Pub Date : 2026-02-01Epub Date: 2026-02-03DOI: 10.1097/CCM.0000000000006985
Lone Musaeus Poulsen, Ole Mathiesen, Bodil Steen Rasmussen, Stine Estrup, Anders Granholm, Lars Peter Kloster Andersen, Nina Andersen-Ranberg, Camilla Bekker Mortensen, Marie Oxenbøll Collet
{"title":"The authors reply.","authors":"Lone Musaeus Poulsen, Ole Mathiesen, Bodil Steen Rasmussen, Stine Estrup, Anders Granholm, Lars Peter Kloster Andersen, Nina Andersen-Ranberg, Camilla Bekker Mortensen, Marie Oxenbøll Collet","doi":"10.1097/CCM.0000000000006985","DOIUrl":"https://doi.org/10.1097/CCM.0000000000006985","url":null,"abstract":"","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":"54 2","pages":"374-375"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146112733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Objective: To assess the impact of the presence of an intermediate care unit (IMCU) on ICU patient-level clinical outcomes as well as hospital-level utilization and cost measures by comparing those with vs without an IMCU, using Japan's nationwide inpatient database.
Design: Nationwide retrospective cohort study.
Setting: Acute-care hospitals in Japan participating in the Diagnosis Procedure Combination Study Group database and Hospital Bed Function Reports from 2016 to 2022.
Patients: A total of 2,278,521 adult patients admitted to the ICU between April 2016 and March 2023.
Interventions: ICU admission to hospitals with vs without an IMCU.
Measurements and main results: Outcomes included patient-level (in-hospital and ICU mortality, ICU readmission, and length of ICU stay) and hospital-year level (resource utilization and hospitalization costs) measures. Among the 2,278,521 eligible ICU patients from 557 hospitals across 2,953 hospital-years, 1,771,000 (77.7%) patients were admitted to hospitals with both an ICU and IMCU. Overall, 14.3% of patients were transferred between the ICU and IMCU, with large variability between hospitals. When estimating participant-average treatment effect for patient-level outcomes, ICU patients in IMCU-equipped hospitals had lower in-hospital mortality (adjusted odds ratio [aOR] 0.94; 95% CI, 0.89-0.99), ICU mortality (aOR 0.87; 0.83-0.92), fewer ICU readmissions (aOR 0.92; 95% CI, 0.85-1.00), and shorter ICU stays (adjusted rate ratio 0.98; 95% CI, 0.98-0.99). When estimating cluster-average treatment effect for hospital-level outcomes, IMCU-equipped hospitals had higher ICU bed occupancy (mean difference: 5.5%, 95% CI, 3.3-7.7%), higher occupancy for life-sustaining therapies (8.6%, 95% CI, 7.7-9.5%), increased reimbursement rates (5.4%, 95% CI, 4.0-6.8%), and increased annual revenue per ICU bed (25 million JPY, 19-31 million JPY).
Conclusions: The presence of an IMCU in ICU-equipped hospitals was associated with improved patient outcomes and more efficient ICU utilization, with only modest increase in hospitalization costs. These findings support integration of the IMCU into critical care systems.
