Embracing the Digital Future of Healthcare - Symposium

Where: The MAGIC Center on the campus of RIT

When: Tuesday, May 16, 2023

Embracing the Digital Future of Healthcare is a one day symposium dedicated to the digital future of healthcare. The symposium will include keynote addresses, expert panels, oral and poster presentations, along with a student idea contest. Topics will include immersive technology such as augmented and virtual reality, the use of artificial intelligence, machine learning, human-computer interactions, computer modeling, robotics, gaming, etc.

Attendees will learn about how these technologies are being deployed for disease prediction and prevention, treatment and rehabilitation, medical imaging, pain management, healthcare education and access, telehealth, wearable, wellness, and many other areas across the healthcare continuum.

The symposium will also include the Digital Future of Healthcare Idea Contest. This contest is specifically designed to promote and advance early-stage concepts that employ technology (especially digital) to improve any dimension of healthcare. All students are encouraged to apply. Deadline to apply is May 10, 2023.

Deadline to submit an 
abstract is May 10, 2023

Deadline to register
is May 8, 2023

Program

8-8:30 a.m.

Registration and Continental Breakfast

8:30-9 a.m.

Welcome
Dr. David C. Munson Jr., President of RIT
Dr. Sarah Mangelsdorf, President of University of Rochester

9-9:30 a.m.

Keynote
Dr. Richard “Chip” Davis, President of Rochester Regional Health
The Future of Healthcare

9:30-10 a.m.

Keynote
Dr. Michael F. Rotondo, CEO University of Rochester Medical Faculty Group
Importance of Digital Healthcare

10-10:15 a.m.

Break

10:15-10:35 a.m.

Invited Talks I
Artificial Intelligence for Healthcare


Gregg T. Nicandri, MD, FAAOS, FAOA
Chief Medical Information Officer
Professor of Orthopaedic Surgery
University of Rochester Medical Center

Linwei Wang, Ph.D.
Professor of Computing and Information Sciences, Golisano College of Computer and Information Sciences
Director of Personalized Healthcare Technology (PHT180)

10:35-10:55 a.m.

Invited Talks II
AR/VR/XR/3D/Gaming for Healthcare


David J. Mitten, M.D.
Director, UR Health Lab
Professor of Orthopedics and Biomedical Engineering
University of Rochester

Robert H. Rice, Jr., Ph.D., LMHC
Associate Professor and Chair of Mental Health Counseling
Core CACREP Faculty
HRSA BHWET Integrated Care Project Director
St. John Fisher University

10:55-11:10 a.m.

Break

11:10-11:30 a.m.

Invited Talks III
Behavioral Health


Caroline Easton, Ph.D.
Professor of Behavioral Health
College of Health Sciences and Technology
Rochester Institute of Technology

Kayla Hunt, Psy.D.
Associate Director, Pediatric Behavioral Health & Wellness Outpatient Services
Associate Director of Informatics in Psychiatry
Assistant Professor of Psychiatry and Neurology
UR Medicine: Pediatric Behavioral Health & Wellness
University of Rochester Medical Center: Department of Psychiatry

11:30 a.m.-11:50 p.m.

Invited Talks IV
Healthcare Access


Michael Hasselberg, PhD, RN, PMHNP-BC
Co-Director, UR Health Lab
Chief Digital Health Officer, University of Rochester Medicine
Associate Professor of Psychiatry, Clinical Nursing, and Data Science
University of Rochester

Matt Huenerfauth, Ph.D.
Professor and Dean of the Golisano College of Computer and Information Sciences
Director of the Linguistic and Assistive Technology Lab
Co-Director of the Center for Accessibility and Inclusion Research
Rochester Institute of Technology

12-1 p.m.

Lunch

1-3 p.m.

Poster Session

3-4 p.m.

Satellite Meetings

Abstracts

Abstracts are a poster presentation of one’s research. Abstracts will be located on the Sound Stage at MAGIC Spell Studios.

 

Presenter: Matthew Altobelli

Cognitive processing styles are the lens through which we view the world. There are three main processing styles: accommodation, overaccommodation, and assimilation. These styles influence how we process information from new events by comparing them to schemas about situations or objects. We are interested in how these processing styles influence the mental well-being of those who experience an adverse event. Language influences thought, and thought influences behavior, so how a person talks or writes about an experience tells how they process information. We analyzed the written narratives of 106 undergraduate students who either witnessed or were a part of an adverse event. We annotated processing style labels at the statement level and ran a statistical analysis to see the link between those styles and mental well-being on measures such as the state-trait anxiety inventory, Beck's depression inventory, positive and negative affect schedule, and brief post-traumatic cognitions inventory. The extracted labels were then used to train and test a machine-learning algorithm that automatically provides these labels to novel statements. The impact of this research adds strength to the current literature on cognitive processing styles and mental well-being. It also provides a new way to analyze written statements for clinicians and researchers.

