Data and Outcomes

The 7 Data Principles that Drive Lark

The 7 Data Principles that Drive Lark
Lark Team

We are passionate about providing scalable virtual care. We pioneered the use of AI for better health. Combining the power of data, behavior change science, and smart devices, Lark’s digital platform provides scalable, personalized coaching 24/7 to help people manage or prevent chronic disease.

Every member interaction and coaching session with the Lark AI coach is a result of years of data collection, thorough analysis, and data-driven insights. Lark dedicated six years of R&D before our 2018 market launch collecting one of the largest user data sets in the industry. This data, plus member-specific data, is analyzed to determine the best behavioral intervention for each user at the time of intervention. 

Lark develops member profiles based on individual level data collected via connected devices, surveys, and in-app conversations to learn about each individual’s unique needs and preferences. We combine these insights with over one billion data points from other users to deliver a personalized coaching experience, identify and enhance clinical outcomes, and drive member engagement using the following seven data principles:

Clustering: Clustering is a powerful machine learning technique for discovering subgroups within heterogeneous data and allows Lark to identify users with similar behavior patterns to drive greater engagement and health outcomes as they progress through the program.

Natural Language Processing: NLP derives meaning from human language. It is used extensively in Lark’s proprietary food logger to enable painless, intuitive meal logging.

Multivariate Analysis: MA statistical techniques provide Lark with a clearer view of behavior variables that are highly correlated, enabling the most useful conversations to be presented to the user at the best time.

Trend Predictions: By extracting underlying data patterns, Lark eliminates the noise from biometrics readings.

Big Data: Lark’s population-level data set includes approximately one billion data points, the largest in the industry.

Active Learning: AL is a subset of ML that focuses on development of learning algorithms and user interactions that enable mapping of data with desired outputs.

Outlier Detection: This data mining step identifies anomalous device reading data which is then evaluated to determine whether it is truly an outlier to be removed from the data set, or represents a potential new pattern.

The Value of Data and AI

Lark’s award-winning AI platform delivers personalized digital health coaching that uses deep data knowledge to help users better manage chronic conditions and improve their health.  We are able to provide instantaneous, unlimited coaching that helps members take small, meaningful steps toward lasting behavior change all based on science.

If you’re interested in learning more about how Lark leverages data to provide personalized AI support and achieve similar or better outcomes than other programs, contact us at modernizecare@lark.com.