What Digital Health Tells Us About Disease

Christine Ngaruiya, MD, uses Natural Language Processing to uncover gender disparities in noncommunicable diseases, while also leading initiatives at the intersection of health and climate change, fostering targeted interventions and policy changes worldwide.

For more than a decade, Ngaruiya has worked toward increasing research and resources for noncommunicable diseases (NCDs), with a focus on racial equity and addressing disparities in health outcomes. Her work comprises epidemiologic assessments and the use of implementation science to develop targeted, context-sensitive interventions in Africa. Among this work is the application of digital health tools for public health interventions as well as exploring the intersection between climate change and the increasing global NCD burden.

Ngaruiya notes there is a substantial disparity in funding between communicable and NCDs, as well as gender inequality and a need for gender-specific targets. Her work helps to improve specificity in creating successful interventions.

Ngaruiya led a study that employed Natural Language Processing (NLP) on 5,358 discharge summaries of patients with acute myocardial infarction (AMI) in Pakistan, partnering with leadership at the Aga Khan University Hospital-Pakistan to investigate gender differences in symptoms and management of ischemic heart disease. The NLP model demonstrated high specificity and sensitivity, and highlighted potential gender disparities in how NCDs are identified and measured. Women with AMI are more likely to present with shortness of breath or gastrointestinal complaints while men are more likely to present with conventional symptoms such as chest pain.

According to Ngaruiya, NLP is underutilized in public health interventions, where accuracy, speed, cost-effectiveness, and scope could identify gender disparities as well as influence policy changes and improve healthcare outcomes in diverse global settings. NLP allows for the extraction of valuable information from large datasets, comprising tens or hundreds of thousands of data points. The manual extraction of specific information from vast datasets is tedious and prone to imperfections. NLP automates this process and teaches itself to be more effective than humans in identifying and extracting relevant data even in the presence of variations, spelling errors, or missing information.

Ngaruiya is also tackling climate change initiatives, collaborating with partners like the National Cancer Institute to explore intersections between health and environmental factors. A recent project involves assessing the impact of climate change on various health aspects and formulating policies in collaboration with the Kenyan government and regional authorities.

Ngaruiya came to Stanford in 2023 to serve as the population and global health research director for the Department of Emergency Medicine, drawn by the university’s diverse resources, multiple centers, and the broader digital community in the Bay Area.


Updates Spring 2024