Presenter: Adam Baker

BZDesign is a design and innovation consultancy in Rochester, NY that has been at the forefront of creating accessible healthcare products and technologies in conjunction with RIT and the LiveAbility Lab. Their collaborative efforts in developing innovative accessibility products for production have resulted in solutions that meet the needs of diverse populations. By leveraging the power of collaboration and design thinking, BZDesign has helped refine products that are inclusive, user-friendly, and effective.

Their approach involves working closely with stakeholders, including patients, healthcare providers, and industry experts, to identify and understand the needs and challenges of different populations. This collaborative approach has enabled BZDesign to design solutions that are both functional and aesthetically appealing, while also meeting regulatory and safety standards, while focusing on improving cost and ease of production. BZDesign's efforts are a testament to the power of collaboration and design thinking in creating healthcare products that are accessible to all.

Some examples of these projects are the Anchor Shaker, ROAM, and Tappo devices:

  • ROAM is a product developed in collaboration with ABVI, D3 Engineering, and Goodwill Finger Lakes that is focused on users with visual impairment. This device helps users identify obstructions between waist and head height that could cause harm or inconvenience.
  • Tappo is a accessible play device meant to analyze, monitor, and improve a variety of abilities ranging from cognitive function, motor skills, and auditory and visual processing.
  • Anchor Shaker is a collaborative musical play product that encourages users to work together, as well as improve cognitive and motor function in a simple analog shaker device.

Presenter: Lizhou Cao

This work presents the design and evaluation of three mini exercise games (exergames) that were developed for virtual reality (VR) and can be played using a head-mounted display (HMD). The aim was to investigate whether immersive display devices like the HMD could enhance exergaming and promote physical activity in comparison to widely-used flat-screen display devices. As VR technologies are increasingly integrated into exergames for fitness, training, rehabilitation, and education, it is important to examine the potential benefits and limitations of such devices.

The exergames developed in this work included core scenarios of aerobic exercises, such as reaction drills, movement control drills, and stretching drills. We analyzed the heart rate data and gathered feedback from participants to evaluate the exercise intensity of the games. The difficulty level of gameplay was adjusted by control parameters, allowing for the exercise intensity level to be modified from moderate to light, or to vigorous. We ensured that the difficulty level of gameplay was identical for both HMD and flat-screen versions of the games to ensure that any differences in engagement and exertion in our experiment were not influenced by variation in gameplay.

Our experimental results showed that the immersive environment provided by the HMD resulted in a more engaging experience than the standard flat screen. However, we found no evidence to suggest that wearing the HMD would be more effective than using the flat screen in increasing the exercise intensity level. Our results also suggested that the VR version of an exergame can better motivate people who do not usually engage in exercises than non-VR exergames.

Presenter: Sergei Chuprov

Data Quality (DQ) and security are crucial in medical applications as they play vital roles in the accurate diagnosis and treatment of patients (e.g., MRI scans) as well as their privacy protection. However, the DQ can be affected by various factors, such as the scanning equipment used, the imaging technical parameters, and the data communication factors. This poses a significant challenge in fusing data from diverse DQ sources in ML training, as various MRI machines might produce scans of varying quality. We propose Federated Learning (FL) as the promising technology to design ML based medical imaging applications, which are both robust to possible DQ diversity and degradation, and allows enhancing users' privacy and security. We report the results of two studies. In the first, we investigate how DQ impacts the performance of ML medical image classifiers. We employ real-world X-ray images of lung diseases and industry-standard ML models such as VGG16, ResNet50, and InceptionV3 to compare their performance under different conditions. According to our findings, even small DQ degradation with packet losses of 2-5% can significantly reduce the accuracy of ML classifiers from 90+% to about 70-80%, making those classifiers unsuitable for medical applications.

In the second, we investigate how the FL employment might improve the robustness of ML models trained on data of diverse DQ. We evaluate ML image classifiers trained on data cohorts of various quality and demonstrate how their performance depends on the input DQ. We then provide our recommendations on how to fuse diverse data to produce more robust ML models trained in a FL manner.

Presenter: Teresa Gibson

Objective: This study aims to describe engagement, utilization and associated employee characteristics with a digital tool for SARS-CoV-2 exposure and infection implemented in a workplace setting.

Methods: Employees of a multi-site utility company were offered return-to-work digital tools including a 7-item daily symptom survey. Employee engagement with the digital tool was defined via measures of frequency, intensity, longevity, and recency. Using multivariable linear regression, we estimated the relationship between engagement and individual employee-level characteristics (age group, sex, race, ethnicity, Body Mass Index, education, chronic conditions, household size, smoker, work location). An engagement index was developed using two approaches: standardized summed scores and confirmatory factor analysis.

