|
08:30
10:00
|
-
Yen-Ta Huang
Speaker
From Guidelines to Bayesian Trials: How Should We Interpret Polymyxin B Hemoadsorption in Sepsis?Polymyxin B hemoadsorption has occupied an unusually contested place in sepsis care, and the way authoritative bodies describe it has shifted over time. This talk traces that evolution—from the 2020 Taiwan AKI consensus through the 2021 and 2026 Surviving Sepsis Campaign guidelines—and uses the changing recommendations as a lens on the methodological debates that underlie them.
From an evidence-based medicine standpoint, the controversy is less about the device than about how we generate and grade evidence. I will first examine the frequentist evidence base, including our network meta-analysis of blood purification therapies for severe infection and sepsis/septic shock published in Critical Care Medicine. I will outline what a network meta-analysis adds—coherent ranking across multiple modalities, indirect comparison, and explicit quantification of heterogeneity—while being candid about its limitations, including reliance on the transitivity assumption, sparse networks, and the fragility of mortality estimates pooled from heterogeneous trials.
I will then turn to the recently reported TIGRIS trial, which used a Bayesian, enrichment-based phase 3 design with prespecified borrowing from the EUPHRATES treatable cohort. I will highlight the strengths of the Bayesian presentation—direct posterior probabilities of benefit, transparent incorporation of prior evidence, and efficiency in a rare, mechanistically defined phenotype—alongside the cautions it demands: sensitivity to prior choice, the open-label design, and the gap between a high posterior probability and a confirmed effect.
Bringing these strands together, the talk offers a multifaceted, methodology-driven reading of polymyxin B hemoadsorption in sepsis, and a transferable framework for clinicians facing the increasingly common situation in which guidelines, frequentist syntheses, and Bayesian trials appear to diverge.從治療執行到品質治理：連結式重症腎臟照護與資料驅動決策
-
Competency Frameworks in Critical Care Nephrology: Milestones and Entrustable Professional Activities (EPAs)
WEIHUNG Lin
Speaker
重症腎臟醫學的能力架構：里程碑與 EPAs 重症腎臟醫學的能力架構：里程碑與可信賴專業活動（EPAs）
林威宏醫師
醫學系副系主任｜教學中心主任｜內科部副部長｜一般內科主任
成功大學醫學院附設醫院
重症腎臟醫學要求臨床醫師在高度不確定、時間敏感且跨專業的情境中，整合急性腎損傷、腎替代治療、電解質與酸鹼異常、液體與血流動力學，以及倫理與溝通等多重能力。傳統以訓練年資、授課主題或病例數為主的課程，往往無法回答最關鍵的問題：受訓醫師究竟能在何種督導程度下，安全且一致地完成真實臨床任務？
本演講將介紹一套適用於重症醫學專科訓練的「重症腎臟醫學核心能力框架」草案。此框架以能力導向醫學教育為基礎，將 ACGME 六大核心能力整合至六個臨床領域，建構18項可信賴專業活動（entrustable professional activities, EPAs）及五階段里程碑。L4定義為完訓時可被信賴執行的目標層級；在臺灣ICU主治醫師覆核文化下，對應於間接督導及次日覆核，L5則代表完訓後的進階專家發展。七項涉及病人安全的核心任務被列為must-pass EPAs，包括腎臟替代療法的適應症與時機、CRRT處方、CRRT劑量調整與治療藥物監測、危急電解質異常、以床邊超音波及VExUS評估液體反應性與靜脈充血、敗血症相關AKI，以及腎臟替代療法撤除與緩和醫療。
評量採多工具、縱向及多來源的整合式評量，結合病例討論、mini-CEX、DOPS、模擬式OSCE、360度回饋與學習歷程；罕見但高風險的臨床情境，可由模擬、病歷回顧或結構式病例討論補足。建議的完訓門檻為七項must-pass EPAs全數達L4、18項中至少15項達L4，其餘至少達L3，並完成跨團隊品質改善成果。
此框架將訓練焦點由「教過多少、做過幾例」轉向「能否在適當督導下可靠完成臨床工作」，並透過可觀察表現、持續回饋與縱向進展，支持可辯護的信賴決策，同時為後續AI輔助學習與可信賴教育系統提供由臨床專業所界定的能力基礎。
關鍵詞：能力導向醫學教育、重症腎臟醫學、可信賴專業活動、里程碑、整合式評量
Competency Frameworks in Critical Care Nephrology: Milestones and Entrustable Professional Activities (EPAs)
Wei-Hung Lin, MD, PhD
Vice Chair, Department of Medicine, College of Medicine, National Cheng Kung University
Director, Education Center
Deputy Director, Department of Internal Medicine
Chief, Division of General Internal Medicine
National Cheng Kung University Hospital (NCKUH), College of Medicine, National Cheng Kung University
