stakes may be too high for trainees to receive “on the job training” via real-life encounters.²⁹  Experiential learning with simulated encounters provides an ideal format for teaching communication skills, as this pedagogical method provides an opportunity to practice new skills in a safe space.^5,24,30,31 

For other institutions seeking to incorporate simulation into their clinical training, our findings show that it is feasible to conduct an experiential workshop during an ambulatory rotation. The curriculum was developed through funded research and is available from the authors on request. The only other major expense was the actors' fees. Faculty facilitators donated their time to this program, and the department of medicine provided refreshments for all participants.

This study has limitations. The OSCEs were neither audiotaped nor videotaped, the faculty facilitator and standardized surrogate collectively completed 1 assessment form for each participant, and interrater reliability was not obtained. The same faculty facilitator evaluated participants during both OSCEs; due to their clinical responsibilities, 3 of 11 faculty facilitators potentially had prior clinical contact with participants, which may have served to introduce bias. For educational purposes, participants benefited more from having the same faculty facilitator present to identify changes in skills and confidence over time. To mitigate this, each OSCE utilized a unique standardized surrogate (actors did not work with interns twice) who took part in the evaluation. While there are limitations to OSCEs, including the high cost of actors, potential “collusion” by students who completed the case discussing its content with others, and the fact that simulation is not “the real world,” they are commonly accepted as robust assessment tools in the clinical setting.^32,33 

Finally, it is possible that training sessions that emphasized communication skills relevant to surrogate decision-making could have led evaluators to assess these items more stringently and introduced evaluation bias.³⁴ 

Finally, we did not assess whether the improvements in communication skills from the workshop are enduring. Future research should assess whether this results in improvements in communication skills when participants are actually faced with leading difficult conversations with actual surrogates.

This pilot study identified and addressed the need for standardized educational programs to teach communication skills specific to surrogate decision-making for adult patients. The findings suggest that an experiential learning intervention can meaningfully improve surrogate decision-making skills.