Results: In 2021, 349 (20.7%) employees consented to participate in the study. Their average age was 49.7 years (SD 11.1), 44.8% were female, 50.7% were White, 32.9% Black, 16.3% other. Employees filled out an average of 3.6 (SD 1.12) of 7 surveys per week and 93.5% filled out surveys throughout the entire time offered (longevity). 33.5% were routed to further screening at least once due to exposure to or symptoms of COVID-19. After adjusting for covariates, age, lower values of BMI, and employee work location at headquarters were associated with higher levels of engagement measures, and the composite engagement index (p<0.05). Regressing the standardized engagement index on employee characteristics, engagement decreased with age (p<0.05).

Results: We found considerable variation in engagement with the digital tool. This study expands knowledge of real-world application of digital solutions using data directly collected from users in a workplace setting. Reference: Gibson, Schubel, Assibey-Mensah et al. (2022) Value in Health 25(7 Supplement): S459-460.

Presenter: Arianna Giguere

Approximately ¼ - ½ million new U.S. stroke patients per year are left with cortical blindness (CB), a type of peripheral vision loss that affects a quarter to a half of their visual field. In many states, CB patients are still legally allowed to drive, and they do. One critical subtask of driving is the task of steering to stay within one's lane. The visual guidance of steering is thought to rely significantly on optic flow (OF), and it is perhaps surprising that CB patients demonstrate greater biases in lane keeping than visually-intact controls (Bowers et al. 2010) given that heading judgments can be accurate using only a portion of the flow field (Warren and Kurtz 1992). One possible explanation for these lane biases is that the damage to the primary visual cortex that causes cortical blindness also acts as a source of noise that biases perception of heading during steering, specifically by corrupting visual input that falls upon the blind field. In this preliminary study, we asked to what extent OF signal strength impacts steering accuracy for participants with and without CB, and how this is impacted by gaze behavior. Fifteen visually-intact subjects and eight CB patients were immersed in a simulated environment seen through an HTC Vive Pro, and they were tasked with staying within a procedurally-generated roadway while traveling at 26.6 m/s (59.5 mph). Between turns, we varied turn direction (right/left), turn radius (35, 55, or 75 m), and the density of texture elements that provide OF information (low, medium, high). Results from our fifteen visually-intact subjects revealed that average signed divergence from road center increased with OF density (low OF: 0.16 meters ±0.36 (SD), medium: 0.29±0.34m, high: 0.34±0.35m), where positive values indicate bias towards the road's inner edge (i.e. corner-cutting). Our findings suggest that lane biases and corner-cutting behavior are strongly impacted by texture density (and thus OF) in the virtual environment, in a manner similar to the effect of increasing global flow rate found by Kountouriotis et al. (2016). Preliminary investigation reveals that CB patients were less susceptible to manipulations of optic flow. Moreover, when turning in the direction of their blind field, CB patients demonstrated different corner-cutting biases and adopted a more exaggerated gaze angle, with the effect of causing more of the future road to fall on the intact portion of their visual field. Future work will be extended to investigate why OF affects CB steering and gaze patterns in this way and how this information can be leveraged in vision rehabilitation work.

Presenter: Paige Hepple

Objective: To characterize the evolution of motor control impairments in wrist flexor and extensor muscle groups over the first 3 months post-stroke.

Background: Upper extremity (UE) disability is common after stroke. Prior studies in individuals with hemiparesis due to stroke have identified four distinct impairments of motor control: (1) decreased maximal muscle activation, (2) delayed muscle activation, (3) motor fatiguability, and (4) abnormal co-activation of antagonistic muscle groups. However, the factors predicting emergence of specific motor control abnormalities are unclear.

Design/Methods: We have developed an electromyographic (EMG) computer interface to collect EMG signals from subjects performing repetitive muscle activations. Our EMG computer interface uses surface EMG signals from the wrist flexor and extensor muscle groups to control a computer game. Successful gameplay requires multiple isometric muscle contractions at precise time points. EMG data are analyzed to identify maximum EMG values, activation delay, and co-activation of antagonist muscle groups during gameplay. We are conducting a longitudinal study over the first 3 months post-stroke to study evolution of each impairment and identify factors that may predict specific motor control abnormalities in the chronic phase after stroke.

Results: Preliminary results suggests that our system is well-tolerated and that subjects early (<10 days) after stroke exhibit decreased EMG amplitude and increased variability in activation onset of wrist flexor and extensors in their affected UE as compared to their unaffected arm. Some participants also show increased co-activation and increased motor fatigue in their paretic arm.

Conclusions: Our EMG-controlled computer interface can be used to collect EMG data across multiple time points. With our current longitudinal study, we will assess the prevalence and individual trajectory of each motor impairment over the first 3 months post-stroke. Ultimately, we hope to use the information gained in this longitudinal study to develop personalized, impairment-specific interventions for motor recovery.

