Predicting Biological Age for a HealthTech App

Background

Humanity, a health tech startup, is dedicated to using AI for aging to help individuals extend their healthspan. The company’s innovative app aims to leverage biological age prediction to assist users in slowing the aging process and living healthier for longer. The challenge is complex, as biological age varies based on lifestyle factors, and the effects of health interventions are not universally the same. To address this, Humanity sought to collaborate with Omdena to create a system that uses AI for aging to provide personalized recommendations for users.

Since launching globally in late 2021, Humanity has seen rapid growth, with 75,000 sign-ups and 50,000 active users. The company’s ambition is to continue scaling and to build a predictive model that can suggest the most impactful actions for users to slow their aging, ultimately improving long-term health outcomes.

Objective

The primary objective of this collaboration was to develop robust biological age prediction models that could analyze users’ lifestyle actions and predict their biological age. The specific goals were:

1. To harness AI for aging by predicting biological age based on activity, health behaviors, and biomarkers.
2. To develop personalized strategies that help users slow their biological aging process.
3. To provide a data-driven system that identifies the most effective actions, combinations, and timing for each user based on their unique attributes.

Approach

Omdena took a comprehensive approach to building predictive aging models:

• Data Collection & Preprocessing: Omdena worked with Humanity to gather high-throughput lifestyle data, such as physical activity, sleep patterns, diet, and meditation habits, along with key biomarkers. This data is crucial for predicting biological age and understanding how various actions influence aging processes.
• Unsupervised Learning Techniques: The team used unsupervised learning methods to compress and analyze this data, isolating the most significant factors affecting aging. This allowed for a more focused approach to predicting biological age and identifying key interventions.
• Supervised Learning & Model Training: Omdena applied supervised learning techniques to develop algorithms that could predict biological age based on user data. The model was designed to assess the impact of lifestyle changes on biological age and to create personalized recommendations.
• Cross-User Matching: To scale the system, Omdena’s team built a mechanism for matching users with similar characteristics, enabling the model to apply insights from one user to others with comparable profiles. This personalization is key in making biological age prediction relevant to each individual.
• Testing & Validation: The final models were rigorously tested using independent datasets to ensure that the predictions were accurate and reliable. The validation process confirmed that the models could effectively forecast the biological age outcomes of specific lifestyle interventions.

Results and Impact

The project’s outcomes were both impressive and impactful:

1. Increased User Engagement: Since the global launch of Humanity’s app, over 75,000 people have signed up, with 50,000 users actively engaged. The app successfully motivated users to increase their health actions by 15% in just 40 days, resulting in an average reduction of 0.49 years in biological age per active user. These results were validated against peer-reviewed biological age prediction models.
2. Predictive Aging Models: The predictive models, built using AI for aging, allowed Humanity’s app to provide personalized recommendations based on each user’s biological age and lifestyle actions. The models assessed which combinations of lifestyle changes were most effective, both individually and in combination, based on the user’s unique health profile.
3. Personalized Action Plans: Through the AI-powered system, the app provided tailored insights into which actions—such as changes in diet, exercise, and sleep—would help slow users’ aging processes. This personalization helped users make better decisions about their health interventions, further reducing their biological age over time.

Future Implications

The results of this collaboration have significant long-term implications for health tech, aging research, and healthcare practices:

1. Scalable Health Management: As Humanity’s user base grows, the biological age prediction models will become even more accurate, helping individuals at a larger scale slow their aging process through tailored recommendations. This will allow millions of users to take control of their health with data-driven insights.
2. Personalized Medicine: The project demonstrates the potential for AI for aging to influence personalized medicine. By tailoring interventions to each person’s unique biology and lifestyle, healthcare could become more proactive, reducing the incidence of age-related diseases and improving quality of life.
3. Advancing Aging Research: The collaboration also pushes forward the field of aging research by using real-world data to understand the effects of different lifestyle interventions on biological age. This could lead to further breakthroughs in how we approach aging.
4. Long-Term Health Impact: With continued development, these biological age prediction models have the potential to significantly reduce the burden of age-related diseases and extend the healthy years of individuals worldwide, improving overall public health outcomes.

In conclusion, Omdena’s work with Humanity has paved the way for a transformative approach to aging. By combining AI for aging with personalized biological age prediction, this collaboration is setting new standards in health tech, with the potential to revolutionize how we manage aging and health in the future.