Assessing the Effectiveness of an N-of-1 Platform Using Study of Cognitive Enhancers

The growing burden of chronic disease in the U.S. and the economics of accountable care are driving a shift toward proactive approaches to disease prevention and health maintenance. At the same time, precision medicine studies continue to reveal substantial heterogeneity in the manifestations of even the most common chronic diseases. The bulk of morbidity and mortality in the U.S. arises from conditions with a significant lifestyle component (e.g. type II diabetes), and responsibility for monitoring and maintaining health largely falls on individuals.

Recent advances in molecular biology, sensors, and digital health technology underlie rapidly growing market availability of products and devices for measuring and monitoring individual health. A vast array of wearable devices, smart home monitors, and health tracking apps provide an unprecedented view of individuals "in the wild" and provide customers with health information once accessible only to researchers. The growing digital health market is generating unprecedented volumes of information to support decision making around individual health, and current trends suggest the demand for personalized health information, tools, and services will continue to grow in the next decade.

What is missing from this technological and scientific growth is a reliable, individualized way to translate data into action. If society wants to prevent diabetes, heart disease, and other chronic illnesses that kill millions of Americans each year, individuals must be empowered to address precursor conditions like obesity, hypertension, and depression. Dialogue around consumer health often fails to address the profound disconnect between measurements and outcomes/goals; e.g. monitoring one's weight is not the same as losing weight, and counting steps is not the same as lowering blood pressure. Data are only useful if they can help individuals identify interventions that work for them. The combination of diet, exercise, drugs/supplements, activities, and lifestyle changes that targets an individual's particular set of health problems is unique to him or her, and it is dependent on a complex web of factors including genetics, environment, and personal lifestyle. If individuals are to benefit from data, they must be able to relate changes in their personal data to targeted adjustments in actions and outcomes. This effectively necessitates conducting a robust trial at the level of the individual to determine the most promising recipe of personal lifestyle adjustments to effect change.

To address these challenges, the researchers have developed a unified statistical framework for producing consistent, interpretable study results from diverse n-of-1 study designs. The analysis framework is the backbone of the initial software platform, which includes modules for study design, e-consent, data ingestion, data analysis, and visualization of results.

To test this platform, the researchers plan to deploy a prototype study that allows individuals to test the cognitive effects of two commonly consumed substances: caffeine and caffeine in combination with a safe, prevalent compound, L-theanine. Each enrolled individual will participate in his/her own n-of-1 trial. After a baseline period where neither treatment is taken, participants will alternate between the two treatments ("caffeine alone" and "caffeine + L-theanine") according to a predefined schedule. Participants will complete a daily cognitive assessment composed of 3 validated cognitive tests administered via the N1 app. The platform will analyze the cognitive assessment data and determine whether there is a statistically meaningful treatment effect for either treatment compared to baseline for any of the 3 cognitive tests for each individual that completes the study.

It is important to state explicitly that the research objectives for this protocol are not related to the efficacy of L-theanine and caffeine. This specific study is designed to allow the researchers to efficiently recruit and enroll subjects so that the underlying statistical methods and software platform for executing n-of-1 trials may be evaluated.