Presenter: Michael Huddleson

Introduction: Alcohol use disorder (AUD) is among the most prevalent mental health diagnoses among homeless populations in the United States . Increasing evidence demonstrates that the Wish, Outcome, Obstacle, Plan (WOOP) intervention can enhance top-down cognitive control to reduce the risk of relapse among individuals with SUD through mental contrasting with implementation intentions (MCII). The current study sought to assess feasibility and efficacy of implementing the WOOP intervention combined with digital technology in a homeless population with AUD. Method: Participants will include residents (N = 12) at a local homeless shelter who will be participating in AUD treatment through an onsite clinic. Clients will provide information about substance use and precursors to use, including craving and self-efficacy, at intake and throughout treatment for up to 10 sessions. It is anticipated that the WOOP intervention will be associated with reductions in AUD indicators over time. Data will contribute to program evaluation and improvement of services at the clinical site.

Presenter: Stephen Jacobs

Federal Agencies, Private Funders and Professional Societies are requiring earlier, broader, and wider Open Dissemination of funded work. This has significant secondary effects on all funded research. The move to a more Open academia will significantly impact policies and practices including tenure and promotion, tech transfer, faculty recruitment, and more.

Presenter: Sheetal Kashid

Background: It has been posited that the Black vs. White disparity in stroke risk may be partly attributable to the relative underdetection of AF in Black compared with White patients. We propose a novel monitoring solution that could minimize the impact of the socio-economic disparities in access to monitoring solutions. We developed a video- based monitoring technology to detect AF without the need for a patient to adopt a dedicated wearable device. It is a low-cost software-only solution running on smartphones. It minimizes requirements for patient compliance with recording procedures since it passively monitors patients while they use their smartphone for other purposes.

Objective: Similar to photoplethysmography, the technology captures a pulsatile signal from a patient's face. Remote video-based cardiac monitoring measures the modulation of ambient light reflected from one's face due to changes in blood volume (hemoglobin) in the upper layers of the skin. In addition to hemoglobin, melanin is also a chromophore of the human skin. Hence, we aim to assess the performance of the technology to measure heart rate (HR) and detect the presence of AF across the whole spectrum of human complexions.

Methods: We enrolled paroxysmal and persistent AF patients. Based on the Fitzpatrick score, we used 3 complexion groups: white (A), tanned and olive-color (B), and brown and black (C). Measurements were conducted under 4 indoor illumination levels (50, 100, 200, and 500 lux). For reference, we used a single-lead EKG. We used 25% of the cohort to train. HR accuracy is assessed using Bland-Altman analyses with paired replicate measures based on one-way repeated ANOVA. The detection performance was measured as an average sensitivity (SENS) and specificity (SPEC) adjusted for repeated measures.

Results: We collected 7,042 recordings from 60 AF patients (age: 67±10 years, 44 men). We enrolled 23, 26, and 11 patients in the complexion groups A, B, and C, respectively. In the validation, 87% of the recordings were conclusive. Overall SENS and SPEC were 96% and 92%. For an illumination >50 lux, the SENS and SPEC for the complexion groups were: [A: 99%,94%], [B: 91%,93%], [C: 96%,90%]. The mean HR measurement errors in reference to EKG were: [A: -2.0 bpm], [b: -0.2 bpm], and [C: 0.3 bpm] with limits of agreements <5 bpm.

Conclusion This monitoring technology enables an accurate measure of HR and detection of AF under indoor illumination and across all human skin complexions.

Presenter: Michael Kuhl

Efficient and effective healthcare delivery is key to realizing the potential of advances in medicine. This research focuses on the operational aspects of healthcare delivery that impact patient outcomes. In particular, we investigate Digital Twins to address issues related to facility design; patient flow; work flow; facility, equipment, and personnel scheduling; etc. A digital twin utilizes a dynamic, simulation modeling and optimization approach to understand the healthcare system behavior, identify opportunities for improvement, and propose and compare alternative solutions in a virtual environment before implementing them in the healthcare system.

Presenter: Diane Lee

We have been producing custom animated shorts using a cast of cartoon characters to deliver moral lessons for elementary school demographics in story-driven, entertaining ways. Each character in the cast embodies a "character trait," and advocates for their assigned trait in various fictional scenarios. Similarly directed animations may be produced under the guidance of pediatric health experts to address various behavioral and mental health themes.

Presenter: Ellen Masters

Group-based behavioral-oriented parent training programs have been shown to have positive effects on child behavior, parenting skills, and parental stress, across both community samples and clinical samples (Barkley 1997; Furlong et al., 2012; Jobe-Shields et al., 2015; Kaminski et al. 2008; Webster-Stratton 1998). Such programs have also been shown to have positive long-term effects on parenting strategies (Heinrichs et al. 2014; Jobe-Shields et al., 2015). However, limited parent engagement has been shown to have negative effects on program evaluation efforts and can ultimately reduce the potential benefits and positive outcomes associated with parent training programs (Dumas et al., 2007). The goal of the current study therefore is to assess possible predictors of parent attendance and engagement in a brief parent training program within a community sample, in order to better understand and promote parental engagement in future parent trainings programs.

