Journal of palliative medicine最新文献

Context: Artificial intelligence (AI), particularly large language models (LLMs), offers the potential to augment clinical decision-making, including in palliative care pharmacy, where personalized treatment and assessments are important. Despite the growing interest in AI, its role in clinical reasoning within specialized fields such as palliative care remains uncertain. Objectives: This study examines the performance of four commercial-grade LLMs on a Script Concordance Test (SCT) designed for pharmacy students in a pain and palliative care elective, comparing AI outputs with human learners' performance at baseline. Methods: Pharmacy students from 2018 to 2023 completed an SCT consisting of 16 clinical questions. Four LLMs (ChatGPT 3.5, ChatGPT 4.0, Gemini, and Gemini Advanced) were tested using the same SCT, with their responses compared to student performance. Results: The average score for LLMs (0.43) was slightly lower than that of students (0.47), but this difference was not statistically significant (p = 0.55). ChatGPT 4.0 achieved the highest score (0.57). Conclusions: While LLMs show potential for augmenting clinical decision-making, their limitations in patient-centered care highlight the necessity of human oversight and reinforce that they cannot replace human expertise in palliative care. This study was conducted in a controlled research setting, where LLMs were prompted to answer clinical reasoning questions despite default safety restrictions. However, this does not imply that such prompts should be used in practice. Future research should explore alternative methods for assessing AI decision-making without overriding safety mechanisms and focus on refining AI to better align with complex clinical reasoning. In addition, further studies are needed to confirm AI's comparative effectiveness, given the sample size limitations.

Background: The Affordable Care Act (ACA) aimed to expand insurance coverage, improve health outcomes, and reduce costs. We assessed the impact of the ACA on hospice or palliative care utilization among patients with stage IV gastrointestinal (GI) cancer. Methods: Individuals diagnosed with stage IV GI cancer between 2007 and 2019 were identified from the Medicare database. An interrupted time series analysis (ITS) examined the impact of ACA on palliative care utilization. Entropy balancing and gamma regression were used to assess the cost implications of not utilizing palliative care. Results: Among the 26,227 stage IV GI cancer Medicare beneficiaries, approximately half (53.9%) were male. Overall, 80.5% of patients used palliative care before death. Utilization increased from 54.3% in 2007 to 84% in 2013 pre-ACA (slope: +0.009; 95% confidence interval [CI]: 0.005-0.012) and from 84.5% in Q1 2014 to 89.7% in Q4 2019 post-ACA (slope: +0.004; 95% CI: 0.0007-0.007), indicating slow progress in palliative care uptake. The ITS model demonstrated that ACA implementation did not affect palliative care utilization (slope: -0.006; 95% CI: -0.017 to +0.004). Patients from minority racial groups (odds ratio [OR]: 0.79; 95% CI: 0.74-0.86) and those in moderate (OR: 0.86; 95% CI: 0.80-0.94) and high (OR: 0.68; 95% CI: 0.62-0.74) Social Vulnerability Index (SVI) counties were less likely to use palliative care in both pre- and post-ACA eras. Palliative care use was associated with $2,633 lower total expenditure. Conclusion: ACA implementation did not improve palliative care utilization for racial minorities and high SVI groups. Targeted efforts are needed to improve access to equitable end-of-life care.

Background: Palliative sedation has major clinical, social and ethical implications. Vicariously witnessing suffering in others is known to be distressing. However, little is understood about how palliative sedation is experienced by relatives and health care professionals. Objectives: To explore the experiences of relatives and health care professionals with palliative sedation. Design: A qualitative design with thematic framework analysis of data collected in semistructured interviews. Setting/participants: Research was conducted in seven specialist in-patient palliative care services in Belgium, Germany, Italy, the Netherlands, and Spain. A bereaved relative and health care provider linked to a deceased patient with cancer who had palliative sedation was recruited. Measurements: We coded transcripts and characterized factors that arose during the process of palliative sedation. Results: We interviewed 66 people (33 relatives and 33 health care professionals) linked to 33 deceased patients. Three main themes were identified: (1) Understanding the aim of palliative sedation among relatives and health care professionals such as to alleviate suffering, dying with dignity, (2) Palliative sedation is a complex process, accompanied by many uncertainties, which can cause distress for both relatives and health care providers such as eligibility of the patient, when to start or the effectiveness of palliative sedation, (3) Sedation involves a period of intense family communication, collaboration, and caregiving with heightened mixed emotions. Opportunities to say goodbye before starting palliative sedation were important. Conclusion: Relatives and health care professionals focused on the effectiveness of palliative sedation in alleviating suffering and offering a dignified death. It was described as complex, with mixed experiences of relief and distress.

