Most people understand that a significant predictor of diabetes outcomes (HbA1c) is good treatment decisions by the treating MD. But is there something more to it? A model developed by MindField Solutions in collaboration with The Cleveland Clinic revealed a breathtaking finding. The research evidenced that patient cognition was more predictive of HbA1c than whether or not the patient was on insulin (p<.05). Through this research, a reliable and valid patient diagnostic tool was developed that provided a cognitive profile of the patient, showing the patient’s likelihood to achieve desired HbA1c based on their cognitive approach to health and disease management. With this model, select patients for further intervention could be identified and provided with predictably meaningful solutions to moderate their approach to health behavior, and consequently, health outcomes.

Sadly, most models of healthcare change are not founded on neuroscience predictive models. For example, the British Medical Journal discussed the failure of a teenage smoking cessation campaign in the UK. The program was based on the popular “Stages of Change” theory of health behavior by Prochaska. The author concludes “Its failure even in this group reinforces the evidence that the acquisition and shedding of a smoking habit in the teenage years is essentially chaotic. Unlike adult quitting, it does not follow any readily definable stages.” But why did the programmers assume the “Stages of Change” theory would be impactful here? There is no data to demonstrate a predictive relationship between the stages and the outcome behavior. The stages of change model is, afterall, a segmentation methodology, rendering it incapable of making predictions. Getting the neuroscience wrong here created significant financial waste, it created significant burden in implementing the operational changes to accommodate the failing program, and no doubt it demotivated teachers and students alike.

The correct neuroscience path may have included a programmatic effort based on Social Learning Theory, for example, and ore importantly, developing a reliable and valid predictive model within the target group prior to implementing a costly operational program.

The same can be seen in the more recent “Just say ‘No’” campaign in US schools. Again, this program was built on faulty assumptions about what drives teenage drug use. No predictive model was used to support the development of the campaign. Many campaign designers are unaware of the differences between segmentation models and predictive models and consequently, misappropriate significant investment behind segmentation models when behavior change requires a predictive model, and usually, one derived from neuroscience.