Presenter: Dennis McCorry

Jam for Hope was a Game and Animation Jam hosted at MAGIC Spell Studios in partnership with the "Dear Jack Foundation" in order to fulfill a wish for Zach Skeikh. Zach is a fourth-year game design and development student who is battling cancer for the third time. Zach wanted to create uplifting animations and games to support others dealing with their own medical and emotional challenges.

Presenter: Jade Myers

Current Compression/Release Stabilized (CRS) prosthetic socket designs may result in soft tissue damage between areas of compression and release. Pressure and shear can contribute to tissue damage even outside of areas where a prosthetic socket contacts the body1 While CRS sockets offer promise for improved user control, increased range of motion (ROM), and more even distribution of pressures across the bone,2.3 the potential for tissue damage due to the abrupt change from areas of compression to open areas of release remains. A cushioned socket containing flexible density-graded lattice structures may allow for a more gradual and controllable transition in firmness between areas of compression and release.

Two initial unit cell types were used to create 3D printed lattice samples. Each cell type was subjected to 8 design alterations affecting density. Samples were compression tested to determine their capacity to meet predefined goals for serving as compression areas and release areas within a socket. The offset diamond unit cell performed best when paired with blend radius density alterations, but did not meet goal thresholds set for release areas. Based on these preliminary results, 36 additional unit cell types were investigated for their potential to serve as compression areas in a socket, and a previous design was altered to include open areas towards which the lattice densities would gradually decrease. Lattice samples were tested for 3D printability based on unit cell type, cell size & thickness. A total of 432 sample conditions were analyzed with the goal of identifying the best-suited lattice candidates for further compression testing and eventual use in a modified CRS socket. Top performing cell types were then investigated further to determine the relationship between unit cell size and thickness with respect to printability. It was observed that as cell size becomes larger, printability decreases, most notably at a thickness less than or equal to 2 mm. Cell sizes 5-15 mm print well at 2-3 mm beam thickness, suggesting lattice choices for further compression testing should be chosen from this range. Going forward, a modified CRS socket will be printed using selected lattice structures for compression areas where the lattice density will taper toward the release openings. Testing will be conducted to determine whether risk of tissue damage can be diminished.

Funding: This work was supported in part through funds from the Orthotic and Prosthetic Education and Research Foundation, Inc. (OPERF).

References:

  1. L. Bennett. "Transferring Load to Flesh," New York University, 1971.
  2. L. Resnik. "Comparison of transhumeral socket designs utilizing patient assessment and in vivo skeletal and socket motion tracking: a case study," Disab and Rehab: Asst Tech, 2016.
  3. R. D. Alley. "Prosthetic sockets stabilized by alternating areas of tissue compression and release," JRRD, 2011.

Presenter: Mary Nguyen

Visualizing the human brain and its structures in three dimensions is a complex and overwhelming task. Sensory pathways such as visual neural pathways, are important but difficult to visualize. The interactive neuroanatomy tool is designed to solve these issues. The tool features an interactive brain that was constructed and modeled using T1-weighted MRI data. Users will learn about the visual neural pathway through brain cross-sections in a three dimensional space. In addition, users will be able to visualize the structures involved in the pathway and be guided through learning modules to understand how vision is processed. With the rising popularity of extended reality in modern education, this tool will allow students or any person interested in Neuroanatomy to learn in a fun and engaging way. Virtual reality enables a user to have the ability to visualize complex structures in the brain in a way that is otherwise, impossible to see in a cadaver lab or web-based resources. By implementing a completely immersive and interactive learning style, students are able to overcome these issues and learn at their own pace with clarity.

Presenter: Michael Nolan

The new Data Management & Sharing Plans being required by NIH funded research programs impact research proposals beyond simply having additional paperwork needed to submit new funding proposals. These new requirements force researchers and their institutions to consider data artifacts to become living pieces of work that may evolve and grow as peers utilize and contribute new additions to these data sets. The process of assessing and effectively utilizing the wide variety of data hosting platforms to develop open data communities requires specialized skill sets, human resourcing, and technical infrastructure. In this poster, we will showcase how Open@RIT's community development methodology can provide NIH funding applicants an edge when it comes to data management and sharing.

Presenter: Natalie Nordlund

Mindfulness-based interventions incorporating art have been found to have various positive impacts on college-age students, including reductions in anxiety, improved mood, and an increased ability to engage in mindfulness-related skills (Carsley & Heath, 2019; Hui & Ma'rof, 2019). In this study, a mindfulness-based group intervention incorporating technology and creative arts has been developed and will be implemented over the course of four weeks with college students. Sessions and activities will incorporate technology and utilize creative/expressive arts, while teaching and practicing mindfulness-related skills. To quantitatively evaluate the impact of the intervention on participants' mindfulness-related skills, pre-post data from the 27-item Balanced Inventory of Mindfulness-related Skills (BIMS) (Padmanabham et al., 2021) will be collected and statistically evaluated. Participant feedback data from a 17-item anonymous participant feedback survey will also be collected and evaluated following the intervention to gain insight regarding participants' experience participating in the group, any feedback/suggestions they may have, and their inclination towards future application of mindful and creative practices and using technology for engaging in those practices. In this presentation, results will be presented and discussed, along with implications and opportunities for further research and future implementation.

