Editorial <\/p>\n\n\n\n
May 21, 2023<\/p>\n\n\n\n
Goals-of-care conversations between clinicians and patients with serious illness have been shown to improve measures of patient and family well-being, while also decreasing health care costs and high-intensity treatments at the end of life.1<\/a><\/sup>,2<\/a><\/sup> Nonetheless, achieving the goal of skillful, timely goals-of-care conversations for hospitalized patients with serious illness remains a challenge due to a multitude of patient, clinician, and system-level factors, including clinician discomfort and time constraints.3<\/a><\/sup><\/a><\/p>\n\n\n\n Prior efforts to improve goals-of-care conversations in the hospital setting have focused on intensive interventions to improve clinicians\u2019 communication skills, adding trained communication facilitators to the clinical team, and deploying standardized communication protocols.1<\/a><\/sup>,4<\/a><\/sup>,5<\/a><\/sup> Although these interventions have improved patient-centered outcomes, not all health systems are able to implement them due to cost or lack of personnel.<\/a><\/p>\n\n\n\n In this issue of JAMA<\/em>, Curtis and colleagues6<\/a><\/sup> demonstrate the value of leveraging technology to nudge clinicians to conduct goals-of-care conversations. In a pragmatic, multicenter, randomized clinical trial (RCT) of 2512 hospitalized older patients with serious illness, stratified by history of dementia, investigators assessed whether a clinician-facing communication-priming patient-specific intervention (Jumpstart Guide) increased the proportion of patients with a goals-of-care discussion documented in the electronic health record (EHR) within 30 days of randomization.<\/a><\/p>\n\n\n\n Clinicians caring for patients in the intervention group were emailed a 1-page Jumpstart Guide containing EHR-populated, patient-specific data such as code status and advanced directives, if available, as well as suggested topics and phrases to use in a goals-of-care conversation. The proportion of patients with documented goals-of-care conversations was higher with the intervention (34.5%) vs usual care (30.4%).6<\/a><\/sup> Notably, the effect size was significantly larger for patients of minoritized race or ethnicity (10.2%) vs non-Hispanic White patients (1.6%).6<\/a><\/sup>There were no differences across groups for the secondary outcomes at 1 month, including intensive care unit admissions, emergency department visits, palliative care consultations, and intensive care unit\u2013 and hospital-free days. Thirty-day survival was high (95%) and equivalent between groups.<\/a><\/p>\n\n\n\n This study has numerous, important strengths. The intervention that was tested is practical, low cost, and readily scalable to hospitals that have reasonably modern EHR systems. Furthermore, the intervention was developed using innovative, human-centered design principles with input from multiple stakeholders.7<\/a><\/sup> The use of natural language processing to detect goals-of-care discussions within EHRs is a novel and important methodological advance to improve the efficiency of research and quality improvement initiatives.<\/a><\/p>\n\n\n\n The RCT was rigorously conducted and included a large, racially and ethnically diverse patient population drawn from both academic and community hospitals. The investigators also enrolled a high proportion of patients with limited English proficiency, who are often underrepresented in RCTs. Furthermore, investigators achieved high rates of ascertainment for all primary and secondary outcomes.<\/a><\/p>\n\n\n\n The study\u2019s major limitation is that, due to the pragmatic nature of the RCT, investigators were unable to determine to what extent clinicians engaged with the intervention (ie, whether they opened and read the email prompt). More granular data from clinicians to understand why many did not conduct goals-of-care conversations despite receiving the prompt could explain the outcomes or illuminate ways to improve the intervention in the future.<\/a><\/p>\n\n\n\n There are several possible explanations for why the intervention resulted in only a modest increase in the occurrence of goals-of-care conversations. First, because the overall mortality rate in the cohort was low (5%), it is possible that many patients were not facing the kind of decisions about life support for which a goals-of-care conversation is likely to be immediately valuable. For such patients, it is possible that clinicians did not believe it would be valuable to engage in broad, future-oriented goals-of-care planning. If so, interventions may also need to target changing clinicians\u2019 beliefs about the value of future-oriented goals-of-care planning. Clinicians will also need clarity about what should be included in these discussions, which will be different from their usual conversations with hospitalized patients.<\/a><\/p>\n\n\n\n Second, it may be that a clinician-facing intervention is insufficient. Serious illness communication involves, at minimum, a dyadic relationship between clinicians and patients. Prior RCTs on communication-priming interventions conducted by this same group of investigators8<\/a><\/sup>,9<\/a><\/sup> included both clinician- and patient-facing components and demonstrated larger effect sizes.<\/a><\/p>\n\n\n\n Third, other interventions may be needed to overcome the considerable structural and administrative barriers that prevent clinicians from conducting goals-of-care discussions when they determine such a conversation is warranted. These barriers may include insufficient time due to high workload and the complexities of scheduling meetings that include family members or other treating clinicians.<\/a><\/p>\n\n\n\n In light of the modest observed effect size of the intervention, some clinicians and hospital administrators may conclude that the intervention should not be broadly adopted. We believe this conclusion would be mistaken. Decisions to adopt new health care interventions do not depend merely on effect size, but instead require consideration of the benefits relative to the costs of adoption. In addition, the intervention had a substantially larger effect for patients with minoritized race or ethnicity. Although the reasons for this differential effect are unclear, one possibility is that the intervention overcame implicit bias by clinicians that posed a barrier to initiating needed goals-of-care conversations for minoritized patients.10<\/a><\/sup>,11<\/a><\/sup> In this sense, the Jumpstart Guide intervention may be a valuable tool to mitigate inequities in serious illness care, which is a pressing public health priority.<\/a><\/p>\n\n\n\n How can the field build on these findings to promote patient-centered care in the context of serious illness? First, although achieving timely goals-of-care conversations is an important step, it is not sufficient because many clinicians do not feel confident in their ability to skillfully conduct these challenging conversations. Indeed, existing literature documents serious deficiencies in these conversations, including ineffective communication about prognosis,12<\/a><\/sup> failures to elicit patients\u2019 values and preferences,13<\/a><\/sup> and reluctance by clinicians to offer recommendations even when requested to do so.14<\/a><\/sup> Additional interventions are urgently needed to ensure that clinicians have these needed communication skills for effective, empathic conversations.15<\/a><\/sup><\/a><\/p>\n\n\n\n Second, researchers and policymakers should abandon the mindset that a single intervention will be developed that cures all that ails end-of-life care in the US. Instead, we think the path forward will necessarily involve multiple interventions directed at multiple levels of the health care delivery system, health care funding, and societal attitudes. These may include public awareness campaigns, regulations that promote advanced care planning, and enhanced training to support clinicians that encourages a culture of acceptance around serious illness communication.16<\/a><\/sup> The present RCT is an excellent example of the value of developing innovative, scalable interventions to overcome discrete barriers to goals-of-care conversations.<\/a><\/p>\n\n\n\n Third, researchers and clinicians should work to develop longitudinal strategies for serious illness conversations and care that bridge the chasm between inpatient and outpatient care. To this end, more clarity is needed about what should be included during inpatient goals-of-care conversations when there are no pressing treatment decisions to be made.<\/a><\/p>\n\n\n\n