Blood Pressure | Spatial Control | Public Perception: UMSI Research Roundup
Monday, 07/28/2025
By Noor HindiUniversity of Michigan School of Information faculty and PhD students are creating and sharing knowledge that helps build a better world. Here are some of their recent publications.
Publications
Loopsim: enrichment analysis of chromosome conformation capture with fast empirical distribution simulation
NAR Genomics & Bioinformatics, September 2025
Gideon Shaked, Haihan Zhang, Zhaolin Zhang, Jiayu Zhou, Johann E Gudjonsson, James T Elder, Matthew T Patrick, Lam C Tsoi
Gene regulation is intricately influenced by the three-dimensional organization of the genome. In particular, chromatin can exist in loop structures that enable long-range regulatory interactions. By utilizing chromosome conformation capture techniques such as Hi-C, valuable information regarding the organization of these loop structures in 3D space can be obtained. Although functional/feature enrichment is now a common downstream analysis for various genomic platforms to provide biological context, tools specifically designed for high-throughput assays that capture chromosome conformation remain relatively limited. Here, we present Loopsim, a command-line application that performs enrichment analysis on Hi-C loop profiles against user-defined regions (available on GitHub at https://github.com/CutaneousBioinf/Loopsim). Loopsim efficiently simulates a background distribution using a distinctive sampling approach that considers loop size, intervals, loop–loop distances, and structure; it then computes statistics based on the empirical null distribution.
Unlocking efficiency in real-world collaborative studies: a multi-site international study with one-shot lossless GLMM algorithm
Npj digital medicine, July 2025
Jiayi Tong, Jenna M. Reps, Chongliang Luo, Yiwen Lu, Lu Li, Juan Manuel Ramirez-Anguita, Milou T. Brand, Scott L. DuVall, Thomas Falconer, Alex Mayer Fuentes, Xing He, Michael E. Matheny, Miguel A. Mayer, Bhavnisha K. Patel, Katherine R. Simon, Marc A. Suchard, Guojun Tang, Benjamin Viernes, Ross D. Williams, Mui van Zandt, Fei Wang, Jiang Bian, Jiayu Zhou, David A. Asch & Yong Chen
The widespread adoption of real-world data has given rise to numerous healthcare-distributed research networks, but multi-site analyses still face administrative burdens and data privacy challenges. In response, we developed a Collaborative One-shot Lossless Algorithm for Generalized Linear Mixed Models (COLA-GLMM), the first-ever algorithm that achieves both lossless and one-shot properties. COLA-GLMM ensures accuracy against the gold standard of pooled data while requiring only summary statistics and completes within a single communication round, eliminating the usual back-and-forth overhead. We further introduced an enhanced version that employs homomorphic encryption to reduce the risks of summary statistics misuse at the coordinating center. The simulation studies showed near-exact agreement with the gold standard in parameter estimation, with relative differences of 7.8 × 10−6 %–3.0% under various cell suppression settings. We also validated COLA‑GLMM on eight international decentralized databases to identify risk factors for COVID‑19 mortality. Together, these results show that COLA‑GLMM enables accurate, low‑burden, and privacypreserving multi‑site research.
Physical activity and diet just-in-time adaptive intervention to reduce blood pressure: a randomized controlled trial
npj Digital Medicine, July 2025
Michael P. Dorsch, Jessica R. Golbus, Rachel Stevens, Brad Trumpower, Tanima Basu, Evan Luff, Kimberly Warden, Michael Giacalone, Sarah Bailey, Gabriella V. Rubick, Sonali Mishra, Predrag Klasnja, Mark W. Newman, Lesli E. Skolarus, Brahmajee K. Nallamothu
Mobile health interventions for behavioral change require large-scale studies to ensure their clinical benefits. We conducted a randomized controlled trial of patients with hypertension to assess the myBPmyLife application in promoting physical activity and lower-sodium foods to lower systolic blood pressure (SBP). 602 participants were randomly assigned to either a control group or an intervention group that received the myBPmyLife application. For the primary outcome, change in SBP over 6 months was −5.2 mmHg in the intervention and −5.7 mmHg in the control group (p = 0.76). For secondary outcomes, the intervention group increased their daily step count by 170 steps, while the control group decreased by 319 steps (p = 0.040). Sodium intake decreased by 1145 mg in the intervention and 860 mg in the control group (p = 0.002). The myBPmyLife application did not reduce SBP over 6 months in hypertension patients despite increasing step counts and lowering sodium intake.
