Computational competitions are the standard for benchmarking medical image analysis algorithms, but they typically use small curated test datasets acquired at a few centers, leaving a gap to the reality of diverse multicentric patient data. To this end, the Federated Tumor Segmentation (FeTS) Challenge represents the paradigm for real-world algorithmic performance evaluation. The FeTS challenge is a competition to benchmark (i) federated learning aggregation algorithms and (ii) state-of-the-art segmentation algorithms, across multiple international sites. Weight aggregation and client selection techniques were compared using a multicentric brain tumor dataset in realistic federated learning simulations, yielding benefits for adaptive weight aggregation, and efficiency gains through client sampling. Quantitative performance evaluation of state-of-the-art segmentation algorithms on data distributed internationally across 32 institutions yielded good generalization on average, albeit the worst-case performance revealed data-specific modes of failure. Similar multi-site setups can help validate the real-world utility of healthcare AI algorithms in the future.

Introduction: Management of daily activities in ICUs is challenging. ICU shift leaders, charge nurses and intensivists have to make several immediate ad hoc decisions to enable the fluent flow of ICU activities. Even though the management of ICU activities is quite well delineated by international consensus guidelines, we know only a little about the content of the real clinical decision making of ICU shift leaders. Methods: We conducted an observational study with the think-aloud technique to describe the ad hoc decision making of ICU shift leaders. The study was performed in two university-affiliated hospital ICUs. Twelve charge nurses and eight intensivists were recruited. Observations were recorded and transcribed for qualitative content analysis using the protocol analysis method. The software program NVivo 7 was used to manage the data. The interrater agreement was assessed with percentages and by Cohen's κ. Results: We identified 463 ad hoc decisions made by the charge nurses and 444 made by the intensivists. During our data collection time, this breaks down to over 230 immediately made decisions per day (24 hours). We divided the ad hoc decision making of ICU shift leaders into two types: process-focused and situation-focused. Process-focused decision making included more permanent information, such as human resources, know-how and material resources, whereas situation-focused decision making included decisions about single events, such as patient admission. We named eight different categories for ICU ad hoc decision making: (1) adverse events, (2) diagnostics, (3) human resources and know-how, (4) material resources, (5) patient admission, (6) patient discharge, (7) patient information and vital signs and (8) special treatments. Conclusions: ICU shift leaders make a great number of complex ad hoc decisions throughout the day. Often this decision making involves both intensivists and charge nurses. It forms a bundle that requires versatile, immediate information for a successful outcome. In the future, we need to investigate which information is crucial for ad hoc decision making. These challenges should also be emphasised when information technology programs for ICU care management are developed.

Healthcare is an information-intensive field, as information is needed to make strategic, tactical, and operational decisions. The purpose of this study was to identify the tactical decisions that middle management healthcare managers make, the information that is available, and the necessary information that is missing using the cardiac care process as an example. Data were collected through focused interviews of nurses and physicians who work in middle management in a secondary healthcare field. The interviews were coded and analyzed using the thematic content analysis method. We identified two main categories of tactical decisions: those concerning the process of care and those concerning the resources for the care. We termed the categories "process decisions" and "resource decisions." The availability of information varied. Much of the necessary information was created and processed manually. Our results show that the collection, mining, and systematic use of information are difficult because of the existence of many types of information systems and their varying abilities to produce and report information. Finally, much of the important information is missing. In conclusion, the information management process in healthcare settings needs to be improved, and a new generation of information system is needed to support tactical decision making in middle management.

Aging is one of the greatest challenges in modern society. The development of wearable solutions for telemonitoring biological signals has been viewed as a strategy to enhance older adults’ healthcare sustainability. This study aims to review the biological signals remotely monitored by technologies in older adults. PubMed, the Cochrane Database of Systematic Reviews, the Web of Science, and the Joanna Briggs Institute Database of Systematic Reviews and Implementation Reports were systematically searched in December 2021. Only systematic reviews and meta-analyses of remote health-related biological and environmental monitoring signals in older adults were considered, with publication dates between 2016 and 2022, written in English, Portuguese, or Spanish. Studies referring to conference proceedings or articles with abstract access only were excluded. The data were extracted independently by two reviewers, using a predefined table form, consulting a third reviewer in case of doubts or concerns. Eighteen studies were included, fourteen systematic reviews and four meta-analyses. Nine of the reviews included older adults from the community, whereas the others also included institutionalized participants. Heart and respiratory rate, physical activity, electrocardiography, body temperature, blood pressure, glucose, and heart rate were the most frequently measured biological variables, with physical activity and heart rate foremost. These were obtained through wearables, with the waist, wrist, and ankle being the most mentioned body regions for the device’s placement. Six of the reviews presented the psychometric properties of the systems, most of which were valid and accurate. In relation to environmental signals, only two articles presented data on this topic. Luminosity, temperature, and movement were the most mentioned variables. The need for large-scale long-term health-related telemonitoring implementation of studies with larger sample sizes was pointed out by several reviews in order to define the feasibility levels of wearable devices.