Presenter: Hayden Orr

Do clear face masks help reduce the communication barriers that Deaf and Hard of Hearing individuals experience? In this study, we attempt to investigate clear mask accessibility factors, and whether the design of four different clear masks could be improved upon.

A survey was conducted (N=60) with Deaf and Sign Language-using professionals in which they were asked to wear four different brands of clear masks and provide user experience feedback. Subsequently, they were invited to join a focus group (N=30) where their comments and thoughts were collected regarding the comfort, design, and accessibility of the clear masks that were provided.

The data we collected revealed that the discomfort of wearing clear masks outweighed the minor benefit of accessibility of the clear masks provided. Deaf participants generally preferred to wear traditional cloth masks over clear masks; although they largely believed that they would benefit from clear masks if the interpreters or presenters were to use them.

There is little research to date concerning how clear masks affect communication barriers, specifically those affecting Deaf and Sign Language-using individuals. This study attempts to demonstrate that there are considerable design flaws that should be improved upon to increase the comfort and accessibility of the clear mask to better prepare sign-language communities for routine health accessibility and the next pandemic

Presenter: Mariam Paracha

A multidisciplinary RIT and NTID project team has worked to evaluate and develop clear face masks to support communication and inclusion of the Deaf and Hard-of-Hearing (DHH) community. The work has commenced with a design charrette to map a project plan and scope, from which a research team was developed to: 1) Determine and support stakeholder needs and capacities, 2) Evaluate clear face masks in the market, 3) Assess regulatory guidance and conformance testing, 4) Elucidate the scientific and technological challenges inherent to clear face masks, including air movement and filtration, 5) Design a clear face mask prototype using more sustainable materials. The goal of the work is to reduce communication challenges in deaf spaces and medical settings, supporting both deaf patients and deaf medical professionals. Preliminary results include: 1) Basic physical and compositional properties of polymers commonly used in 8 commercially available clear face masks were measured. In addition, application-specific properties of the materials in the clear portion of these masks, such as their wettability and haze were determined. 2) Existing PPE regulatory systems and testing protocols were evaluated to frame compliance and efficacy goals. 3) Reflection of light at different wavelengths in the visible light spectrum on large samples at multiple angles was used to quantify the visibility of the user's face through clear face masks. This test is on large samples and aims at the user's perspective. 4) Existing clear and solid masks were evaluated with regards to construction, sizing, and fit to inform new ideation of a more inclusive clear mask design. 5) Survey-based data on four well-known clear face mask brands were collected from Deaf and Sign Language-using professionals (n=60). The outcome of their user experiences was that the discomfort of wearing clear masks outweighed the minor benefits of accessibility. These initial findings support developing preliminary prototypes of clear face masks that are more effective, safe, and accessible to facilitate communication between DHH individuals and hearing individuals.

Presenter: Stephen Pellow

Nothing leaves a lasting impression more in our youth than seeing their creativity and passion come to life especially when it helps a loved one or close friend.

Sixth grader, Mitchel, learned how to tap into his creativity through his advanced students' class at Brighton Twelve Corners Middle School (TCMS). Following Torrance's 4 criteria for creativity, students are given the task of coming up with a creative idea that would benefit others in their community, an idea they present in a video. Through a partnership with Kids Miracle Making Club (KMMC), a project-based, service-learning organization and the Rochester Institute of Technology, KMMC evaluates the ideas for potential prototype design projects conducted by RIT's Multidisciplinary Senior Design (MSD) program - teams of RIT seniors tapping into their engineering expertise to convert an idea into a physical prototype.

Mitch's grandfather has diabetes and suffers from peripheral neuropathy. To help his grandfather Mitchel came up with a footwear idea that would help reduce the impact of the neuropathy. The idea was proposed by KMMC and accepted by MSD with a team of engineers developing a physical prototype of the design. With support from KMMC, Mitchel was designated co-client participating in the MSD team's 8-step client gate reviews.