{"title":"Outcome, Process, Utilization, and Cost Measurements of Patients Admitted to the ICU in Hospitals With Vs. Without an Intermediate Care Unit: A Nationwide Inpatient Database Study.","authors":"Hiroyuki Ohbe, Daisuke Kudo, Yuya Kimura, Hiroki Matsui, Kiyohide Fushimi, Hideo Yasunaga, Shigeki Kushimoto","doi":"10.1097/CCM.0000000000006962","DOIUrl":"10.1097/CCM.0000000000006962","url":null,"abstract":"<p><strong>Objective: </strong>To assess the impact of the presence of an intermediate care unit (IMCU) on ICU patient-level clinical outcomes as well as hospital-level utilization and cost measures by comparing those with vs without an IMCU, using Japan's nationwide inpatient database.</p><p><strong>Design: </strong>Nationwide retrospective cohort study.</p><p><strong>Setting: </strong>Acute-care hospitals in Japan participating in the Diagnosis Procedure Combination Study Group database and Hospital Bed Function Reports from 2016 to 2022.</p><p><strong>Patients: </strong>A total of 2,278,521 adult patients admitted to the ICU between April 2016 and March 2023.</p><p><strong>Interventions: </strong>ICU admission to hospitals with vs without an IMCU.</p><p><strong>Measurements and main results: </strong>Outcomes included patient-level (in-hospital and ICU mortality, ICU readmission, and length of ICU stay) and hospital-year level (resource utilization and hospitalization costs) measures. Among the 2,278,521 eligible ICU patients from 557 hospitals across 2,953 hospital-years, 1,771,000 (77.7%) patients were admitted to hospitals with both an ICU and IMCU. Overall, 14.3% of patients were transferred between the ICU and IMCU, with large variability between hospitals. When estimating participant-average treatment effect for patient-level outcomes, ICU patients in IMCU-equipped hospitals had lower in-hospital mortality (adjusted odds ratio [aOR] 0.94; 95% CI, 0.89-0.99), ICU mortality (aOR 0.87; 0.83-0.92), fewer ICU readmissions (aOR 0.92; 95% CI, 0.85-1.00), and shorter ICU stays (adjusted rate ratio 0.98; 95% CI, 0.98-0.99). When estimating cluster-average treatment effect for hospital-level outcomes, IMCU-equipped hospitals had higher ICU bed occupancy (mean difference: 5.5%, 95% CI, 3.3-7.7%), higher occupancy for life-sustaining therapies (8.6%, 95% CI, 7.7-9.5%), increased reimbursement rates (5.4%, 95% CI, 4.0-6.8%), and increased annual revenue per ICU bed (25 million JPY, 19-31 million JPY).</p><p><strong>Conclusions: </strong>The presence of an IMCU in ICU-equipped hospitals was associated with improved patient outcomes and more efficient ICU utilization, with only modest increase in hospitalization costs. These findings support integration of the IMCU into critical care systems.</p>","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":" ","pages":"234-245"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145488132","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01Epub Date: 2025-11-11DOI: 10.1097/CCM.0000000000006964
Khalid S Alotaibi, Naveed Saleem, Timothy Arthur Chandos Snow, Pietro Arina, Alex Dyson, Mervyn Singer
Objectives: To identify the impact of early hypothermic temperature control (HTC) initiation and achievement on neurologic outcomes after cardiac arrest. Secondary endpoints assessed: 1) overall survival, 2) body temperature at hospital admission, 3) time taken to reach targeted temperature, and 4) adverse events. Subgroup analyses assessed: 1) HTC induction method, 2) presenting rhythms (shockable vs. nonshockable), and 3) timing of HTC initiation (during cardiopulmonary resuscitation [CPR] vs. post-return of spontaneous circulation [ROSC]).
Data sources: PubMed, Cochrane Library, U.S. National Library of Medicine, MedRxiv, BioRxiv.
Study selection: Prospective randomized controlled trials enrolling comatose adult cardiac arrest patients that assessed the efficacy of early cooling (defined by initiation of HTC within 30 min of ROSC) against late cooling.
Data extraction: Two reviewers independently conducted study selection, data extraction, and assessment of evidence quality. Full texts were jointly reviewed, with discrepancies resolved by a third reviewer through discussion and consensus.
Data synthesis: Of 7269 citations, 11 met the inclusion criteria. Early cooling did not improve neurologic outcomes (risk ratio [RR], 1.01; 95% CI, 0.94-1.14) nor survival (RR, 1.01; 95% CI, 0.92-1.11). Nasal evaporative cooling initiated during resuscitation may benefit patients with shockable rhythms (RR, 1.40; 95% CI, 1.00-1.96), while administering cold fluid during CPR was associated with a higher risk of rearrest. Cooling was initiated in the prehospital phase in ten studies, nine of which failed to achieve target temperature (32-34°C) by hospital admission. The median time to reach target temperature ranged from 38 to 360 minutes.