COLA-GLM: collaborative one-shot and lossless algorithms of generalized linear models for decentralized observational healthcare data
Npj digital medicine, July 2025
Qiong Wu, Jenna M. Reps, Lu Li, Bingyu Zhang, Yiwen Lu, Jiayi Tong, Dazheng Zhang, Thomas Lumley, Milou T. Brand, Mui Van Zandt, Thomas Falconer, Xing He, Yu Huang, Haoyang Li, Chao Yan, Guojun Tang, Andrew E. Williams, Fei Wang, Jiang Bian, Bradley Malin, George Hripcsak, Martijn J. Schuemie, Yun Lu, Steve Drew, Jiayu Zhou, David A. Asch, Yong Chen
Clinical insights from real-world data often require aggregating information from institutions to ensure sufficient sample sizes and generalizability. However, patient privacy concerns only limit the sharing of patient-level data, and traditional federated learning algorithms, relying on extensive back-and-forth communications, can be inefficient to implement. We introduce the Collaborative One-shot Lossless Algorithm for Generalized Linear Models (COLA-GLM), a novel federated learning algorithm that supports diverse outcome types via generalized linear models and achieves results identical to a pooled patient-level data analysis (lossless) with only a single round of aggregated data exchange (one-shot). To further protect aggregated institutional data, we developed a secure extension, secureCOLA-GLM, utilizing homomorphic encryption. We demonstrated the effectiveness and lossless property of COLA-GLM through applications to an international influenza cohort and a decentralized U.S. COVID-19 mortality study. COLA-GLM and secure-COLA-GLM offer a scalable, efficient solution for decentralized collaborative learning involving multiple data partners and diverse security requirements.
Designing for Instant Convenience: Dark Stores, Spatial Control and Worker Quantification in On-demand Delivery Platforms
DIS '25: Proceedings of the 2025 ACM Designing Interactive Systems Conference, July 2025
Abhishek Sekharan, Matthew Bui, Julie Hui
This paper critically examines the design practices of platforms like DoorDash that promise to deliver convenience to consumers’ doorsteps within minutes. Based on ethnographic observations in DoorDash dark stores and qualitative interviews with DoorDash workers, we find that instant delivery platforms have several work- and customer-specific impacts due to how platforms govern space and broader geographies. These platforms computationally remap city spaces to reorient and accelerate the flow of goods and people across them, subsequently exacerbating intra-urban disparities in access to food and groceries. Such centralized surveillance and management of space is made possible through hyper-quantified labor processes that severely limit worker agency inside dark stores and intensify working conditions. Furthermore, minute-tracking of worker output and mobilities, and their hyper-quantified evaluations eclipse much of the invisible work that sustains the veneer of instant deliveries. In arguing for a worker-centered design of gig platforms and their service delivery models, we end with three crucial dimensions that critical design researchers and platform designers need to account for when making things ’convenient.’
Designing for Discourse: Social Media, Socio-Technical Rhetorical Strategies, and Affirmative Action Discussions
DIS '25: Proceedings of the 2025 ACM Designing Interactive Systems Conference, July 2025
Cassidy Pyle, Nicole B. Ellison, Nazanin Andalibi
Social media platforms enable diverse users to engage in everyday political talk with (un)known audiences. Platform features and affordances may shape political discussions and how audiences make sense of them, potentially shifting political attitudes. Using affirmative action (AA) – a controversial, identity-centric higher education policy – as a context for analysis, we investigate social media features’ and affordances’ role in AA discussions. Our qualitative content analysis of over 38,000 social media posts and comments across Reddit, Twitter/X, and TikTok demonstrates how features (e.g., Green Screen) and affordances (e.g., visibility) shape the presentation of external evidence and cues on social media that help users determine information veracity. We introduce socio-technical rhetorical strategies to describe rhetorical devices enabled by platform features and affordances and consider how these strategies are used to express and refute racism online. Finally, we suggest ways that social media designers may leverage visibility, navigability, and association affordances to enhance users’ ability to make sense of and safely experience AA discussions.
