University of Toronto

Health & Care

Figure from "From patient voices to policy: data analytics reveals patterns in Ontario's Hospital feedback"
Patterns in over 120,000 de-identified patient reviews from 45 Ontario hospitals. From From patient voices to policy (PLOS Digital Health 2026).

Health is shaped as much by social, cultural, and economic conditions as by clinical care. We design and study technologies for care in low-resource settings, approaches to mental health that go beyond Western individualistic models, and the barriers that stigma and gender norms create in access to care.

Mobile Care for Children with Autism

Caring for children with autism spectrum disorder is especially challenging in low- and middle-income countries such as Bangladesh. mCARE is a mobile-based care system that routinely collects behavioural and developmental data, supporting data-driven decisions by families and practitioners (Pervasive Health 2024). Using mCARE data, we have applied machine learning to inform developmental milestone achievement (JMIR Medical Informatics 2021) and studied patterns of risk and resilience during COVID-19 lockdowns (Scientific Reports 2021).

Caregivers also face stigma. Working with 29 parents and nine teachers at special needs schools in Dhaka, "We Need More Power to Stand Up" (ICTD 2020) designs to combat it at its deeper roots.

Related papers

Mental Health and Wellbeing

Process of Generating Personalized LLM-Enhanced Narratives: User inputs their circumstances through structured and open-ended responses, which are combined with
Features for drafting an email to future self: (a) an open-ended text box for drafting the message, (b) a feature enabling open-ended modifications to the draft
Left: how personalized LLM-enhanced narratives are generated from a user's circumstances (DIS 2025). Right: an LLM-supported tool for drafting a message to one's future self (CHI 2024).

Large language models can personalize support in ways traditional interventions cannot. Understanding the Role of LLMs in Personalizing and Scaffolding Strategies to Combat Academic Procrastination (CHI 2024, Best Paper Honorable Mention) explores what users expect from such tools, and Perfectly to a Tee (DIS 2025) uses LLMs to tailor stories about overcoming personal struggles, studied with 346 young adults. We also contributed to a crowdsourced megastudy of 12 digital single-session interventions for depression (Nature Human Behaviour 2026).

Outside the West, mental health needs different framings. Studying Kaan Pete Roi, a suicide prevention helpline in Bangladesh, "What's the Point of Having this Conversation?" (COMPASS 2023) argues for decolonizing mental health services, and an analysis of 1,000 helpline calls connects mental health with the Sustainable Development Goals (ICTD 2022).

Related papers

Access to Care, Gender, and Stigma

Figure from "From patient voices to policy: data analytics reveals patterns in Ontario's Hospital feedback"
Figure from "From patient voices to policy: data analytics reveals patterns in Ontario's Hospital feedback"
Trends in patient feedback across Ontario hospitals between 2015 and 2022, from From patient voices to policy.

Access to care is shaped by stigma, gender norms, and local knowledge. A ten-month ethnography in three villages, The Ancient Elephant in the Room (ICTD 2024), examines barriers to reproductive care for rural Bangladeshi women, and "Parar-daktar Understands My Problems Better" (CSCW 2019) explores how people in 13 villages access healthcare. Computing and the Stigmatized (CHI 2024, Best Paper Award) studies trust, surveillance, and space with sex workers in Bangladesh, and "Deprived of support because we're born women" (CHI 2026) examines the struggles of women entrepreneurs.

At scale, From patient voices to policy (PLOS Digital Health 2026) uses natural language processing on over 120,000 patient reviews from 45 Ontario hospitals to surface patterns that can inform policy.

Related papers