Conclusions: Early HTC showed no neurologic nor survival benefit in cardiac arrest survivors. Nasal evaporative cooling during CPR may benefit those with shockable rhythms. Delays in reaching the target temperature may underlie the failure of HTC. Future studies should evaluate whether achieving target temperature, for example, within 30 minutes, rapidly improves outcomes.
{"title":"Does Time to Achieve a Targeted Body Temperature Matter for Survivors of Cardiac Arrest? A Systematic Review and Meta-Analysis.","authors":"Khalid S Alotaibi, Naveed Saleem, Timothy Arthur Chandos Snow, Pietro Arina, Alex Dyson, Mervyn Singer","doi":"10.1097/CCM.0000000000006964","DOIUrl":"10.1097/CCM.0000000000006964","url":null,"abstract":"<p><strong>Objectives: </strong>To identify the impact of early hypothermic temperature control (HTC) initiation and achievement on neurologic outcomes after cardiac arrest. Secondary endpoints assessed: 1) overall survival, 2) body temperature at hospital admission, 3) time taken to reach targeted temperature, and 4) adverse events. Subgroup analyses assessed: 1) HTC induction method, 2) presenting rhythms (shockable vs. nonshockable), and 3) timing of HTC initiation (during cardiopulmonary resuscitation [CPR] vs. post-return of spontaneous circulation [ROSC]).</p><p><strong>Data sources: </strong>PubMed, Cochrane Library, U.S. National Library of Medicine, MedRxiv, BioRxiv.</p><p><strong>Study selection: </strong>Prospective randomized controlled trials enrolling comatose adult cardiac arrest patients that assessed the efficacy of early cooling (defined by initiation of HTC within 30 min of ROSC) against late cooling.</p><p><strong>Data extraction: </strong>Two reviewers independently conducted study selection, data extraction, and assessment of evidence quality. Full texts were jointly reviewed, with discrepancies resolved by a third reviewer through discussion and consensus.</p><p><strong>Data synthesis: </strong>Of 7269 citations, 11 met the inclusion criteria. Early cooling did not improve neurologic outcomes (risk ratio [RR], 1.01; 95% CI, 0.94-1.14) nor survival (RR, 1.01; 95% CI, 0.92-1.11). Nasal evaporative cooling initiated during resuscitation may benefit patients with shockable rhythms (RR, 1.40; 95% CI, 1.00-1.96), while administering cold fluid during CPR was associated with a higher risk of rearrest. Cooling was initiated in the prehospital phase in ten studies, nine of which failed to achieve target temperature (32-34°C) by hospital admission. The median time to reach target temperature ranged from 38 to 360 minutes.</p><p><strong>Conclusions: </strong>Early HTC showed no neurologic nor survival benefit in cardiac arrest survivors. Nasal evaporative cooling during CPR may benefit those with shockable rhythms. Delays in reaching the target temperature may underlie the failure of HTC. Future studies should evaluate whether achieving target temperature, for example, within 30 minutes, rapidly improves outcomes.</p>","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":" ","pages":"312-323"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145488148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01Epub Date: 2025-12-09DOI: 10.1097/CCM.0000000000006978
Samuel A Tisherman, Antoinette Spevetz, J Christopher Farmer, Rahul Kashyap, Ed Michener, Stefan W Leichtle, Roshni Sreedharan, Samantha S Strickler, Cherylee W J Chang, Angel Coz Yataco, Joseph Cuschieri, David Dudzinski, Timothy Ellender, Lillian L Emlet, Brenda G Fahy, Bram J Geller, Erin Hennessey, Krista L Kaups, Mark T Keegan, May M Lee, Deepa Malaiyandi, Christopher P Michetti, Nicholas M Mohr, Vivek K Moitra, Kevin O'Neil, Tina L Palmieri, Pauline K Park, Abhijit Pathak, P B Raksin, Angela Hays Shapshak, Shahla Siddiqui, Deborah M Stein, Shelly D Timmons, Paul M Vespa, Brian T Wessman, Erica D Wittwer
Objectives: In the United States, training for physicians who manage critically ill adult patients (intensivists) evolved through parallel subspecialty critical care medicine (CCM) pathways with significant commonality. The Society of Critical Care Medicine Adult Critical Care Physician Core Knowledge and Skills Task Force aimed to delineate the common core knowledge and skills required of all intensivists.