Critical care nephrology requires clinicians to integrate acute kidney injury, renal replacement therapy, electrolyte and acid–base disorders, fluid and hemodynamic management, and ethically complex communication in time-sensitive, uncertain, and interprofessional settings. Traditional curricula based on training duration, covered topics, or case numbers do not adequately answer a fundamental question: What clinical work can a trainee be trusted to perform, and at what level of supervision?
This lecture introduces a proposed competency-based framework for postgraduate training in critical care nephrology. The framework maps the six Accreditation Council for Graduate Medical Education core competencies across six clinical domains and organizes authentic workplace practice into 18 entrustable professional activities (EPAs) with five developmental milestone levels. Level 4 (L4) represents the graduation target—interpreted in the Taiwanese ICU context as indirect supervision with next-day attending review—whereas L5 denotes aspirational post-training expertise. Seven safety-critical EPAs require achievement of L4: determining the indications and timing of renal replacement therapy; prescribing continuous renal replacement therapy (CRRT); adjusting CRRT dose and therapeutic drug monitoring; managing life-threatening electrolyte disorders; assessing fluid responsiveness and venous congestion using point-of-care ultrasound (POCUS) and venous excess ultrasound (VExUS); managing sepsis-associated AKI; and leading decisions on withdrawal of renal replacement therapy and palliative care.
Assessment is programmatic, longitudinal, and multisource, combining case-based discussion, mini-CEX, direct observation of procedural skills, simulation-based OSCEs, multisource feedback, and a longitudinal portfolio. Simulation, chart review, and structured case conferences can address low-frequency, high-risk clinical exposures. Proposed graduation standards require all seven must-pass EPAs and at least 15 of 18 EPAs to reach L4, with the remainder at L3 or above, plus documented interprofessional quality-improvement work.
By shifting the focus from content coverage and case counting to observable performance, feedback, progression, and defensible entrustment decisions, the framework provides a clinically defined foundation for subsequent AI-enabled learning and trustworthy education systems.
Keywords: competency-based medical education; critical care nephrology; entrustable professional activities; milestones; programmatic assessment
Yen-Ta Huang
Moderator
From Guidelines to Bayesian Trials: How Should We Interpret Polymyxin B Hemoadsorption in Sepsis?Polymyxin B hemoadsorption has occupied an unusually contested place in sepsis care, and the way authoritative bodies describe it has shifted over time. This talk traces that evolution—from the 2020 Taiwan AKI consensus through the 2021 and 2026 Surviving Sepsis Campaign guidelines—and uses the changing recommendations as a lens on the methodological debates that underlie them.
From an evidence-based medicine standpoint, the controversy is less about the device than about how we generate and grade evidence. I will first examine the frequentist evidence base, including our network meta-analysis of blood purification therapies for severe infection and sepsis/septic shock published in Critical Care Medicine. I will outline what a network meta-analysis adds—coherent ranking across multiple modalities, indirect comparison, and explicit quantification of heterogeneity—while being candid about its limitations, including reliance on the transitivity assumption, sparse networks, and the fragility of mortality estimates pooled from heterogeneous trials.
