CAVA
Why is this project worthy of an award?
because
What else would you like to share about your design? Why is it unique and innovative?
Obama
Who worked on the project?
Bill Young
View the project video:
Clinical Decision Support for Artificial Heart Implant
Company Carnegie Mellon University
Introduction Date September 25, 2018
Project Website
Why is this project worthy of an award?
The artificial heart (ventricular assist device, VAD) is an implantable electromechanical device used to partially replace the function of a heart. It is the last therapeutic treatment for end-stage heart failure patients. VADs were expected to extend patients lives for several years. However, many who received VADs died shortly after the implant. Machine Learning (ML) can offer a powerful way to foresee implant outcomes and help doctors better identify patients who are more likely to benefit from the therapy. Interestingly, in the past 30 years, almost all such tools failed when migrating from labs to clinical practice. ML remains the least used feature in the Electronic Medical Records (EMR) systems. While ML systems often assume doctors would recognize when they need help, walk up to computers and seek advice, research showed that clinicians often perceive no need for data support. When facing difficult cases, clinicians rely on teamwork. Team discussions take place in meetings, through phone calls as well as during impromptu hallway chats. Clinical decisions rarely happen in front of a computer, nor at a single time point or by a single user. In this light, we designed a radically different decision support system that takes the form of a presentation-slide generator. It extracts patient data from EHR, automatically populates presentation slides, meeting printouts and structured emails for clinician team’s decisions. The system embeds ML predictions into these day-to-day documents and offers links to further information on key factors influencing the prediction and most likely causes of patient demise. As FastCompany puts it, the artifact of this design is ingenious. [1] Yet this simple design effectively situates machine intelligence in the chaotic, human-centered clinical practice. We believe three key ideas are essential to the success. One, the system motivates adoption by automating many manual, day-to-day paperwork for the busy clinicians. Second, it embeds ML predictions seamlessly and unobtrusively into clinician workflow -- meetings, printouts, emails -- whenever and wherever decisions get made. In doing so, the system does not slow the decision process down but allows clinicians to learn to trust a system that mostly confirms what they already think. Lastly, when the ML predictions contradict attending physicians’ opinion, the tool provides a new mechanism for lower status clinicians (nurse practitioners, fellows, and medical students) to question a decision as they are not expressing their own disagreements, but are instead noting that the algorithm is offering different advice. This system thus can inform clinicians in a way that makes them feel better at doing their jobs, instead of being substituted by inhuman technology. There have been incredible technical advances in AI in recent years, yet too often it seems that AI’s success in the real world depends solely on algorithm performance. This project sheds a different light on this issue. It demonstrates a provocative yet pragmatic solution to the challenge of human-AI collaboration in the context of high-risk health care. [1] https://www.fastcodesign.com/90157144/an-ingenious-approach-to-designing-ai-that-doctors-trust
What else would you like to share about your design? Why is it unique and innovative?
This design is grounded in rare observations of four artificial heart implant centers and physicians from other disciplines. Today, most intelligent systems for clinicians are created in labs with a focus on algorithm accuracy, not on the felt experience of clinician users. Hospitals are hostile environments to user-centered design methods and practices. Strict rules regarding health data security (as enforced by HIPAA and IRB) and legal concerns about algorithms biasing clinician judgments make it extremely difficult for designers to gain access to clinicians, patients or the real clinical work environment. In this project, we hard-earned access to four hospitals through our collaboration with biomedical engineers and multiple research institutions. We were able to closely observe the ways clinicians make implant decisions at different touchpoints (hospital ward rounds, clinic visits, hallway conversations, decision meetings, etc) for weeks. The insights gained through these observations and a co-design process with physicians led to our unique yet grounded design solution. 2) This design is generalizable for many other high-risk clinical interventions. Designing technology systems that can work for many hospitals is extremely difficult. This is because clinical decisions and clinician work environments differ significantly across hospital sites and healthcare disciplines. Therefore, new technology adoption often happens slowly in the medical world. Our collaboration with four different hospitals sensitized us to this challenge of design generalizability. Our design solution -- a decision support tool that embeds ML predictions in emails, meeting presentations and printouts -- can be easily modified for facilitating many other high-consequence, interventional healthcare decisions. This is because the clinical decisions on aggressive, end-of-life interventions almost always involve evaluations of multiple organs as well as patient’s social conditions, therefore require clinicians of multiple specialities to take part. Embedding AI in these multidisciplinary meetings, -- formal or informal, technology-mediated or paper-based -- is an effective way to overcome clinician acceptance and adoption barriers for clinical decision support tools of many kinds. Our design, though originated from Artificial Heart Implant Centers, has attracts interests of physicians from a range of disciplines, from pediatrics, intensive care, emergency department, kidney transplant, esophageal surgery, gastroenterology and hepatology, and more.
Who worked on the project?
Qian Yang (Research Fellow in Digital Health, Ph.D. student), John Zimmerman (Researcher, Professor), both at Carnegie Mellon University Human-Computer Interaction Institute.
View the project video:
CollectiveVR
Why is this project worthy of an award?
