{
  "abstract": "Background Uncertainty around a patient’s prognosis at the end of life remains a major barrier to timely palliative-care involvement and alignment of treatments with patient goals. Artificial intelligence (AI)-based tools have recently emerged to provide structured mortality predictions and identify patients at risk of deterioration to support clinical decision-making.Objective This mini-review summarises recent literature evaluating AI-based prognostic and decision-support tools in end-of-life and palliative care, focusing on predictive accuracy, early implementation outcomes, communication effects and ethical considerations.Methods A set of peer-reviewed articles published between 2020 and 2025 was identified through a targeted narrative search of major medical databases. Included studies examined prognostic model validation, early-warning systems, implementation outcomes, communication impacts and ethical analyses.Results AI-based models consistently produced more accurate short-term mortality predictions than traditional scoring systems or clinician judgement. Early evidence also suggests that integrating AI decision support into clinical workflows can increase identification of patients and initiation of appropriate palliative care, including within generalist settings. However, data are limited regarding the impact of AI-based prognostic tools on treatment intensity or chemotherapy use. Preliminary qualitative work indicates that AI-generated summaries may assist communication among healthcare teams, though concerns persist regarding transparency, bias and over-reliance on algorithms.Conclusions AI-driven prognostic models show promise in improving risk identification and facilitating earlier engagement with palliative care. Nonetheless, the current evidence base is preliminary. Future research should include prospective trials and strengthened ethical frameworks to ensure that the integration of AI-based prognostic tools into end-of-life decision-making is both safe and equitable.",
  "authors": [
    {
      "affiliations": [
        "Saudi Board in Family Medicine, The Joint Program of Family Medicine in Jeddah, Saudi Arabia Ministry of Health, Jeddah, Makkah Province, Saudi Arabia",
        "Bupa CareConnect, Bupa Arabia for Cooperative Insurance, Jeddah, Makkah Province, Saudi Arabia"
      ],
      "name": "Abdullah Alabbasi"
    },
    {
      "affiliations": [
        "Saudi Arabia Ministry of Health, Jeddah, Makkah Province, Saudi Arabia"
      ],
      "name": "Muhanad Alzahrani"
    },
    {
      "affiliations": [
        "Department of Surgery, King Saud University College of Medicine, Riyadh, Riyadh Province, Saudi Arabia"
      ],
      "name": "Faris Sultan"
    },
    {
      "affiliations": [
        "Department of Medicine, University of Utah Health Hospitals and Clinics, Salt Lake City, Utah, USA"
      ],
      "name": "Mohammed Sayes"
    }
  ],
  "title": "Artificial intelligence-powered predictive tools to improve end-of-life decision-making: mini-review",
  "uid": "200e8bc1-618d-503a-a734-3878bb5c5c89"
}