I will then turn to the recently reported TIGRIS trial, which used a Bayesian, enrichment-based phase 3 design with prespecified borrowing from the EUPHRATES treatable cohort. I will highlight the strengths of the Bayesian presentation—direct posterior probabilities of benefit, transparent incorporation of prior evidence, and efficiency in a rare, mechanistically defined phenotype—alongside the cautions it demands: sensitivity to prior choice, the open-label design, and the gap between a high posterior probability and a confirmed effect.
Bringing these strands together, the talk offers a multifaceted, methodology-driven reading of polymyxin B hemoadsorption in sepsis, and a transferable framework for clinicians facing the increasingly common situation in which guidelines, frequentist syntheses, and Bayesian trials appear to diverge.從治療執行到品質治理：連結式重症腎臟照護與資料驅動決策
-
Training the Critical Care Nephrology in the AI Era: Redesigning Learning and Assessment
-
From AI Tools to a Trustworthy Education System
-
WEIHUNG Lin
Speaker
重症腎臟醫學的能力架構：里程碑與 EPAs 重症腎臟醫學的能力架構：里程碑與可信賴專業活動（EPAs）
林威宏醫師
醫學系副系主任｜教學中心主任｜內科部副部長｜一般內科主任
成功大學醫學院附設醫院
重症腎臟醫學要求臨床醫師在高度不確定、時間敏感且跨專業的情境中，整合急性腎損傷、腎替代治療、電解質與酸鹼異常、液體與血流動力學，以及倫理與溝通等多重能力。傳統以訓練年資、授課主題或病例數為主的課程，往往無法回答最關鍵的問題：受訓醫師究竟能在何種督導程度下，安全且一致地完成真實臨床任務？
本演講將介紹一套適用於重症醫學專科訓練的「重症腎臟醫學核心能力框架」草案。此框架以能力導向醫學教育為基礎，將 ACGME 六大核心能力整合至六個臨床領域，建構18項可信賴專業活動（entrustable professional activities, EPAs）及五階段里程碑。L4定義為完訓時可被信賴執行的目標層級；在臺灣ICU主治醫師覆核文化下，對應於間接督導及次日覆核，L5則代表完訓後的進階專家發展。七項涉及病人安全的核心任務被列為must-pass EPAs，包括腎臟替代療法的適應症與時機、CRRT處方、CRRT劑量調整與治療藥物監測、危急電解質異常、以床邊超音波及VExUS評估液體反應性與靜脈充血、敗血症相關AKI，以及腎臟替代療法撤除與緩和醫療。
評量採多工具、縱向及多來源的整合式評量，結合病例討論、mini-CEX、DOPS、模擬式OSCE、360度回饋與學習歷程；罕見但高風險的臨床情境，可由模擬、病歷回顧或結構式病例討論補足。建議的完訓門檻為七項must-pass EPAs全數達L4、18項中至少15項達L4，其餘至少達L3，並完成跨團隊品質改善成果。
此框架將訓練焦點由「教過多少、做過幾例」轉向「能否在適當督導下可靠完成臨床工作」，並透過可觀察表現、持續回饋與縱向進展，支持可辯護的信賴決策，同時為後續AI輔助學習與可信賴教育系統提供由臨床專業所界定的能力基礎。
關鍵詞：能力導向醫學教育、重症腎臟醫學、可信賴專業活動、里程碑、整合式評量
Competency Frameworks in Critical Care Nephrology: Milestones and Entrustable Professional Activities (EPAs)
Wei-Hung Lin, MD, PhD
Vice Chair, Department of Medicine, College of Medicine, National Cheng Kung University
Director, Education Center
Deputy Director, Department of Internal Medicine
Chief, Division of General Internal Medicine
National Cheng Kung University Hospital (NCKUH), College of Medicine, National Cheng Kung University
Critical care nephrology requires clinicians to integrate acute kidney injury, renal replacement therapy, electrolyte and acid–base disorders, fluid and hemodynamic management, and ethically complex communication in time-sensitive, uncertain, and interprofessional settings. Traditional curricula based on training duration, covered topics, or case numbers do not adequately answer a fundamental question: What clinical work can a trainee be trusted to perform, and at what level of supervision?
