Confidence with and Barriers to Serious Illness Communication: A National Survey of Hospitalists

Introduction

Improving serious illness communication is a national priority; decades of research show lapses in prognosis and goals of care communication, ¹ including in hospitals. ² Recent work in university-based Canadian hospitals indicates that hospital clinicians continue to struggle with serious illness communication in challenging situations, for example, when patients and families have trouble accepting a poor prognosis or understanding treatments. ³

Over the past 20 years, the hospitalist model—in which physicians whose professional focus is the care of hospitalized patients direct inpatient care—has become predominant in U.S. hospitals.^4,5 The hospitalist professional association, the Society of Hospital Medicine (SHM), identifies palliative care as a core competency. ⁶ Yet, previous national surveys found insufficient training of hospitalists in palliative care. ⁷ Recent small studies at academic centers indicate that hospitalists miss opportunities to engage in prognosis and goals of care discussions, ⁸ and lack confidence in complex tasks ⁹ such as responding to patient or family desire for interventions that they feel are not medically indicated. ¹⁰

Methods

We conducted a web-based survey of a national sample of hospitalists, distributed in May 2016 through SHM, aimed at understanding hospitalists' experiences and needs in serious illness communication. Hospitalists and palliative care physicians with clinical and academic experience in hospital medicine, palliative care, and education developed the survey. It was informed by existing literature about the roles of hospitalists in palliative care,^9–14 the SHM palliative care core competencies ⁸ and a focus group at the 2016 SHM Annual Meeting. Nine hospitalists from six institutions piloted a draft survey; they provided feedback on clarity, appropriateness, and completion time, with a goal of 10 minutes. The final survey (Appendix 1) contained 39 items.

SHM distributed an email containing a nonunique link to the survey in SurveyMonkey to a random selection of 4000 members, one-third of the SHM membership. To obtain the perspectives of hospitalists who were not SHM members, we encouraged participants to share the survey link with colleagues. Thus, the respondents are a combination of a randomized and convenience sample. The survey remained open for one month and one reminder was sent. To encourage participation, respondents were given the option to provide their email addresses to be entered for a chance to win a gift certificate. Institutional Review Board review was not sought as the purpose of the survey was to improve the quality of resources provided to hospitalists by their professional society. Based on the distributions, we dichotomized responses as follows: most/every shifts versus never/rarely/some shifts; confident/very confident versus not at all confident/not very confident/neutral; moderate/extreme barrier versus not a barrier/minimal barrier.

Results

Three hundred thirty-two hospitalists from 43 states completed the survey (Table 1). Though total recipients cannot be directly quantified, we estimate the response rate at ∼8% (332 of 4000 to whom SHM sent the email). Participants had practiced for a mean of 9.4 years and worked in a range of settings, including community and academic centers and large and small hospitals. Nearly a third reported limited access to palliative care services. Almost half reported that hospitalists at their institutions had been asked by administrators to be involved in efforts to improve the provision of palliative care; 16% attended on their hospital palliative care consultation service, and 7% were board-certified in Hospice and Palliative Medicine.

Participants often had concerns about patient and family understanding of prognosis, and the appropriateness of code status and medical interventions (Table 2). Confidence in serious illness communication varied by task (Table 3). Most felt confident or very confident in discussing goals of care and prognosis. Fewer reported being confident or very confident in responding to patients or families who had not accepted the seriousness of a patient's illness, to counsel patients or families who request medically inappropriate treatments, and in managing conflict among patients or families and the healthcare team. A minority felt confident or very confident to use self-care techniques to reduce burnout and compassion fatigue.

Most frequently cited barriers to prognosis and goals of care discussions were lack of time, lack of prior outpatient discussions, prognostic expectations from other physicians on the patient's care team, and difficulty finding records of previous discussions (Table 4). Many also reported lack of long-term relationships with patients and frequent handoffs between hospitalists as significant barriers.

Discussion

We conducted a national survey of U.S. hospitalists' experiences in serious illness communication, which was completed by 332 hospitalists from 43 states. Our results mirror those of a recent large study by You et al. of Canadian hospital-based clinicians, ³ as well as a smaller study of hospitalists at two U.S. academic centers, ¹⁰ in that our respondents reported lower confidence in what might be considered more complex or emotionally charged serious illness communication tasks, such as when patients or families have not accepted a poor prognosis or request treatments that the hospitalist feels are medically inappropriate, or when there is conflict among patients, families, and clinicians. However, in comparison to the results of You, our participants rated lack of time and differences in prognostic expectations among clinicians as more significant barriers. We also identified additional barriers, such as handoff frequency, which were not previously described in the literature.

There is a paradox between the high confidence that our participants expressed in conducting goals of care discussions and the findings of studies that indicate concern about the quality of these discussions,^1,10,11 including that hospitalists miss opportunities to have substantive discussions of prognosis and goals of care.^8,9 That most respondents reported education in palliative care is heartening; however, that many still felt unprepared to engage in more advanced serious illness communication tasks indicates that there is still need for improvement. Further investigations may benefit from measuring outcomes other than self-reported confidence. This is especially true given our finding that hospitalists reported frequently encountering scenarios for which they reported lower confidence, such as communicating with patients and families who do not adequately understand a prognosis.

Our report has a number of limitations. First, our study combines a randomized and convenience sample. We cannot distinguish responses from SHM members and those who were forwarded the survey, and cannot accurately calculate a response rate, though we estimate it to be ∼8%. It is also likely that, in addition to the low response rate, respondents had a greater-than-average interest in palliative care, and so our results may overestimate the level of confidence of hospitalists nationally or be otherwise skewed, based on this bias. Although this sample may not represent the majority of hospitalists, it remains one of the largest recent cohorts to be queried. Third, the survey explored hospitalists' self-rated confidence, as opposed to objective measures of experience, knowledge, or skill. Additionally, the survey tool was generated based on a prior survey, the input of an expert group of clinicians in hospital medicine and palliative care, and pilot-tested; it was, however, not formally validated.

As the physicians caring for many of the most seriously ill patients in the United States, hospitalists are well-positioned to be providers of primary palliative care, ¹² especially where hospitals lack adequate or, in many cases, any palliative care consultation services.^13,14,15 Our results suggest the need to develop opportunities to support hospitalists in leading serious illness communication: by integrating serious illness communication into hospitalists' workflows with accompanying time to have discussions, improving documentation systems, fostering communication between inpatient and outpatient clinicians, and developing targeted training on challenging communication scenarios.