Towards Accounting for Nonhuman Agency in Technology Design for Aging
Alisha Pradhan, Shaan Chopra, Pooja Upadhyay, Robin Brewer, Amanda Lazar
Till date, research on technology design to support aging has largely adopted human-centered approaches, such as user-centered and participatory design. While this body of work continues to yield insights on technologies for older individuals, recent research is beginning to question whether other ways of knowing, that are not solely “humanistic”, can yield new understandings on technologies for aging. We pick up this call, and present emerging findings from our interviews with older adults around their use of home-based intelligent voice assistants. Building upon these findings, we discuss how attuning to the nonhuman agency and the larger material world can offer new insights on technology design for aging, such as, speak to issues of technology non-use by older adults. We discuss implications of assigning responsibility and accountability beyond older adults to nonhuman actors and how it can speak to the deficit-centered narrative that often gets associated with aging.
Designing Psychologically Safe Online Spaces: Supporting Aspiring Entrepreneurs in Financially-Constrained Communities
DIS '25: Proceedings of the 2025 ACM Designing Interactive Systems Conference, July 2025
Aarti Israni, Tawanna R Dillahunt
In fields like Human-Computer Interaction designing for safety often focuses on mitigating interpersonal harms such as online harassment. While essential, this emphasis may overlook the psychological aspects of safety, including trust. These aspects are especially important in marginalized contexts. This study explores the factors that foster psychological safety among financially-constrained aspiring entrepreneurs. Using a case-study approach, we analyzed data (10 interviews, 3 focus groups) from 21 financially-constrained aspiring entrepreneurs who engaged with a community-based sociotechnical platform. We identified four key factors that shaped participants’ sense of psychological safety: shared identity, interpersonal trust, perceived expertise, and shared accountability. These insights inform design strategies emphasizing shared accountability and leveraging collective identities to empower users in underserved communities. While centered on financially-constrained entrepreneurs, these strategies offer broader applications and should be considered when creating online spaces that promote connection, collaboration, and psychological safety across diverse contexts.
Rewarding Trust: How Reward Power Shapes Security Robot Acceptance
Proceedings of the Human Factors and Ergonomics Society Annual Meeting, July 2025
As security robots take on more societal roles, public resistance can hinder their effectiveness. This study examines how a security robot’s ability to offer rewards (“reward power”) affects public acceptance and trust, which is vital for integrating robots into communities. Using a between-subjects experiment with 106 participants, we tested the impact of high versus low reward power through online video interactions. The results showed that reward power significantly increased robot acceptance by fostering trust during initial interactions. This research contributes to the field of human-security-robot interaction by highlighting the importance of reward power in building trust and acceptance. These findings provide design guidelines for improving public trust and acceptance, essential for successful real-world deployments of security robots.
The Vibes are Off: Considering Embodied Reflections by TBIPOC to Account for Displacement and Discomfort in Makerspaces
DIS '25 Companion: Designing Interactive Systems Conference, July 2025
F. Ria Khan, Apryl Williams, Tawanna R. Dillahunt, Oliver L. Haimson
Makerspaces are identified in HCI to have great potential in fostering diverse participation in technology and computing–considering making as a democratic form of innovation. However, growing research also indicates many current makerspaces fail to address non-white and non-cisheteronormative perspectives. Prior works suggest embodiment as a core but seldom understood consideration for intersectional inclusion. Current trends in technologies and computing also stifle such considerations through two phenomena: broader implications of “woman lite” thinking, and what this provocation defines as “techno-disembodiments.” To combat perpetuating these phenomena in makerspaces, we posit looking to bodily and sensory responses, or embodied reflections, from communities of Trans and Black, Indigenous and People of Color (TBIPOC). Further, we examine prior works in visual auto-ethnography and diary studies for approaches to inquire about embodied reflections. In considering embodied reflections of TBIPOC communities, researchers can gain insights on decentering cisheteronormative whiteness to afford broader inclusion in maker culture.
Designing markets, governing data: Engineering value in the American healthcare system
Big Data and Society, July 2025
In crucial sectors like healthcare, education, and housing, policymakers are turning to the tools of market design to incentivize public and private actors to more efficiently and effectively produce the public good. Although market design has been a key policymaking tool for decades, datafication is increasingly central to this technocratic tinkering. This article explores a project of datafied market redesign in the U.S. healthcare industry, demonstrating that emerging federal health data regulations are designed to enable the state to more precisely quantify, and thereby incentivize, the production of “valuable” care. This case study demonstrates how both the public good and crucial data infrastructures are constrained through their enactment within market-based modes of governance. As this data-solutionism for extractive markets becomes a more prevalent mode of governance—particularly in areas like climate change—we must find alternative mechanisms for collectively defining the public good, and for achieving corporate accountability beyond financial incentive structures.