This lecture introduces a proposed competency-based framework for postgraduate training in critical care nephrology. The framework maps the six Accreditation Council for Graduate Medical Education core competencies across six clinical domains and organizes authentic workplace practice into 18 entrustable professional activities (EPAs) with five developmental milestone levels. Level 4 (L4) represents the graduation target—interpreted in the Taiwanese ICU context as indirect supervision with next-day attending review—whereas L5 denotes aspirational post-training expertise. Seven safety-critical EPAs require achievement of L4: determining the indications and timing of renal replacement therapy; prescribing continuous renal replacement therapy (CRRT); adjusting CRRT dose and therapeutic drug monitoring; managing life-threatening electrolyte disorders; assessing fluid responsiveness and venous congestion using point-of-care ultrasound (POCUS) and venous excess ultrasound (VExUS); managing sepsis-associated AKI; and leading decisions on withdrawal of renal replacement therapy and palliative care.
Assessment is programmatic, longitudinal, and multisource, combining case-based discussion, mini-CEX, direct observation of procedural skills, simulation-based OSCEs, multisource feedback, and a longitudinal portfolio. Simulation, chart review, and structured case conferences can address low-frequency, high-risk clinical exposures. Proposed graduation standards require all seven must-pass EPAs and at least 15 of 18 EPAs to reach L4, with the remainder at L3 or above, plus documented interprofessional quality-improvement work.
By shifting the focus from content coverage and case counting to observable performance, feedback, progression, and defensible entrustment decisions, the framework provides a clinically defined foundation for subsequent AI-enabled learning and trustworthy education systems.
Keywords: competency-based medical education; critical care nephrology; entrustable professional activities; milestones; programmatic assessment
Yen-Ta Huang
Speaker
From Guidelines to Bayesian Trials: How Should We Interpret Polymyxin B Hemoadsorption in Sepsis?Polymyxin B hemoadsorption has occupied an unusually contested place in sepsis care, and the way authoritative bodies describe it has shifted over time. This talk traces that evolution—from the 2020 Taiwan AKI consensus through the 2021 and 2026 Surviving Sepsis Campaign guidelines—and uses the changing recommendations as a lens on the methodological debates that underlie them.
From an evidence-based medicine standpoint, the controversy is less about the device than about how we generate and grade evidence. I will first examine the frequentist evidence base, including our network meta-analysis of blood purification therapies for severe infection and sepsis/septic shock published in Critical Care Medicine. I will outline what a network meta-analysis adds—coherent ranking across multiple modalities, indirect comparison, and explicit quantification of heterogeneity—while being candid about its limitations, including reliance on the transitivity assumption, sparse networks, and the fragility of mortality estimates pooled from heterogeneous trials.
I will then turn to the recently reported TIGRIS trial, which used a Bayesian, enrichment-based phase 3 design with prespecified borrowing from the EUPHRATES treatable cohort. I will highlight the strengths of the Bayesian presentation—direct posterior probabilities of benefit, transparent incorporation of prior evidence, and efficiency in a rare, mechanistically defined phenotype—alongside the cautions it demands: sensitivity to prior choice, the open-label design, and the gap between a high posterior probability and a confirmed effect.
Bringing these strands together, the talk offers a multifaceted, methodology-driven reading of polymyxin B hemoadsorption in sepsis, and a transferable framework for clinicians facing the increasingly common situation in which guidelines, frequentist syntheses, and Bayesian trials appear to diverge.從治療執行到品質治理：連結式重症腎臟照護與資料驅動決策
-
Room 103
|