Design: A master list of content areas and procedural skills was compiled from all CCM subspecialty program requirements and blueprints of the certification examinations. Using a modified Delphi approach, participants were asked to categorize the knowledge items as "advanced knowledge is essential," "general, but not advanced, knowledge is essential," or "knowledge is not essential." Procedures were categorized as "intensivist performs routinely," "intensivist only performs in an emergency," or "intensivist knows" about the procedure.
Setting: Representatives from CCM stakeholder organizations, including accreditation and certification organizations, critical care societies, and program directors' societies, were invited to participate.
Subjects: Members of the Adult Critical Care Physician Core Knowledge and Skills Task Force of the Society of Critical Care Medicine.
Interventions: For the first two rounds of the modified Delphi process, Research Electronic Data Capture was used. For the third and fourth rounds, the process was completed through online meetings with Zoom (Zoom Video Corporations, San Jose, CA) utilizing Zoom's polling feature.
Measurements and main results: A total of 541 items were determined to be essential, with 145 requiring advanced knowledge and 323 requiring general knowledge. For 73 items, consensus regarding advanced vs. general could not be achieved, but they remained essential. Only eight items were felt to be nonessential. Of the 16 procedures, most were categorized as "intensivist performs."
Conclusions: The large number of items included in the list of essential knowledge and skills demonstrates the complexity of modern CCM. Utilization of a common framework across the subspecialties of CCM could lead to greater harmonization among the fellowship program requirements and certification examinations.
目的:在美国,管理重症成人患者的医生(重症医师)的培训是通过平行的亚专科重症医学(CCM)途径发展起来的,具有显著的共性。重症医学学会成人重症医师核心知识和技能工作组旨在描述所有重症医师所需的共同核心知识和技能。设计:从所有CCM子专业项目要求和认证考试蓝图中编制了内容领域和程序技能的主列表。使用改进的德尔菲方法,参与者被要求将知识项目分类为“高级知识是必要的”,“一般但不高级,知识是必要的”或“知识不是必要的”。程序被分类为“重症医师例行执行”、“重症医师只在紧急情况下执行”或“重症医师知道”该程序。环境:CCM利益相关者组织的代表被邀请参加,包括认证和认证组织、重症监护协会和项目主任协会。受试者:重症医学学会成人重症监护医师核心知识和技能工作组成员。干预措施:在前两轮改进的德尔菲过程中,使用了Research Electronic Data Capture。第三轮和第四轮的过程是通过Zoom (Zoom Video corporation, San Jose, CA)利用Zoom的投票功能与Zoom进行在线会议完成的。测量和主要结果:共确定了541项基本知识,其中145项需要高级知识,323项需要一般知识。在73个项目中,无法就先进与一般达成共识,但它们仍然至关重要。只有8个项目被认为是不必要的。在这16项手术中,大多数被归类为“强化手术”。结论:基本知识和技能清单中包含的项目数量之多显示了现代CCM的复杂性。在CCM的子专业之间使用一个共同的框架可以使奖学金计划的要求和认证考试更加协调一致。
{"title":"Determination of Adult Critical Care Physician Core Knowledge and Skills: Results of a Multidisciplinary, Modified Delphi Process.","authors":"Samuel A Tisherman, Antoinette Spevetz, J Christopher Farmer, Rahul Kashyap, Ed Michener, Stefan W Leichtle, Roshni Sreedharan, Samantha S Strickler, Cherylee W J Chang, Angel Coz Yataco, Joseph Cuschieri, David Dudzinski, Timothy Ellender, Lillian L Emlet, Brenda G Fahy, Bram J Geller, Erin Hennessey, Krista L Kaups, Mark T Keegan, May M Lee, Deepa Malaiyandi, Christopher P Michetti, Nicholas M Mohr, Vivek K Moitra, Kevin O'Neil, Tina L Palmieri, Pauline K Park, Abhijit Pathak, P B Raksin, Angela Hays Shapshak, Shahla Siddiqui, Deborah M Stein, Shelly D Timmons, Paul M Vespa, Brian T Wessman, Erica D Wittwer","doi":"10.1097/CCM.0000000000006978","DOIUrl":"10.1097/CCM.0000000000006978","url":null,"abstract":"<p><strong>Objectives: </strong>In the United States, training for physicians who manage critically ill adult patients (intensivists) evolved through parallel subspecialty critical care medicine (CCM) pathways with significant commonality. The Society of Critical Care Medicine Adult Critical Care Physician Core Knowledge and Skills Task Force aimed to delineate the common core knowledge and skills required of all intensivists.</p><p><strong>Design: </strong>A master list of content areas and procedural skills was compiled from all CCM subspecialty program requirements and blueprints of the certification examinations. Using a modified Delphi approach, participants were asked to categorize the knowledge items as \"advanced knowledge is essential,\" \"general, but not advanced, knowledge is essential,\" or \"knowledge is not essential.