Background: Patient-provider interactions could inform care quality and communication but are rarely leveraged because collecting and analyzing them is both time-consuming and methodologically complex. The growing availability of large language models (LLMs) makes these analyses more feasible, though their accuracy remains uncertain. Objectives: Assess an LLM's ability to analyze patient-provider interactions. Design: Compare a human's and an LLM's codings of clinical encounter transcripts. Setting/Subjects: Two hundred and thirty-six potential symptom discussions from transcripts of clinical encounters with 92 patients living with cancer in the mid-Atlantic United States. Transcripts were analyzed by GPT4DFCI in our hospital's Health Insurance Portability and Accountability Act compliant infrastructure instance of GPT-4 (OpenAI). Measurements: Human and an LLM-coded transcripts to determine whether a patient's reported symptom(s) were discussed, who initiated the discussion, and any resulting recommendation. We calculated Cohen's κ to assess interrater agreement between the LLM and human and qualitatively classified disagreements about recommendations. Results: Interrater reliability indicated "strong" and "moderate" agreement levels across measures: Agreement was strongest for whether the symptom was discussed (k = 0.89), followed by who initiated the discussion (k = 0.82), and the recommendation provided (k = 0.78). The human and LLM disagreed on the presence and/or content of the recommendation in 16% of potential discussions, which we categorized into nine types of disagreements. Conclusions: Our results suggest that LLMs' abilities to analyze clinical encounters are equivalent to humans. Thus, using LLMs as a research tool may make it more feasible to analyze patient-provider interactions, which could have broader implications for assessing and improving care quality, care inequities, and provider communication.

Background: Despite caring for patients with serious illnesses, gastroenterology (GI) fellows rarely receive training in serious illness conversations (SIC). Objectives: To describe the development, implementation, and assessment of GITalk, a novel SIC training for GI fellows. Design: GITalk was based on the REMAP framework and involved two simulated encounters. One case involved a patient with decompensated cirrhosis, and the other case was about feeding tube placement in someone with moderate to severe dementia. Setting and Subjects: GI fellows in an academic medical center in the USA. Measurements: Demographics of the participants, evaluation of the course content, and self-assessed preparedness for SIC. Results: A total of 23 GI fellows participated over 4 consecutive years. Participants had significantly higher mean post-training self-assessed preparedness scores compared to pre-training across all 9 survey questions. 91% of participants strongly agreed with the statement: "I would recommend this training to other fellows." Conclusions: Participants in GITalk reported substantial improvement in self-assessed preparedness for navigating SIC.

The Journal of Palliative Medicine's "Tell Us More: The Palliative Care Oral History Project," seeks to tell the story of Hospice and Palliative Care through informal interviews with pivotal leaders in the field. In each episode, hosts Dr. Ricky Leiter, Dr. Billy Rosa, and research assistant Dr. Yilong Peng sit down with an HAPC luminary and do what our field does best-ask questions, listen, and reflect. In the first episode, Drs. Leiter and Rosa interviewed Dr. Charles von Gunten, Clinical Professor of Medicine at the University of California, San Diego, and Editor Emeritus of JPM. What follows is a transcript of their conversation, edited lightly for clarity.

Background and Objective: Palliative care (PC) has demonstrated the ability to improve patient outcomes and quality of life among patients with life-threatening diseases. The current study aimed to investigate the impact of inpatient PC on health care utilization and expenditure among patients with advanced gastrointestinal (GI) cancer. Methods: Patients diagnosed with advanced GI cancer (2007-2019) were identified from the SEER-Medicare database in the United States. Multivariable regression with entropy balancing was used to analyze the association between inpatient PC and outcomes, including readmission, length of stay (LOS), and expenditure. Results: Among 48,100 patients diagnosed with advanced GI cancer (colon: n = 23,080, 48.0%; pancreas: n = 12,280, 25.5%; rectum: n = 6497, 13.5%; biliary: n = 3551, 7.3%; liver: n = 2692, 5.6%), 1277 (2.65%) received PC. Median patient age was 77 (72-83) with most being female (n = 25,687, 53.4%). Patients with PC were more likely to be discharged to skilled nursing facility (SNF) (42.3% vs. 17.9%) and less likely to get readmitted within 30 days (18.2% vs. 28.1%). On adjusted analysis, patients with PC had higher costs at index admission (mean difference [β]: $1,494, 95% confidence interval [CI] $1,394-$1,594) but lower 90-day expenditure (β: -$3,037, 95% CI: -$3,279 to -$2,796). PC was also linked with lower odds of readmission (odds ratio 0.39, 95% CI: 0.33-0.46) and cumulative 90-day LOS (β: -1.31, 95% CI: -1.62 to -0.99). Conclusion: Following inpatient PC, patients with advanced GI cancer experienced fewer readmissions, days in hospital, and lower costs. Integrating PC into cancer care is vital to enhance patient outcomes while alleviating the strain on health care resources.