What is preventing academic and industry collaborations in the development of health promotion technologies?

May 15, 2018

By Kate Wolin & Sherry Pagoto, Society of Behavioral Medicine

Eleven behavioral scientists working in academia, non-profits, and industry were asked to discuss tensions that stymie progress when it comes to behavioral scientists and industry working together. Tensions were defined in the request as a “specific type of challenge whereby there are often two competing demands that both have value.”

The following is a summary of the eight thematic areas that were generated in the discussion, followed by synergies and opportunities to overcome these tensions and advance the field of digital health.


Participants noted that academics and industry professionals hold conscious and unconscious biases about the “other side” regarding its respective values and goals, with each side judging their value system as superior and underestimating the flaws inherent in their respective value systems (see Table 1 for examples).




Industry cares about money, we care about people.            

Both groups are functioning under incentive systems (e.g., grant funding, publications or revenue goals) that could overshadow altruistic motivation.

Companies deceive patients

The motivators and incentives for both sides can be hidden from patients.

Industry doesn’t care about patient privacy

Both groups have strong moral and financial incentives to protect privacy and to create safe ways for information/data to flow when authorized.


Academics care too much about conflicts of interest.

Loyalties can cloud judgment or obstruct progress regardless of position. For example, the tobacco industry history of hiding data about product harms and doing so through partnerships with researchers gives scientists reason to be wary.

Academics don’t care about impact.

Ability to demonstrate that what they do delivers outcomes is integral to success for both groups.

Academics ignore scale and reach.

Both sides are seeking to identify solutions that are effective and serve the needs of the whole population spectrum.

Participants noted a prevailing bias on both sides that they are more data-driven than the other. This suggests both a values opportunity (data is important to everyone) and a tension around who is doing it and doing it well. Both sides should be concerned that simply using data is not sufficient to be data-driven, as careful consideration must be given to whether it is the right data and that data are being analyzed appropriately and interpreted correctly. This requires knowledge and skills that come from training and both groups should be aware of the ability to generate numbers without understanding their underlying assumptions and limitations.


Industry hides negative findings

Academics have long struggled to find venues for sharing null results and their incentive structures do not reward null studies, making this a challenge for everyone.

Industry is not concerned with proving results.

Many of the modern digital health companies are engaged in randomized clinical trials to demonstrate their program efficacy.


Academics take a kitchen sink approach to program development and don’t know what parts do/don’t work.

Optimization strategies are evolving in academia (e.g., MOST) and industry (agile), with both using data to identify efficacious strategies.

Academics are overly concerned with defining terms.

A clear definition of metrics, including ability to reliably and validly assess the domain, benefits everyone.

Certainty and action

Tension exists regarding the extent of certainty that is necessary to taking responsible action. Academics are reluctant to express certainty and tend to require a high level of certainty before acting, whereas business culture favors releasing innovations early and testing and iterating ’in the wild. Tensions around certainty also affect our goals and plans. Academics plan small steps because funders demand supporting evidence to go to the next step. In industry, bolder leaps are taken because funders don’t necessarily have such criteria but instead move on innovative ideas that have high potential to solve a major problem and generate profits.

What metrics matter?

Academia and industry focus on different metrics. Examples of important industry metrics offered including click rates, referrals, steps (physical activity) and conversions. Industry also has a need to measure cost in any undertaking to evaluate pricing if nothing else. By contrast, cost analyses are rarely primary in an academic analysis, only considered once efficacy is well-established. Academic research studies tend to focus on clinical outcomes that may take months or years to demonstrate meaningful change (e.g., weight loss, diabetes risk). Related, but separate, the attendees noted that product or program evaluations in industry rarely last more than a few months.

Pace of progress

The pace at which each group moves and decides when a program or product is ready for dissemination differs. Industry has a history of releasing ‘MVPs’ (minimally viable products) and iterating (again a reflection of the focus on an agile development process with frequent A/B testing). This results in many release cycles over a product lifecycle, as is the case with smartphone apps undergoing frequent bug fixes and updates. In contrast, academia conducts small pilots that are used to inform the definitive product. This is reflected in a “finished” product that is offered for others to adopt whole, as in the case of the Diabetes Prevention Program.

Lost in translation

Many of behavioral science’s most successful interventions were developed and tested before digital health became the force that it is today. This naturally creates tensions around what can be modified when digitizing the approach while retaining the original intervention and its effects. Perhaps the best-known example of this is the digital adaptation of the Diabetes Prevention Program. The original program relies on handouts, homework assignments, and face-to-face meetings with a trained behavioral or nutrition counselor. In this case, industry’s digital programs have needed to demonstrate to the FDA, via robust clinical trials, comparable outcomes. Tension exists beyond whether or not a program can be delivered digitally and remain efficacious and whether it can be changed in other ways, such as decreasing its duration, or modifying measurement tools.

Values again come into play, as industry may be willing to accept a slightly lower product efficacy if a digital program can be deployed at a fraction of the cost. Values may also be why industry uses science in developing a product but doesn’t include it in a brief product description. Academics rarely discuss a new program or intervention without referencing the foundational science.

Finally, scientific findings published behind paywalls, in lower profile journals or presented at academic conferences may not be accessible to industry. This may result in industry failing to include known academic research findings in product development or, worse, “re-inventing the wheel” by developing and testing approaches that have been well studied.

Who owns what?

The group acknowledged that “ownership” is a challenge both groups face, though they may differ in what they think is worth owning. Both sides are navigating the issues that surround data ownership and sharing, particularly as it relates to healthcare data, where the patient may ultimately be the true data owner.

What is behavioral science?

Industry and academia may have different conceptualizations of what behavioral science is. Industry’s introduction to behavioral science has been thanks to influencers like BJ Fogg and popular press books like Thaler’s Nudge, Gladwell’s Outliers, and Kahneman’s Thinking Fast and Slow. These sources while extremely valuable—particularly in moving the field beyond marketing science—represent small subdomains of behavioral science, most notably behavioral economics. Behavioral science represents a broad knowledge base and includes the fields of social psychology, clinical psychology, sociology, anthropology and cognitive science, to name a few. Widening the scope even further, related fields including epidemiology, public health, health promotion sciences, exercise science, and nutrition science, also contribute to our understanding of health behavior. Academics have not effectively communicated the richness and depth of behavioral and related sciences to non-academic audiences.

Synergies and opportunities

In spite of the many tensions, many areas of synergy, opportunities for collaboration, and ways to learn from each other emerged. Synergies included a shared desire to invest in what works and to change health outcomes. Opportunities for collaboration include data analytics, product testing, and improvement.  They require networks that allow the respective groups to know about each other and an open, non-defensive conversation about expectations and responsibilities.

The hope is that by identifying often unspoken tensions and assumptions, collaborative conversations and initiatives can move forward leveraging the best of what industry and academia bring to the table. The Society of Behavior Medicine has begun collaborations with Computing Community Consortium (CCC) and the Working Group on Interactive Systems in Healthcare (WISH). In launching a new relationship with the Personal Connected Health Alliance, SBM hopes to bring together behavioral scientists and industry as partners in developing new, evidenced based health solutions.