\" Procedures were categorized as \"intensivist performs routinely,\" \"intensivist only performs in an emergency,\" or \"intensivist knows\" about the procedure.</p><p><strong>Setting: </strong>Representatives from CCM stakeholder organizations, including accreditation and certification organizations, critical care societies, and program directors' societies, were invited to participate.</p><p><strong>Subjects: </strong>Members of the Adult Critical Care Physician Core Knowledge and Skills Task Force of the Society of Critical Care Medicine.</p><p><strong>Interventions: </strong>For the first two rounds of the modified Delphi process, Research Electronic Data Capture was used. For the third and fourth rounds, the process was completed through online meetings with Zoom (Zoom Video Corporations, San Jose, CA) utilizing Zoom's polling feature.</p><p><strong>Measurements and main results: </strong>A total of 541 items were determined to be essential, with 145 requiring advanced knowledge and 323 requiring general knowledge. For 73 items, consensus regarding advanced vs. general could not be achieved, but they remained essential. Only eight items were felt to be nonessential. Of the 16 procedures, most were categorized as \"intensivist performs.\"</p><p><strong>Conclusions: </strong>The large number of items included in the list of essential knowledge and skills demonstrates the complexity of modern CCM. Utilization of a common framework across the subspecialties of CCM could lead to greater harmonization among the fellowship program requirements and certification examinations.</p>","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":" ","pages":"224-233"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145707287","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01Epub Date: 2025-12-09DOI: 10.1097/CCM.0000000000006980
Brett Abbenbroek, Balasubramanian Venkatesh
{"title":"Sepsis-Bridging the Gap Between Public Perception and Clinical Urgency.","authors":"Brett Abbenbroek, Balasubramanian Venkatesh","doi":"10.1097/CCM.0000000000006980","DOIUrl":"10.1097/CCM.0000000000006980","url":null,"abstract":"","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":" ","pages":"362-364"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145707491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01Epub Date: 2025-12-19DOI: 10.1097/CCM.0000000000006983
Dominique Piquette, Christie Lee
{"title":"Designing Critical Care Training for Real-World Practice: Capturing Complexity in a Meaningful and Useful Way.","authors":"Dominique Piquette, Christie Lee","doi":"10.1097/CCM.0000000000006983","DOIUrl":"10.1097/CCM.0000000000006983","url":null,"abstract":"","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":" ","pages":"354-357"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145793411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01Epub Date: 2026-02-03DOI: 10.1097/CCM.0000000000006943
Shiuan-Chih Chen, Ming-Cheng Lin
{"title":"Reconsidering Rapid Response Teams in Sepsis: Context-Dependent Value Beyond Mortality.","authors":"Shiuan-Chih Chen, Ming-Cheng Lin","doi":"10.1097/CCM.0000000000006943","DOIUrl":"https://doi.org/10.1097/CCM.0000000000006943","url":null,"abstract":"","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":"54 2","pages":"392-393"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146112695","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-01Epub Date: 2025-11-24DOI: 10.1097/CCM.0000000000006968
Luiz E V Silva, Hunter A Gaudio, Viveknarayanan Padmanabhan, Rodrigo M Forti, Lingyun Shi, McKenna Mason, Takayuki Sueishi, Matthew P Kirschen, Wesley B Baker, Todd J Kilbaugh, Ryan W Morgan, Fuchiang Rich Tsui, Tiffany S Ko
Objectives: To determine whether high-resolution (HighRes) and multimodal integration of physiologic signals improve prediction of return of spontaneous circulation (ROSC) during pediatric cardiopulmonary resuscitation (CPR) compared with low-resolution (LowRes) and single-modality approaches.
