Nutrition For All: Our Solution for De-risking Disease in the Hispanic Community

A culturally humble and affordable digital platform to empower at-risk families, particularly ethnic minority and low-income families, to improve nutrition.

What's the problem?

The Hispanic community faces greater risks in chronic diseases and have higher rates of obesity. From a lifestyle perspective, this population lacks access to clinical care, healthy eating information and falls behind in health literacy. For some subsets of this population, a language barrier further exacerbates access to information and care. Other minority and low-income groups face similar challenges.

Why should we solve it?

Hispanics are the largest minority in the United States and have higher rates of type 2 diabetes (T2D) in both adults (80% higher than non-Hispanic whites [NHWs]) and children (fivefold higher than NHW). This conveys a huge cost to society. It was estimated by the American Diabetes Association (ADA) that the economic costs of diabetes in the United States in 2017 had increased by 26% from 2012 because of increased prevalence of diabetes and the increased cost per person with diabetes. (1) Care for people with diabetes accounted for 1 in 4 spent dollars in health in the United States representing $327 billion when combining direct medical costs and reduced productivity.

Higher incidence and prevalence is partly due to sociocultural factors, such as lower income and decreased access to education and health care, as well as a genetic susceptibility to obesity and higher insulin resistance. Lack of affordable, nutritious food is also a serious problem in the Latino community. According to Healthy People 2020, Hispanics are 3.5 times more likely to be obese than non-Hispanic Asians (44.9% versus 12.5%), who scored the best rate as a racial/ethnic group, and less likely to consume the recommended daily servings of fruits and vegetables. Health disparities were also observed among levels of physical activity.\

Addressing nutrition and health education via an accessible platform based on established nutrition data minimizes these problems and promotes accessible, affordable, personalized, and quality care.

What tools do we propose to develop?

This project outlines the development of a digital platform to empower low-income Hispanic families with easy access to diet and nutrition data to minimize health risks of diabetes, cardiovascular conditions, obesity, and cancer, among others. The platform would provide easy access to nutrition facts and incorporate gamification and cognitive behavioral change models to incentivize families to eat healthier. It would recommend culturally, seasonally, environmentally appropriate and economically accessible recipes, health tips, and healthy ingredient alternatives so they can enjoy foods of their preferred cultural origin while improving their health.

The platform would take user-entered data to assess health risk based on diet. Over time, it would provide assessments of increased or decreased risk based on their food choices as measured by our machine learning models. We propose using a human-centric approach to generate a positive user experience. To this end, we propose testing various models for logging food easily. For example, we propose exploring the development of a photo upload feature where users take pictures of their food and the application would provide real time feedback and assessment on health risk through machine learning. The application would then collect risk assessment data which would be available for NIH researchers and scientists through cloud computing. We also propose incorporating geospatial datasets such as SafeGraph. This application will feature recommendations where to shop and get ingredients and even feature price comparisons.

What tools did we use?

Python, Jupyter Notebook, AWS, Google Sites

What datasets did we use?

openweathermap, open data DC grocery store locations

What datasets would we like to use?

USDA Grocery Stores Dataset, Food Data Central, Kroger API Dataset

Team members

Isabel Otero (co-lead, nutritionist)

Tom Coyle (co-lead, system administrator)

Carol Gu (writer, programmer)

Rhonda Moore (NIH, dataset researcher)

Timothy Chirava (clinician, dataset researcher)

Rohan Narain (NIH, dataset researcher)

Acknowledgements

Allissa Dillman (NIH, codeathon organizer)

Ricardo Sandoval (computer scientist, programming mentor)

Edwin Lock (computer scientist, programming mentor)

Duncan McElfresh (health applications expert, mentor)

Lynn Kirabo (human computer interaction, mentor)