Epidemiology of Connectional Silence in Palliative Care Serious Illness Conversations

Context Deep human connection can reduce suffering in serious illness, particularly when it happens during conversations with one’s clinical team amid the isolating, confusing, and often terrifying experience of acute hospitalization in advanced cancer. Some moments of conversational silence are markers of human connection, but we know very little empirically about their prevalence, frequency, or associated characteristics. Purpose To describe the epidemiology of Connectional Silence in palliative care conversations with people who are hospitalized for advanced cancer. Methods As part of a multisite cohort study, we audio-recorded initial palliative care consultations involving 216 hospitalized people with advanced cancer. Using tandem machine-learning and human coding, we identified each conversational pause lasting at least two seconds in the 6,415 minutes of audio. We distinguished episodes of Connectional Silence from other pauses and sub-categorized each Connectional Silence into two mutually-exclusive types: Invitational or Emotional. We described the prevalence, frequency, and distribution of each conversational silence using standard epidemiological methods. Results Among all 6,600 two-second silences, we observed 232 (3.5 %) Invitational Connectional Silences (ICS) and 315 (4.8 %) Emotional Connectional Silences (ECS). The per-conversation prevalence of ICS and ECS were 37% (79/216) and 42% (90/216), respectively. The prevalence of ICS was associated with clinician expectations for shorter survival time (ORadj: 2.99; 95% CI: 1.29, 6.94). ECS prevalence was associated with higher versus lower patient pre-consultation ratings of spiritual pain (ORadj: 2.09, 95% CI: 1.01, 4.36). Conclusion Both Invitational and Emotional Connectional Silence are common during inpatient palliative care conversations and associated with patients’ level of spiritual pain (Emotional) and survival prognosis (Invitational)