The results - A sixth grader learning the value of healthcare technology while watching his idea to help his grandfather and others come to life. A RIT engineering design team combining their engineering expertise to develop a patentable footwear to help prevent foot ulcerations which lead to further complications associated with diabetic peripheral neuropathy - https://www.rit.edu/engineering/seniordesign/projects/diabetic-neuropathy-footwear-device

Presenter: Dan Phillips

Deploying the tubes, cords and other implements in an arthroscopic procedure, so they can maintain their sterile condition, can be challenging, especially if the cords and tubes become tangled. This situation was presented as a problem by Dr. David Carrier and Aaron Burton from Rochester Regional Health as part of the RIT IdeaLab program in the Spring 2021. IdeaLab is a weekend long event where interdisciplinary teams of students meet with the clients to develop innovative solutions to health care and accessibility challenges. If a concept looks promising, further concept development, including proof of concept and prototype solutions, can be carried out at RIT in a 10-week summer program known as Studio930. Like IdeaLab, the development is carried out by an interdisciplinary team comprised of design and engineering students working with faculty coaches and, most importantly, the client, in this case, Dr. Carrier and the surgical staff, nursing and technicians, involved in the actual arthroscopic procedures. The RIT students were allowed to visit the surgical suite and observe the arthroscopic procedures. The concept and prototype that was developed was enthusiastically received by Dr. Carrier. It also included accessory devices such as clamps and tube organizers to facilitate the placement of the tubes and cords in the surgical area. The logical next step would be to refine the prototype design so that it could actually be considered for manufacturing. This was made possible by the awarding of a grant from the Kilian J. and Caroline F. Schmitt Foundation that enabled the hiring of RIT students by a local design firm, BZ Design, through the RIT cooperative employment program. This enabled the students to work on refining the design for manufacturing in a professional environment with experts in design for manufacturing. The result was a realizable design of a sterilizable container that arranged the tubes, cords and implements in a fashion that enable the surgical team to deploy the necessary items in an organized and efficient fashion that can save time and effort. The client was presented with a prototype device and a design package that included estimated costs and recommended manufacturing resources that would enable an informed decision as to next steps that could include a Limited Rate Initial Production (LRIP) as well as potential intellectual property considerations.

Presenter: Leon Reznik

Machine learning based image classifiers are widely employed in the medical industry with applications including disease diagnosis, tumor classification, treatment planning, and many others. In many cases medical images and data need to be transmitted from their collection points to cloud-based services, which perform their recognition and other tasks over communication channels and networks. Over the communication, the images could be negatively impacted by network conditions or be a subject to an adversarial attack resulting in their quality degradation. This paper evaluates the impact of input data quality on the recognition performance of industrial ML image classifiers like YOLO, Faster-RCNN, Google Cloud Vision, and AWS Rekognition on image datasets. In this paper, we consider two major types of attacks: diverse zero-knowledge attacks and knowledge-based attacks against ML systems. We investigate how image corruption caused by the conventional network attacks, such as DDoS, affect the performance of object detectors. In addition, we review the detection effectiveness of image classifiers impacted by adversarial attacks. Finally, we review the results of our empirical study, and provide recommendations based on the observed performance. We found that DDoS attacks have the most substantial effect on the detectors' performance. As a generic way to improve resilience against network attacks, we advise switching transport protocols depending on network conditions with recommendations given.

Presenter: Jennifer Schneider

The COVID-19 pandemic highlighted the ill-prepared and brittle PPE lifecycle, encompassing challenges with supply & selection, access, design, performance, inclusion, and waste generation, in addition to the long-faced issues with "shrink and pink" (not just downscaled) and "no and throw" (nothing for you and one use) approaches. Most of these issues were not new, but all were exacerbated by this global crisis (National Academy of Science). Between 2019 and 2020 alone, global demand for PPE increased 300-400% (UK Aid). The ecosystem for this highly regulated product class is complex and has high impact. This ongoing research project has several interconnected components: 1) improving the global regulatory landscape to be responsive yet protective, 2) creating designs that are broadly inclusive, as women, children, and disabled (Deaf and Hard of Hearing [DHH], communication disorders, etc.) currently lack accessible options 3) developing new technologies and materials that can serve multiple hazards, 4) creating new technical capacity for fit and performance (sensors, print to fit, etc.), and 5) improving supply, lifespan and sustainability. In 2021, with support of the National Technical Institute for the Deaf, RIT kicked off the Clear Mask project to respond to the lack of inclusive PPE for our DHH community. From that work, a broader research initiative has commenced, to form a consortium (engineers, physicists, material scientists, industrial designers, scientists, and more from across the university) to leverage RIT expertise across the PPE landscape addressing this broad, critical need. Simply, we must do better.

Initial research questions include:

  • To what extent are US and global PPE standards aligned and responsive?
  • What gaps and challenges must be resolved?
  • How can new performance-based testing and qualification methods be developed to ensure safety and health?
  • What characteristics must next gen PPE materials have to serve the broader user base?
  • How must we expand, serve and engage with stakeholders to ensure health?
  • How can design drive an improved product lifespan across manufacturing, use and sustainability?

