February 16, 2018 | Written by: Elizabeth Koumpan
Share this post:
People may be surprised to learn that in 2010, humanity passed an important milestone. Obesity became a bigger public-health problem than hunger. This crisis is not just a pressing health concern. Obesity has caused a severe rise in heart and lung disease, diabetes, lifestyle-related cancers, and other non-communicable diseases.
There is lack of knowledge and understanding of nutrition; as a result, unhealthy food environments exist in schools, public institutions and elsewhere. As a society, we face significant health problems, with diets being a key contributing factor. To address these issues, we launched an Academy of Technology (AoT) Initiative: Precision Nutritional Cognitive Insights Based on Different Data Sources.
|IBM’s Academy of Technology ‘Initiatives’ are cross-business line projects that provide actionable insights to advance understanding of key technical areas and trends helping engage our clients in technical pursuits of mutual value.
We can create a system that determines the positive effects of foods by tracking the health outcomes of patients who have eaten certain foods, and providing insights at a population level. To accomplish this, however, we are challenged by lack of integration between clinical, nutritional, behavioral/physical activity, and genomic data. At IBM, we have full set of capabilities to address and meet these challenges.
The standard American diet is jammed with pro-inflammatory foods, but proper nutrients can reduce inflammation. DNA can be damaged by outside forces, but some foods enhance our body’s ability to self-repair. Over 200 genes have been identified as possible contributors to human obesity. At the same time, genetic variation may affect food likes and dislikes and nutrition.
The food we eat has the power to influence six major neurotransmitters in the brain, and a low-quality diet can change brain chemistry and its function. Understanding simple biochemistry will help us become healthier and adjust our diets according to our personal needs. For example, structural information about individual regulatory proteins and biochemical and metabolical pathways can help us investigate a number of human diseases, such as Alzheimer, Parkinson, and Type 2 diabetes.
Analyzing such data will help us understand the benefits of specific food and related chemistry components. We can provide more accurate dietary nutrition recommendations if we can apply such understandings to account for specific medical disorders, or specific genetic conditions, taking into consideration environmental characteristics, such as location, weather, cultural, historical and psychological factors.
The challenge of this approach requires reducing the complexity of highly diverse nutrition science and nutrition information. Today we use only limited data sources to gather information, and collected personal data is not always aggregated and comparable to similar use cases across populations.
Using cognitive computing, data science, machine learning and other techniques to collect, aggregate and analyze data from multiple sources, we can reduce nutrition data complexity to better deliver understandable and personalized recommendations for users.
We can build nutrition models that estimate human requirements and nutrients derived from menus created for specific diseases, and used by specific populations based on geography (geo location, population density), demographic (age, race), physiological (personality, social class), education (awareness regarding the types of food and beverages that help maintain a balanced diet and proper health), and medical pre-condition (genetics, family history, personal habits – smoking, drinking, and others). We can then train the models on newly-collected data. We can also use precision food consumption and nutritional habits to predict disease onset likelihood and suggest ways to avoid it by generating recipes using ingredients that have been shown to minimize the chance of developing certain conditions.
Across the nutrition domain, cognitive computing and artificial intelligence will help us:
- Understand the correlation between patient medical data and diets over time
- Analyze the possibility that specific nutrient and dietary components associated with specific diets have a positive or negative impact on specific diseases
- Evaluate insights of the effects of specific foods on disease
- Understand how nutrition may reduce the efficacy of some medicines
Cognitive systems will learn and build knowledge from various sources to understand the relationships between food, chemical components, eating habits, diet, environment, medical history, genomes, and specific diseases.
As Thomas Edison said, “The doctor of the future will no longer treat the human frame with drugs, but rather will cure and prevent disease with nutrition.” We add to this – with cognitive systems and artificial intelligence, that helps us to facilitate how optimal diets can be achieved – while understanding about good food, specific practices, awareness, motivation, and supporting better disease prevention and management.
To learn more about the finding of this Academy Initiative download the complete Academy report.
To learn more about what IBM has developed in this space, please explore the following links:
We encourage our clients to visit our website and to contact us at email@example.com to inquire further about AoT Initiatives and partnering with our client-value teams.
These are the opinions of the author Elizabeth Koumpan and while a distinguished member of our Academy and IBM, all thoughts expressed are solely her own.