It is no small thing to leave a mark on our fast-changing, complex world. As architects, we design for people in one of the most permanent design fields. This lasting impact makes inclusive design our foremost value and responsibility. To do this, we need to develop a deep awareness of the social, cultural and environmental needs of the people we serve. Our goal is to transform the design process through empathy, by pioneering new methods of understanding. The proliferation of the digital tool set has given us the ability to transcend traditional architectural practice for the first time in generations. Leveraging our multidisciplinary studio, we are advancing virtual reality to better understand the people who inhabit our buildings by simulating experiences related to vision, mobility and communication. We can’t expect to truly comprehend an individual’s contextual and cumulative experiences, but if we are able to tap into others’ abilities and sensorial comprehension, our designs can be more sensitive to the broader community and cause fewer impediments in the built environment. CollectiveVR is our approach to transform the architectural creative process through diverse understandings. By merging the immersive nature of virtual reality with the ability to modify our experiential abilities, we hope to embrace the extraordinary collective of our unique human perspectives. Through the creation and application of experience modifiers hosted in a virtual reality toolset, we gain insights and means to communicate in a more profound manner. CollectiveVR encourages us to explore the many ways in which people might experience the spaces we design, creating a new framework for our design processes. Our responsibility as architects is both inspiring and humbling: to create a positive, lasting impact through design. CollectiveVR elevates our ability to iterate, understand and communicate ideas through a widened frame of reference. By enabling architects to understand through experience, it creates a common ground for us to learn from one another. CollectiveVR is not a shortcut to empathy, but by tapping into different perspectives, we broaden our understanding of others.
What else would you like to share about your design? Why is it unique and innovative?
Research of the populations that occupy space is a cornerstone of CollectiveVR and has led to profound personal experiences and insightful collective studies. The first perspectives of the CollectiveVR platform were initiated by a study of the aging population. In the next decade, one quarter of the United States population will be over 65 years of age; a demographic that is often afflicted with a variety of acute and chronic conditions. Cataracts, Glaucoma and Age Related Macular Degeneration are the most common vision impairments experienced by this demographic. Using a virtual reality headset, we developed proprietary software to create experience modifiers around these three eye conditions. We then paired these experience modifiers with a simulation body suit to explore the unbuilt. Today the platform is occupied by an increasing collection of perspectives related to movement, communication and vision. As we continue to research and gain insight into diverse ways that our designs are experienced, we can add, change and modify the modules that occupy the CollectiveVR platform. It is easily customized to specific projects, designers and creative processes, allowing for diverse possibilities in its use and integration. The design of CollectiveVR is centered on a simple user interface, hosted within a virtual wrist watch that feels both familiar and accessible. Any experience modifier or tool is a maximum of three clicks, gazes or voice cues away from the home interface, creating a clear user experience that does not impede on the sense of immersion. Through the CollectiveVR operating system, we can document and analyze our interactions – both with each other and with the proposed architecture. On this foundation of fresh understanding, with quantitative and qualitative insights, we can collaborate, design and innovate for enduring impact in a way never before possible.
Who worked on the project?
Jonathan Bartling (Principal), Adam Barnstorff, Anupam Das, Rich Firkins, Nicolas Ramirez, Lene-Mari Sollie, Ryan Spiering, Jared Widner
View the project video: https://www.youtube.com/watch?v=9tmmnHk2sfM
ColorFuse: An integrated, automated manufacturing process
Why is this project worthy of an award?
ColorFuse is a direct surface coating process for car exteriors that combines plastic injection molding and painting in a single step. Exterior sections made of plastic, such as door trim, sunroofs, and lighting fixtures, are commonly injection-molded and then need to be prepped before being painted. With ColorFuse, plastic parts have the color already infused within them, which allows designers a new freedom to create elements with elegant surface aesthetics. Parts can be made with sharp edges, partly grained surfaces with precise accents and styling lines that are high gloss and scratch resistant. From an economic standpoint, ColorFuse can be integrated into existing manufacturing processes to save car manufactures 35% compared to the cost of wet painting processes.
What else would you like to share about your design? Why is it unique and innovative?
ColorFuse gives car designers new freedom to create shapes and textures for car exteriors that would not be otherwise possible. It unleashes car designers’ creativity to make eye-catching designs without the restraints imposed by painting parts in a separate step. Plastic elements are more seamlessly integrated with the rest of the car to give it a more eye-catching look.
Who worked on the project?
R&D team in Ruti Switzerland
View the project video:
Co-Minkan
Company The Committee of Co-Minkan
Introduction Date August 1, 2017
Project Website https://www.co-minkan.org/
Why is this project worthy of an award?
Co-Minkan is a new form of social education for everyone. Japan needs to tackle a variety of social issues caused by aging population, and loneliness. Though the civic participation is necessary, most of the people aren’t interested. Japan has its own social education system but it is not catered for current needs and social issues. Co-Minkan is redesigned to suit to the people in this era and lower the limits to welcome anyone who would like to learn about their area, community, people and their own lives through dialogues with other people or experience the activities.
What else would you like to share about your design? Why is it unique and innovative?
Co-Minkan is not just to learn but also to gather and connect to each other to make bonds amongst the people who live near-by so that it creates the sense of belongings in a community and avoid social isolation or loneliness. Co-Minkan is a mobile package and it contains a kit such as tea set and stationery to start a mini gathering. It can be opened to start at anywhere by anyone with their own resources. If you have an empty houses or disused spaces you can open up this package to start your own gathering.
Who worked on the project?
Ms Noriko Deno, Co-Founder, Ms Arisa Nishigami, Co-Founder, Mr Taro Yokoyama, Co-Founder and Mr Ryo Yamazaki, Advisor.
View the project video:
