ICU Cockpit: a platform for collecting multimodal waveform data, AI-based computational disease modeling and real-time decision support in the intensive care unit

Boss, J. M.; Lüthy, Rahel; Suter, Susanne K.; Muroi, Carl

doi:10.1093/jamia/ocac064

Cited by 16 publications

(11 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data Collection For the development of the OD models, we prospectively collected blurred anonymized video data from cameras (AXIS M1065-L) directed onto the bedsides of a 12bed neurocritical care unit at the University Hospital Zurich. The blurred video streams have a resolution of 640×400 pixels at 25 frames per second, collected by a dedicated research IT infrastructure [16]. The video data streams are blurred using a software solution for video stream conversion (FFmpeg, https://ffmpeg.org/) with a box blur filter (boxblur=6:1).…”

Section: Experimental Setup and Resultsmentioning

confidence: 99%

Video Object Detection for Privacy-Preserving Patient Monitoring in Intensive Care

Emberger,

Boss,

Baumann

et al. 2023

2023 10th IEEE Swiss Conference on Data Science (SDS)

Self Cite

View full text Add to dashboard Cite

Patient monitoring in intensive care units, although assisted by biosensors, needs continuous supervision of staff. To reduce the burden on staff members, IT infrastructures are built to record monitoring data and develop clinical decision support systems. These systems, however, are vulnerable to artifacts (e.g. muscle movement due to ongoing treatment), which are often indistinguishable from real and potentially dangerous signals.Video recordings could facilitate the reliable classification of biosignals using object detection (OD) methods to find sources of unwanted artifacts. Due to privacy restrictions, only blurred videos can be stored, which severely impairs the possibility to detect clinically relevant events such as interventions or changes in patient status with standard OD methods. Hence, new kinds of approaches are necessary that exploit every kind of available information due to the reduced information content of blurred footage and that are at the same time easily implementable within the IT infrastructure of a normal hospital. In this paper, we propose a new method for exploiting information in the temporal succession of video frames. To be efficiently implementable using off-the-shelf object detectors that comply with given hardware constraints, we repurpose the image color channels to account for temporal consistency, leading to an improved detection rate of the object classes. Our method outperforms a standard YOLOv5 baseline model by +1.7% mAP@.5 while also training over ten times faster on our proprietary dataset. We conclude that this approach has shown effectiveness in the preliminary experiments and holds potential for more general video OD in the future.

show abstract

Section: Experimental Setup and Resultsmentioning

confidence: 99%

Video Object Detection for Privacy-Preserving Patient Monitoring in Intensive Care

Emberger,

Boss,

Baumann

et al. 2023

2023 10th IEEE Swiss Conference on Data Science (SDS)

Self Cite

View full text Add to dashboard Cite

show abstract

“…The training pipeline of the dynamic model selected C-reactive protein, Creatine kinase, Creatinine, Interleukin-6, Leukocytes, Lymphocytes ratio, Middle corpuscular volume, Serum osmolality, and Neutrophils ratio. In the ROC analysis, the final voting model showed an AUC of 0.747 ± 0.094, while the dynamic and static models had an AUC of 0.729 ± 0.083 and 0.679 + 0.131, respectively 35 .…”

Section: Methodsmentioning

confidence: 96%

“…Survey: 95 clinicians from the N-ICU participated in the online survey at the start of the design process (the response rate was high at 88.79%). To set the stage, clinicians received a video-based online instruction explaining the general purpose of the DCIP and how the system will be integrated as a plug-and-play application into the already existing ICU cockpit dashboard 35 . Participants were then presented with a patient scenario depicting an aSAH patient with ambiguous symptoms of an upcoming DCI, followed by scenario-based questions about (Q1) which actions they would take to assess the risk of DCI in the described patient without the assistance of the DCIP; (Q2) actions they would take to assess the risk of a DCI in this patient with the assistance of the DCIP; and (Q3) which factors would help them establish trust in the DCIP.…”

Section: Methodsmentioning

confidence: 99%

Solving the explainable AI conundrum by bridging clinicians’ needs and developers’ goals

Bienefeld

Boss

Lüthy³

et al. 2023

npj Digit. Med.

Self Cite

View full text Add to dashboard Cite

Explainable artificial intelligence (XAI) has emerged as a promising solution for addressing the implementation challenges of AI/ML in healthcare. However, little is known about how developers and clinicians interpret XAI and what conflicting goals and requirements they may have. This paper presents the findings of a longitudinal multi-method study involving 112 developers and clinicians co-designing an XAI solution for a clinical decision support system. Our study identifies three key differences between developer and clinician mental models of XAI, including opposing goals (model interpretability vs. clinical plausibility), different sources of truth (data vs. patient), and the role of exploring new vs. exploiting old knowledge. Based on our findings, we propose design solutions that can help address the XAI conundrum in healthcare, including the use of causal inference models, personalized explanations, and ambidexterity between exploration and exploitation mindsets. Our study highlights the importance of considering the perspectives of both developers and clinicians in the design of XAI systems and provides practical recommendations for improving the effectiveness and usability of XAI in healthcare.

show abstract

“…Evaluation of different tree ensembles showed that Extremely Randomized Trees 26 performed best in the dataset. In the ROC analysis, the final voting model showed an AUC of 0.747 + 0.094, while the dynamic and static models had an AUC of 0.729 + 0.083 and 0.679 + 0.131, respectively 27 .…”

Section: System Descriptionmentioning

confidence: 97%

Solving the Explainable AI Conundrum: How to Bridge the Gap Between Clinicians Needs and Developers Goals

Bienefeld

Lüthy²,

Brodbeck

et al. 2022

Preprint

View full text Add to dashboard Cite

Explainable AI (XAI) is considered the number one solution for overcoming implementation hurdles of AI/ML in clinical practice. However, it is still unclear how clinicians and developers interpret XAI (differently) and whether building such systems is achievable or even desirable. This longitudinal multi-method study queries (n=112) clinicians and developers as they co-developed the DCIP – an ML-based prediction system for Delayed Cerebral Ischemia. The resulting framework reveals that ambidexterity between exploration and exploitation can help bridge opposing goals and requirements to improve the design and implementation of AI/ML in healthcare.

show abstract

ICU Cockpit: a platform for collecting multimodal waveform data, AI-based computational disease modeling and real-time decision support in the intensive care unit

Cited by 16 publications

References 25 publications

Video Object Detection for Privacy-Preserving Patient Monitoring in Intensive Care

Video Object Detection for Privacy-Preserving Patient Monitoring in Intensive Care

Solving the explainable AI conundrum by bridging clinicians’ needs and developers’ goals

Solving the Explainable AI Conundrum: How to Bridge the Gap Between Clinicians Needs and Developers Goals

Contact Info

Product

Resources

About