arXiv subjects

Alyssa Donawa

Publications and source records attributed to Alyssa Donawa.

3 recordsLinked to original sources

arXiv2026

On-Device Language Models for Privacy-Preserving Stress Prediction: A Multimodal Evaluation on Mobile Health

Stress is a pervasive determinant of mental health and a key target for mobile health interventions. On-device language models (ODLMs) offer privacy-preserving inference without cloud dependency, yet their feasibility for health prediction under mobile resource constraints remains underexplored. We evaluate ODLMs for multi-modal stress prediction using zero-shot prompting, measuring predictive accuracy alongside latency and throughput. Our results show that objective sensor features marginally outperform subjective self-reports on average, and that lightweight sub-2B models achieve low latency with predictable resource usage. Our findings highlight both the promise and the practical constraints of ODLMs for mobile mental health.

cs.LG↗

arXiv2021

Addressing the Need for Remote Patient Monitoring Applications in Appalachian Areas

There is a need to address the urban-rural disparities in healthcare regarding equal access and quality of care. Due to higher rates of chronic disease, reduced access to providers, and a continuous decline in rural hospitals, it is imperative that Appalachian cancer patients adopt the use of health information technology (HIT). The NCCN Distress Thermometer and Problem List (DT) is under-utilized, not patient-centered, does not consider provider needs, and is outdated in the current digital landscape. Digitizing patient distress screening poses advantages, such as allowing for more frequent screenings, removing geographical barriers, and rural patient autonomy. In this paper, we discuss how knowledge gained from patient-centered design led to the underpinnings of developing a rural remote patient monitoring app that provides delightful and insightful experiences to users.

cs.HC↗

arXiv2020

Scaling Blockchains to Support Electronic Health Records for Hospital Systems

Electronic Health Records (EHRs) have improved many aspects of healthcare and allowed for easier patient management for medical providers. Blockchains have been proposed as a promising solution for supporting Electronic Health Records (EHRs), but have also been linked to scalability concerns about supporting real-world healthcare systems. This paper quantifies the scalability issues and bottlenecks related to current blockchains and puts into perspective the limitations blockchains have with supporting healthcare systems. Particularly we show that well known blockchains such as Bitcoin, Ethereum, and IOTA cannot support transactions of a large scale hospital system such as the University of Kentucky HealthCare system and leave over 7.5M unsealed transactions per day. We then discuss how bottlenecks of blockchains can be relieved with sidechains, enabling well-known blockchains to support even larger hospital systems of over 30M transactions per day. We then introduce the Patient-Healthchain architecture to provide future direction on how scaling blockchains for EHR systems with sidechains can be achieved.

cs.CR↗