Design: Retrospective analysis of experimental data using machine learning models for outcome prediction.
Setting: Laboratory setting with pediatric swine models of cardiac arrest.
Subjects: A total of 187 pediatric swine undergoing standardized cardiac arrest and CPR protocols.
Interventions: Animals were monitored using multiple physiologic signals during CPR, including aortic blood pressure (ABP), right atrial pressure (RAP), capnography, and electrocardiography. No therapeutic interventions were evaluated.
Measurements and main results: Four data approaches were evaluated: 1) Waveform-HighRes (100 Hz waveforms); 2) Compression-HighRes (compression-by-compression physiologic series); 3) Waveform-LowRes (15-s averaged waveforms); and 4) Compression-LowRes (15-s averaged compression-by-compression series). Models were developed to predict ROSC using segments 2-4, 2-6, 2-8, and 2-10 minutes of CPR, using both single and combined signal modalities. Area under the receiver operating characteristic curve (AUROC) was used to evaluate models' performance. In early CPR (2-4 min), Compression-HighRes outperformed both LowRes approaches for ABP (AUROC, 0.74 [0.65-0.82] vs. 0.65 [0.55-0.74] and 0.54 [0.44-0.64]) and RAP (0.70 [0.62-0.79] vs. 0.61 [0.51-0.70] and 0.57 [0.48-0.66]; p < 0.05). In multimodal models, LowRes data performed comparably to HighRes models (AUROC, 0.76-0.79). Across time points, ABP-based model performance improved, reaching AUROC 0.90 (0.84-0.95) for the full CPR period (2-10 min)-comparable to the multimodal model (0.89 [0.83-0.95]).
Conclusions: HighRes monitoring improved early ROSC prediction for individual signals, especially ABP and RAP. However, combining multiple modalities compensates for lower resolution, enabling comparable predictive performance. These findings support data-driven strategies for selecting physiologic targets and technical requirements in physiology-directed CPR.
目的:确定与低分辨率(低分辨率)和单模态方法相比,高分辨率(HighRes)和多模态生理信号整合是否能改善小儿心肺复苏(CPR)期间自然循环恢复(ROSC)的预测。设计:使用机器学习模型对实验数据进行回顾性分析,以预测结果。环境:实验室环境与儿童猪心脏骤停模型。研究对象:共有187头儿科猪接受了标准化的心脏骤停和心肺复苏术治疗。干预措施:动物在心肺复苏术中使用多种生理信号进行监测,包括主动脉压(ABP)、右心房压(RAP)、血管造影和心电图。未评估任何治疗干预措施。测量和主要结果:评估了四种数据方法:1)波形-高分辨率(100 Hz波形);2)压缩-高分辨率(逐压缩生理系列);3)波形-低分辨率(15-s平均波形);4)压缩-低分辨率(15秒平均压缩-压缩系列)。使用单一和联合信号模式,开发了使用CPR 2-4、2-6、2-8和2-10分钟预测ROSC的模型。采用受试者工作特征曲线下面积(AUROC)评价模型的性能。在CPR早期(2-4分钟),Compression-HighRes优于ABP (AUROC, 0.74 [0.65-0.82] vs. 0.65[0.55-0.74]和0.54[0.44-0.64])和RAP (0.70 [0.62-0.79] vs. 0.61[0.51-0.70]和0.57 [0.48-0.66];p < 0.05)。在多模态模型中,低分辨率数据的表现与高分辨率模型相当(AUROC, 0.76-0.79)。跨时间点,基于abp的模型性能有所提高,在整个心肺复苏术期间(2-10分钟)达到AUROC 0.90(0.84-0.95),与多模态模型(0.89[0.83-0.95])相当。结论:高分辨率监测改善了个体信号的早期ROSC预测,尤其是ABP和RAP。然而,结合多种模式可以补偿较低的分辨率,从而实现可比较的预测性能。这些发现支持了生理导向CPR中选择生理靶点和技术要求的数据驱动策略。