Leveraging Group-Level Signals for Robust Many-Domain Generalization
Advances in Knowledge Discovery and Data Mining, June 2025
Bohan Zhang, Yachuan Liu, Qiaozhu Mei, Paramveer S. Dhillon
Many-domain generalization poses a significant challenge for machine learning models, as real-world data often comprises numerous domains in both training and unseen test sets. Traditional approaches that focus on a limited number of domains fail to capture this complexity, leading to suboptimal performance. We propose a novel group-wise reweighting strategy that leverages diverse group-level features—including label entropy, representation statistics, and gradient properties—to address this problem. By determining group importance during training based on these features, our approach overcomes limitations of existing methods that rely solely on group error. Our results demonstrate significant improvements in worst-group and tail performance on out-of-sample test data across multiple datasets, validating our selective upweighting strategy. We further implement a learning-to-rank framework that integrates multiple group features. While this approach achieves substantial gains over empirical risk minimization, the challenges in consistently outperforming individual features highlight the inherent difficulties in achieving transferable robustness amid varying group characteristics across datasets. Finally, SHAP analysis confirms the heterogeneous importance of different features, emphasizing the need for adaptive strategies.
A community health worker-delivered intervention (STEPS) to support chronic pain self-management among older adults in an underserved urban community: protocol for a randomized trial
Trials, June 2025
Mary R. Janevic, Rebecca Lindsay, Elizabeth Brines, Kimberlydawn Wisdom, Sheria G. Robinson‑Lane, Robin Brewer, Susan L. Murphy, John Piette, Leslie Grijalva, Michael Anderson, Jaye Clement, Courtney Latimer
Background: Older adults in disadvantaged urban communities contend with chronic psychosocial and environmental stressors that contribute to high levels of chronic pain-related disability. African American older adults are especially at risk due to the health-damaging effects of structural racism. The purpose of this study is to test the ef‑ cacy of a chronic pain self-management intervention tailored for this context. STEPS (Seniors using Technology to Engage in Pain Self-management) is a community health worker (CHW)-led chronic pain self-management pro‑ gram designed for older adults living in underserved communities. It is a 7-week intervention that includes (a) brief videos presenting pain self-management skills; (b) weekly telephone calls with a CHW to support the practice of new skills and goal setting; and (c) tracking daily step counts using a wearable activity tracker. CHWs also screen for social needs and make appropriate community referrals.
Methods: We will randomly assign 414 participants to the STEPS intervention or a control condition in a 1:1 ratio, stratifying by gender and age group. We hypothesize that participants in the STEPS intervention will have greater improvements in pain interference and pain intensity, and a more positive Global Impression of Change immediately following the intervention and at 12 months from baseline. Control group members are invited to attend a workshop covering key intervention content after the final data collection point.
Discussion: Growing evidence supports the effectiveness of CHWs as culturally sensitive liaisons between healthcare systems and underserved communities. If the STEPS program is shown to significantly improve pain-related outcomes, STEPS could be integrated into healthcare systems to more comprehensively treat chronic pain while reduc‑ ing barriers to care and promoting non-pharmacological pain management strategies.
Counting days is a spacing incentive that unlocks the potential of low GPA students
npj, Science of Learning, June 2025
Iman Yeckeh Zaare, Paul Resnick
Spacing and retrieval practice enhance learning, but students often underuse these strategies. We tested a simple grading incentive, which we call Counting Days, in two RCTs: one randomizing 143 students within a course and another randomizing 71 instructors. The “counting questions” control condition awarded points for each practice question answered, while the “counting days” treatment assignment awarded points for each day that a student answered a set of questions. In the within-class experiment, the counting days group earned higher exam scores, mediated by spacing practice over more days. Spacing was especially beneficial for lower-GPA students: the correlation between course exam scores and GPA in prior courses was significantly lower for the counting days group. In the between-instructor experiment, there was no way to compare learning outcomes between instructors, but both the number of days and a number of questions practiced were significantly higher under the counting days condition.
