Now It’s Personal: How Precision Behavioral Tools Could Dramatically Improve Healthcare Outcomes

Contributors: Gil Kerbs, MA, WG’13 (Wharton/Lauder Institute, Chinese Track) and Yael Hallak
To learn more about Gil and Yael, click here.


70% of annual deaths and 75% of U.S. healthcare expenditures are due to chronic diseases, conditions mostly preventable by employing better nutrition, sleep, exercise habits, and stress management. A paradigm shift is required: how can we create better outcomes while reducing healthcare costs?

Preventive medicine emerges as our only chance against fierce enemies such as obesity, heart disease, and diabetes. Prevention has a unique trait: instead of a one-time intensive treatment, it is the addition of small decisions made on a day-to-day basis, while healthy, that determine long-term health.

The most influential of those decisions is what you eat. For example, the CDC estimates that merely reducing U.S. sodium intake to 2,300 mg a day could save 280,000 to 500,000 lives and nearly $100 billion over the next decade.

Personalized nutrition 
What exactly constitutes a 'healthy diet'? Is there a fixed list of foods that would 'solve' our problems? Sadly, that is not the case.

First, many clinical factors determine what is healthy and desirable for whom. For example, while most people are advised to eat more leafy green vegetables, those exact foods could cause pain for a Crohn's disease patient. Another good example is the ketogenic diet ("Keto"): while clinically proven to help epileptic patients, the high protein intake might not benefit someone with kidney disease.  

Second, nutrition is not only clinical but also psychological: the last time you had a beer was not because you clinically needed one. Nutrition provides us with some great pleasures, such as the nostalgia of eating your mom's chocolate cake or with useful tools to bond with others, such as having a cup of coffee while having a heart-to-heart conversation. Eating habits are also correlated with other factors: stress, mood, and sleep.

Thus, for personalized nutrition to work, it has to take into account not only clinical factors such as bloodwork, symptoms, etc., but also psychological factors such as personality, environment, motivations, and goals.

Take something as straightforward as a personalized meal plan, a recommended list of meals provided by a nutritionist. Assume our patient is a sales representative, regularly moving between her customers. If the nutritionist disregards behavioral factors, the proposed meal plan may include a 10 am yogurt snack. As our patient is on-the-go from early morning and cannot refrigerate her food, this meal is not feasible, and in real-time, the patient will opt for a snack at the nearest vending machine. Another example is avoiding favorite foods. While it might be optimal to avoid chocolate altogether while trying to lose weight, a good nutritionist may even encourage you to consume it moderately, as deprivation may cause you to give up on the whole process.

In this article, we will share a few behavioral insights gained while working on nutrition personalization.

Insights on behavioral personalization

Personalized does not mean unique in every dimension.
Many think personalization is like a fingerprint: a unique print for each person, causing concerns that these efforts cannot scale. We found it does not quite work like that.

Personalization happens when you consider multiple factors that characterize and influence behavior, and then split each factor into a few groups. For example, occupation may be divided into many sub-groups such as "on-the-go," "clerical," "manual labor," etc. While a person might not be special by being a salesperson when you look at the entire factor-set, she might be extremely unique: an Indian-American married mother of three, who works as a salesperson in a high-pressure environment, runs marathons, and views food as strictly functional. Insights relevant to that specific group could be seen as extremely personalized.  

The same goes for potential solutions. Behavioral solutions to clinical problems are not an endless universe. A few methods to increase water consumption might cover the needs of 95% of the population. The challenge then becomes which solution to assign to whom. The good news? Conducting this process correctly may increase the effectiveness of your solutions dramatically, the same as you would improve sales effectiveness by good segmentation and targeting.

Make data relevant
Most people do not understand nor react well to pure clinical data. Very few care about how much vitamin C they consume or the ferritin level in their blood. They do care, however, how that is relevant specifically to their health. To increase compliance, instead of communicating 'what,' start communicating 'why,' even if simplification is required. For example, an indicator of immune resilience with insights on how to improve it is preferred to a dry list of clinical factors that contribute to immune health.

Trial and error 
As any general practitioner (GP) knows, improving one's health can be a journey without a clear map of the road ahead. 'A person like you' might be likely to respond to a particular treatment, but upon starting, we find it is not working. This is partly because psychology/biology is complex, and we have yet to map all factors. It is also because of random effects: your boss might be edgy due to her father's illness, causing you increased stress. Moreover, any behavioral scientist will echo that even strong-willed people are likely to have small failures along the way. It is the ability to bounce back that creates long-term success. Trial and error play a significant role in overcoming those issues. Any personalized solution should acknowledge this inherent difficulty and react to it. 'Not working for you? That happens to many. It's not a problem! Let's try something else!'

Combining clinical and behavioral factors for health services personalization is the future. This is a big part of Newt, the personalized nutrition company we founded. We believe companies making clinical solutions not only accurate but also behaviorally adapted will eventually own healthcare. Would you like to join us in a movement that might turn the tide across the globe?

Contact Gil at: [email protected]

Contact Yael at: [email protected]