{"title":"Machine Learning-Based Return of Spontaneous Circulation Prediction During Cardiopulmonary Resuscitation in a Swine Model of Cardiac Arrest: Effect of Data Resolution and Multimodal Physiological Waveforms.","authors":"Luiz E V Silva, Hunter A Gaudio, Viveknarayanan Padmanabhan, Rodrigo M Forti, Lingyun Shi, McKenna Mason, Takayuki Sueishi, Matthew P Kirschen, Wesley B Baker, Todd J Kilbaugh, Ryan W Morgan, Fuchiang Rich Tsui, Tiffany S Ko","doi":"10.1097/CCM.0000000000006968","DOIUrl":"https://doi.org/10.1097/CCM.0000000000006968","url":null,"abstract":"<p><strong>Objectives: </strong>To determine whether high-resolution (HighRes) and multimodal integration of physiologic signals improve prediction of return of spontaneous circulation (ROSC) during pediatric cardiopulmonary resuscitation (CPR) compared with low-resolution (LowRes) and single-modality approaches.</p><p><strong>Design: </strong>Retrospective analysis of experimental data using machine learning models for outcome prediction.</p><p><strong>Setting: </strong>Laboratory setting with pediatric swine models of cardiac arrest.</p><p><strong>Subjects: </strong>A total of 187 pediatric swine undergoing standardized cardiac arrest and CPR protocols.</p><p><strong>Interventions: </strong>Animals were monitored using multiple physiologic signals during CPR, including aortic blood pressure (ABP), right atrial pressure (RAP), capnography, and electrocardiography. No therapeutic interventions were evaluated.</p><p><strong>Measurements and main results: </strong>Four data approaches were evaluated: 1) Waveform-HighRes (100 Hz waveforms); 2) Compression-HighRes (compression-by-compression physiologic series); 3) Waveform-LowRes (15-s averaged waveforms); and 4) Compression-LowRes (15-s averaged compression-by-compression series). Models were developed to predict ROSC using segments 2-4, 2-6, 2-8, and 2-10 minutes of CPR, using both single and combined signal modalities. Area under the receiver operating characteristic curve (AUROC) was used to evaluate models' performance. In early CPR (2-4 min), Compression-HighRes outperformed both LowRes approaches for ABP (AUROC, 0.74 [0.65-0.82] vs. 0.65 [0.55-0.74] and 0.54 [0.44-0.64]) and RAP (0.70 [0.62-0.79] vs. 0.61 [0.51-0.70] and 0.57 [0.48-0.66]; p < 0.05). In multimodal models, LowRes data performed comparably to HighRes models (AUROC, 0.76-0.79). Across time points, ABP-based model performance improved, reaching AUROC 0.90 (0.84-0.95) for the full CPR period (2-10 min)-comparable to the multimodal model (0.89 [0.83-0.95]).</p><p><strong>Conclusions: </strong>HighRes monitoring improved early ROSC prediction for individual signals, especially ABP and RAP. However, combining multiple modalities compensates for lower resolution, enabling comparable predictive performance. These findings support data-driven strategies for selecting physiologic targets and technical requirements in physiology-directed CPR.</p>","PeriodicalId":10765,"journal":{"name":"Critical Care Medicine","volume":"54 2","pages":"257-269"},"PeriodicalIF":6.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146112685","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}