The administrative burden of medication affordability resources: an environmental scan with implications for health informatics to advance health equity
Jamia, June 2025
Marcy G. Antonio, Jennylee Swallow, Rachel Richesson, Christine Carethers, Antoinette B. Coe, Divya Jahagirdar, Yung-Yi Huang, Tammy Toscos, Mindy Flanagan, Tiffany C. Veinot
Objective: To characterize and demonstrate how to reduce the administrative burden experienced by patients when navigating medication affordability resources in the United States.
Materials and Methods: Informed by administrative burden theory, we conducted an environmental scan of medication affordability resources for atrial fibrillation, and four common comorbidities (diabetes, heart failure, hypertension, and lipid disorder). We systematically searched for resources (eg, patient assistance programs, savings cards and nonprofit support) and extracted information about types, eligibility criteria, needed documentation, and application processes. Results: We identified 66 resources across 12 categories across the five conditions. The resources’ varied eligibility criteria, application processes, and requirements for providing sensitive financial documents could introduce multiple administrative costs for patients.
Discussion: The volume and complexity of medication affordability resources and related application processes may create substantial administrative burden for patients that could prevent their use—especially when prescribed multiple medications.
Conclusion: Medication affordability resource informatics tools that reduce administrative burden could advance equitable medication access.
Perceived legitimacy of layperson and expert content moderators
PNAS Nexus, May 2025
Cameron Martel, Adam J. Berinsky, David G. Rand, Amy X. Zhang, Paul Resnick
Content moderation is a critical aspect of platform governance on social media and of particular relevance to addressing the belief in and spread of misinformation. However, current content moderation practices have been criticized as unjust. This raises an important question—who do Americans want deciding whether online content is harmfully misleading? We conducted a nationally representative survey experiment (n = 3,000) in which US participants evaluated the legitimacy of hypothetical content moderation juries tasked with evaluating whether online content was harmfully misleading. These moderation juries varied on whether they were described as consisting of experts (e.g. domain experts), laypeople (e.g. social media users), or nonjuries (e.g. computer algorithm). We also randomized features of jury composition (size and necessary qualifications) and whether juries engaged in discussion during content evaluation. Overall, participants evaluated expert juries as more legitimate than layperson juries or a computer algorithm. However, modifying layperson jury features helped increase legitimacy perceptions—nationally representative or politically balanced composition enhanced legitimacy, as did increased size, individual juror knowledge qualifications, and enabling juror discussion. Maximally legitimate layperson juries were comparably legitimate with expert panels. Republicans perceived experts as less legitimate compared with Democrats, but still more legitimate than baseline layperson juries. Conversely, larger lay juries with news knowledge qualifications who engaged in discussion were perceived as more legitimate across the political spectrum. Our findings shed light on the foundations of institutional legitimacy in content moderation and have implications for the design of online moderation systems.
Identifying progression subphenotypes of Alzheimer’s disease from large-scale electronic health records with machine learning
Journal of Biomedical Informatics, May 2025
Manqi Zhou, Alice S. Tang, Hao Zhang, Zhenxing Xu, Alison M.C. Ke, Chang Su, Yu Huang, William G. Mantyh, Michael S. Jaffee, Katherine P. Rankin, Steven T. DeKosky, Jiayu Zhou, Yi Guo, Jiang Bian, Marina Sirota, Fei Wang
Objective: Identification of clinically meaningful subphenotypes of disease progression can enhance the understanding of disease heterogeneity and underlying pathophysiology. In this study, we propose a machine learning framework to identify subphenotypes of Alzheimer’s disease progression based on longitudinal real-world patient records.
Methods: The framework, dynaPhenoM, extracts coherent clinical topics across patient visits and employs a timeaware latent class analysis to characterize subphenotypes. We validated dynaPhenoM using three patient databases with a total of 3952 AD patients across the United States, demonstrating its effectiveness in revealing mild cognitive impairment (MCI) progression to AD.
Results: Our study identified five subphenotypes associated with distinct organ systems for disease progression from MCI to AD, including common subtypes across cohorts—respiratory, musculoskeletal, cardiovascular, and endocrine/metabolic—as well as a cohort-specific digestive subtype.
