- Award ID(s):
- 1763509
- PAR ID:
- 10379021
- Date Published:
- Journal Name:
- CHI Conference on Human Factors in Computing Systems Extended Abstracts
- Page Range / eLocation ID:
- 1 to 7
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
Emergency medical services (EMS) teams are first responders providing urgent medical care to severely ill or injured patients in the field. Despite their criticality, EMS work is one of the very few medical domains with limited technical support. This paper describes a study conducted to examine technology opportunities for supporting EMS data work and decision-making. We transcribed and analyzed 25 simulation videos. Using the distributed cognition framework, we examined EMS teams' work practices that support information acquisition and sharing. Our results showed that EMS teams leveraged various mechanisms (e.g., verbal communication and external cognitive aids) to distribute cognitive labor in managing, collecting, and using patient data. However, we observed a set of prominent challenges in EMS data work, including lack of detailed documentation in real time, situation recall issues, situation awareness problems, and challenges in decision making and communication. Based on the results, we discuss implications for technology opportunities to support rapid information acquisition, integration, and sharing in time-critical, high-risk medical settings.more » « less
-
Abstract Background In prehospital emergency care, providers face significant challenges in making informed decisions due to factors such as limited cognitive support, high-stress environments, and lack of experience with certain patient conditions. Effective Clinical Decision Support Systems (CDSS) have great potential to alleviate these challenges. However, such systems have not yet been widely adopted in real-world practice and have been found to cause workflow disruptions and usability issues. Therefore, it is critical to investigate how to design CDSS that meet the needs of prehospital providers while accounting for the unique characteristics of prehospital workflows.
Methods We conducted semi-structured interviews with 20 prehospital providers recruited from four Emergency Medical Services (EMS) agencies in an urban area in the northeastern U.S. The interviews focused on the decision-making challenges faced by prehospital providers, their technological needs for decision support, and key considerations for the design and implementation of a CDSS that can seamlessly integrate into prehospital care workflows. The data were analyzed using content analysis to identify common themes.
Results Our qualitative study identified several challenges in prehospital decision-making, including limited access to diagnostic tools, insufficient experience with certain critical patient conditions, and a lack of cognitive support. Participants highlighted several desired features to make CDSS more effective in the dynamic, hands-busy, and cognitively demanding prehospital context, such as automatic prompts for possible patient conditions and treatment options, alerts for critical patient safety events, AI-powered medication identification, and easy retrieval of protocols using hands-free methods (e.g., voice commands). Key considerations for successful CDSS adoption included balancing the frequency and urgency of alerts to reduce alarm fatigue and workflow disruptions, facilitating real-time data collection and documentation to enable decision generation, and ensuring trust and accountability while preventing over-reliance when using CDSS.
Conclusion This study provides empirical insights into the challenges and user needs in prehospital decision-making and offers practical and system design implications for addressing these issues.
-
In this paper we present the promise of the Cognitive Work Analysis (CWA) methodology, particularly abstraction hierarchy modeling, in the foster care domain. There is increasing interest in applying machine learning decision aids to foster care decision making, but that interest is accompanied by concerns that those aids may perpetuate systemic bias or be largely context-blind. Modeling the work conducted at different levels of the domain offers unique insights into where bias may enter the system as well as possible design implications for these future decision aids. This project models two major areas of work in the domain, management of individual cases and management of overall programs offered. These work areas are then considered in the first 3 levels of the abstraction hierarchy to display the promise that this model can hold for the domain in future work, particularly when supported with more naturalistic studies.more » « less
-
On the Variety and Veracity of Cyber Intrusion Alerts Synthesized by Generative Adversarial NetworksMany cyber attack actions can be observed but the observables often exhibit intricate feature dependencies, non-homogeneity, and potential for rare yet critical samples. This work tests the ability to model and synthesize cyber intrusion alerts through Generative Adversarial Networks (GANs), which explore the feature space through reconciling between randomly generated samples and the given data that reflects a mixture of diverse attack behaviors. Through a comprehensive analysis using Jensen-Shannon Divergence (JSD), conditional and joint entropy, and mode drops and additions, we show that the Wasserstein-GAN with Gradient Penalty and Mutual Information (WGAN-GPMI) is more effective in learning to generate realistic alerts than models without Mutual Information constraints. The added Mutual Information constraint pushes the model to explore the feature space more thoroughly and increases the generation of low probability yet critical alert features. By mapping alerts to a set of attack stages it is shown that the output of these low probability alerts has a direct contextual meaning for cyber security analysts. Overall, our results show the promising novel use of GANs to learn from limited yet diverse intrusion alerts to generate synthetic ones that emulate critical dependencies, opening the door to data driven network threat models.more » « less
-
Bluetooth-based item trackers have sparked apprehension over their potential misuse in harmful stalking and privacy violations. In response, manufacturers have implemented safety alerts to notify victims of extended tracking by unknown item trackers. In this study, we specifically investigate the anti-stalking mechanism of Apple's AirTag. We identify and analyze potential triggers of safety alerts that have not been examined in previous research, such as the local time, the victim's device model, AirTag's battery life, and the distance between the AirTag and the victim's device. Furthermore, we demonstrate a novel possibility of developing a stealthy cloned AirTag capable of tracking victims directly on the Find My app while circumventing safety alerts on the victim’s device. Our experiments demonstrate that, despite regular updates to the public key and MAC address, our cloned AirTag can provide real-time location updates even with a four months old key, thereby highlighting the challenges in designing a robust anti-stalking framework. Furthermore, we propose practical solutions to mitigate stalking risks from cloned AirTags and enhance the existing anti-stalking safeguards for AirTags. These suggestions seek to provide a foundation for similar Bluetooth-based item trackers to improve their anti-stalking protections while ensuring optimal tracking efficiency. We conducted rigorous experiments to validate our findings, ensuring their accuracy and reliability. Our evaluation highlights that safety alerts take over 8 hours to appear during the day and are more prompt during the night, particularly after 11 pm.more » « less