Looking Beyond Mortality Prediction: Primary Care Physician Views of Patients' Palliative Care Needs Predicted by a Machine Learning Tool.

OBJECTIVES: To assess primary care physicians' (PCPs) perception of the need for serious illness conversations (SIC) or other palliative care interventions in patients flagged by a machine learning tool for high 1-year mortality risk. METHODS: We surveyed PCPs from four Brigham and Women's Hospital primary care practice sites. Multiple mortality prediction algorithms were ensembled to assess adult patients of these PCPs who were either enrolled in the hospital's integrated care management program or had one of several chronic conditions. The patients were classified as high or low risk of 1-year mortality. A blinded survey had PCPs evaluate these patients for palliative care needs. We measured PCP and machine learning tool agreement regarding patients' need for an SIC/elevated risk of mortality. RESULTS: Of 66 PCPs, 20 (30.3%) participated in the survey. Out of 312 patients evaluated, 60.6% were female, with a mean (standard deviation [SD]) age of 69.3 (17.5) years, and a mean (SD) Charlson Comorbidity Index of 2.80 (2.89). The machine learning tool identified 162 (51.9%) patients as high risk. Excluding deceased or unfamiliar patients, PCPs felt that an SIC was appropriate for 179 patients; the machine learning tool flagged 123 of these patients as high risk (68.7% concordance). For 105 patients whom PCPs deemed SIC unnecessary, the tool classified 83 as low risk (79.1% concordance). There was substantial agreement between PCPs and the tool (Gwet's agreement coefficient of 0.640). CONCLUSIONS: A machine learning mortality prediction tool offers promise as a clinical decision aid, helping clinicians pinpoint patients needing palliative care interventions.

Early serious illness communication in hospitalized patients: A study of the implementation of the Speaking About Goals and Expectations (SAGE) program.

BACKGROUND: Early conversations about patients' goals and values in advancing serious illness (serious illness conversations) can drive better healthcare. However, these conversations frequently happen during acute illness, often near death, without time to realize benefits of early communication. METHODS: The Speaking About Goals and Expectations (SAGE) Program, adapted from the Serious Illness Care Program, is a multicomponent intervention designed to foster earlier and more comprehensive serious illness conversations for patients admitted to the hospital. We present a quality improvement study of the SAGE Program assessing older adults admitted to a general medicine service at the Brigham & Women's Hospital in Boston, Massachusetts. Our primary outcomes included the proportion of patients with at least one documented conversation, the timing between first conversation documented and death, the quality of conversations, and their interprofessional nature. Secondary outcomes assessed evaluations of the training and hospital utilization. RESULTS: We trained 37 clinicians and studied 133 patients split between the SAGE intervention and a comparison population. Intervention patients were more likely to have documented serious illness conversations (89.1% vs. 26.1%, p < 0.001); these conversations occurred earlier (mean of 598.9 vs. 180.8 days before death, p < 0.001) and included more key elements of conversation (mean of 6.56 vs. 1.78, p < 0.001). CONCLUSIONS: This study demonstrated significant differences in the frequency and quality of serious illness conversations completed earlier in the illness course for hospitalized patients. IMPLICATIONS: Programs designed to drive serious illness conversations earlier in the hospital may be an effective way to improve care for patients not reached in the ambulatory setting. LEVEL OF EVIDENCE: Prospectively designed trial, non-randomized sample.

Improving Communication About Serious Illness in Primary Care: A Review.

IMPORTANCE: The Institute of Medicine recently called for systematic improvements in clinician-led conversations about goals, values, and care preferences for patients with serious and life-threatening illnesses. Studies suggest that these conversations are associated with improved outcomes for patients and their families, enhanced clinician satisfaction, and lower health care costs; however, the role of primary care clinicians in driving conversations about goals and priorities in serious illness is not well defined. OBJECTIVE: To present a review of a structured search of the evidence base about communication in serious illness in primary care. EVIDENCE REVIEW: MEDLINE was searched, via PubMed, on January 19, 2016, finding 911 articles; 126 articles were reviewed and selected titles were added from bibliography searches. FINDINGS: Review of the literature informed 2 major topic areas: the role of primary care in communication about serious illness and clinician barriers and system failures that interfere with effective communication. Literature regarding the role that primary care plays in communication focused primarily on the ambiguity about whether primary care clinicians or specialists are responsible for initiating conversations, the benefits of primary care clinicians and specialists conducting conversations, and the quantity and quality of discussions. Timely and effective communication about serious illness in primary care is hampered by key clinician barriers, which include deficits in knowledge, skills, and attitudes; discomfort with prognostication; and lack of clarity about the appropriate timing and initiation of conversations. Finally, system failures in coordination, documentation, feedback, and quality improvement contribute to lack of conversations. CONCLUSIONS AND RELEVANCE: Clinician and system barriers will challenge primary care clinicians and institutions to meet the needs of patients with serious illness. Ensuring that conversations about goals and values occur at the appropriate time for seriously ill patients will require improved training, validation, and dissemination of patient selection tools, systems for conducting and revisiting conversations, accessible documentation, and incentives for measurement, feedback, and continuous improvement.