Preliminary Results:

  • Regulatory Analysis
  • Analysis of currently available clear mask products
  • Testing parameters
  • Design criteria for next gen clear and other PPE

National Academies of Sciences, Engineering, and Medicine. 2022. Frameworks for Protecting Workers and the Public from Inhalation Hazards. Washington, DC: The National Academies Press. https://doi.org/10.17226/26372

UK Aid. Covid-19 - PPE Demand & Supply Perspectives. Mar. 2021, https://tinyurl.com/2r8f53n4.

Presenter: David Schwartz

RIT can provide a vital link for the upcoming conference, especially regarding "gaming." As RIT's School of Interactive Games and Media (IGM) Director, I receive frequent requests for research collaboration with our faculty and hundreds of students, especially for healthcare. Given the confusion of what game design and development entail, my proposed talk would provide a primer to facilitate partnerships more quickly.

RIT's world-renowned game design and development program has produced published research, including creative work, spanning various topics. While some research has focused on emerging and new forms of games, other work has applied to other research domains, like cybersecurity, physics, education, marketing, religion, and more. For more information, see https://www.rit.edu/computing/school-interactive-games-and-media/research and games.rit.edu (scroll down to the research links below "Study").

For my proposed talk, I would cover the basics of games and game research, what gamification can and can't do (and might do), research resources, and places to publish. I especially wish to cover how to engage with RIT's enormous population of game students who have a great interest in helping people but lack a background in healthcare technology. Students can rapidly prototype low-cost solutions that help focus researchers and developers. I will provide all the resources to all attendees to share with their organizations.

[See https://drive.google.com/drive/u/2/folders/1LzfyNWhKAef4jjAIfZvB5AlDOkfkdai1 and https://home.liebertpub.com/publications/games-for-health-journal/588/overview for two examples.]

Presenter: Megan Scroger

Substance use disorders (SUDs) and intimate partner violence (IPV) are highly comorbid and require integrated, innovative interventions (Crane et al., 2014). Technology-assisted interventions can improve in-person treatment engagement, homework completion, and understanding of concepts and skills, while reinforcing skill acquisition and practice (Carroll & Kiluk, 2017). Technology can also standardize the dose of cognitive-behavioral therapy, (CBT) while maintaining cost-effectiveness and individualization (Carroll & Kiluk, 2017). Avatars, or digital health coaches, are well-suited to assisting with SUD and IPV and can aid in administering CBT content, practice exercises, and symptom/behavior monitoring (Easton et al., 2018). Customizing avatar features may also garnish personal investment, increase motivation, and deploy skills in a non-confrontational manner. We hypothesized that a novel avatar-assisted CBT intervention (RITch®-CBT) would be acceptable to adult patients and feasible for use in an inpatient treatment setting. This study utilized qualitative and quantitative methods to test acceptability and feasibility of a two-session avatar-assisted, integrated CBT platform for adults with SUD and IPV. The platform includes a customizable avatar, standardized CBT-based session content (i.e., functional analysis, narrated coping skill activities and exercises) and between session monitoring of thoughts, feelings, and behaviors (e.g., mood, cravings, substance use). Ten participants completed digital sessions and provided their feedback regarding the platform. All participants reported that they agreed or strongly agreed with positive statements about the platform (i.e., avatar was genuine/relatable, helpful content, rewarding) and 90% were willing to use it with their therapist, enjoyed interacting with the avatar, and thought it was easy to use. The RITch®-CBT avatar platform showed preliminary acceptability within this small sample with polysubstance use and psychiatric symptoms. Participant feedback regarding the platform has been incorporated and testing in outpatient settings across demographic groups is ongoing.

Presenter: Matthew Thai

Ensuring healthcare data security and privacy is crucial in embracing the Digital Future of healthcare. This is especially so in acute care which provides short-term care under urgent medical situations. Due to its emergency response nature, acute care prioritizes speed and accessibility so that patient medical records can be quickly accessed to ensure proper care. This priority for speed naturally puts patient data at a vulnerable spot as security and privacy are sacrificed to "break the glass" and access needed data. The novel mechanism, Acute Care: Attribute-Based Access Control (AC-ABAC) system, was recently proposed as a high-granularity role-based access control management solution for healthcare providers to dynamically access patient data with higher levels of data security and privacy. To evaluate the AC-ABAC model, this project implemented its core features as an experiment, including a lightweight encryption protocol to further secure data transactions. Preliminary assessment of this project is that AC-ABAC represents a promising approach, as it provides considerable data security and privacy benefits within acute care situations, and that further research is merited.

Presenter: Zhihong Zhang

Infant mortality is an essential indicator of a society's degree of development, yet it is challenging to predict. To predict infant death, this study first examines the impact of the COVID-19 pandemic on infant mortality. Then, a machine learning model called XGBoost with the original imbalanced dataset is employed to predict infant and neonatal mortality and classify the risk of death into high, medium, and low-risk groups. Furthermore, this study analyzes monthly data to assess the predictive ability of our prediction model and identify the important predictors of infant and neonatal death, which can provide direct guidance on the prevention of infant death.