Conclusion: Our study unravels the complexity and heterogeneity of the progression from MCI to AD. These findings highlight disease progression heterogeneity and can inform both diagnostic and therapeutic strategies, thereby advancing precision medicine for Alzheimer’s disease.
Pre-prints, Working Papers, Articles, Workshops and Talks
What Happens When Data Centers Come to Town?
Science, Technology, and Public Policy (Ford School of Public Health), July 2025
Duc Tuan "Terry" Nguyen, Ben Green
The rapid growth of data centers, with their enormous energy and water demands, necessitates targeted policy interventions to mitigate environmental impacts and protect local communities. To address these issues, states with existing data center tax breaks should adopt sustainable growth policies for data centers, mandating energy audits, strict performance standards, and renewable energy integration, while also requiring transparency in energy usage reporting. “Renewable energy additionality” clauses should ensure data centers contribute to new renewable capacity rather than relying on existing resources.
The Teacher’s Dilemma: Balancing Trade-Offs in Programming Education for Emergent Bilingual Students
arXiv, June 2025
Emma R. Dodoo, Tamara Nelson-Fromm, Mark Guzdial
K-12 computing teachers must navigate complex trade-offs when selecting programming languages and instructional materials for classrooms with emergent bilingual students. While they aim to foster an inclusive learning environment by addressing language barriers that impact student engagement, they must also align with K-12 computer science curricular guidelines and prepare students for industry-standard programming tools. Because programming languages predominantly use English keywords and most instructional materials are written in English, these linguistic barriers introduce cognitive load and accessibility challenges. This paper examines teachers' decisions in balancing these competing priorities, highlighting the tensions between accessibility, curriculum alignment, and workforce preparation. The findings shed light on how our teacher participants negotiate these trade-offs and what factors influence their selection of programming tools to best support EB students while meeting broader educational and professional goals.
Modeling Public Perceptions of Science in Media
arXiv, June 2025
Jiaxin Pei, Dustin Wright, Isabelle Augenstin, David Jurgens
Effectively engaging the public with science is vital for fostering trust and understanding in our scientific community. Yet, with an ever-growing volume of information, science communicators struggle to anticipate how audiences will perceive and interact with scientific news. In this paper, we introduce a computational framework that models public perception across twelve dimensions, such as newsworthiness, understandability, importance, and surprisingness. Using this framework, we create a large-scale science news perception dataset with 10,489 annotations from 2,101 participants who rated 1,506 science news stories. This dataset, sourced from diverse US and UK populations, provides valuable insights into public responses to scientific information across multiple domains. We further develop natural language processing models that predict public perception scores with a strong performance. Leveraging the dataset and model, we examine public perception of science from two perspectives: (1) Perception as an outcome: What factors affect the public perception of scientific information? (2) Perception as a predictor: Can we use the estimated perceptions to predict public engagement with science? We find that individuals’ frequency of science news consumption is the strongest driver of perception, whereas demographic factors exert minimal influence. More importantly, through a large-scale analysis and carefully designed natural experiment on Reddit, we demonstrate that the estimated public perception of scientific information has direct connections with the final engagement pattern. Posts with more positive perception scores receive significantly more comments and upvotes, which is consistent across different scientific information and for the same science, but are framed differently. Overall, this research underscores the importance of nuanced perception modeling in science communication, offering new pathways to predict public interest and engagement with scientific content.
Validation of the Critical Reflection and Agency in Computing Index: Do Computing Ethics Courses Make a Difference?
arXiv, June 2025
Aadarsh Padiyath, Casey Fiesler, Mark Guzdial, Barbara Ericson
Computing ethics education aims to develop students’ critical reflection and agency. We need validated ways to measure whether our efforts succeed. Through two survey administrations (N=474, N=464) with computing students and professionals, we provide evidence for the validity of the Critical Reflection and Agency in Computing Index. Our psychometric analyses demonstrate distinct dimensions of ethical development and show strong reliability and construct validity. Participants who completed computing ethics courses showed higher scores in some dimensions of ethical reflection and agency, but they also exhibited stronger techno-solutionist beliefs, highlighting a challenge in current pedagogy. This validated instrument enables systematic measurement of how computing students develop critical consciousness, allowing educators to better understand how to prepare computing professionals to tackle ethical challenges in their work.
RELATED
Keep up with research from UMSI experts by subscribing to our free research roundup newsletter!