So you want to develop a health app?

March 5, 2018

As Director of the UConn Center for mHealth and Social Media, I am approached a lot by academics interested in developing a health app but not sure where to begin. Here I describe 10 considerations before diving into that health app.

  1. Does it already exist? A helpful first step is to perform an environmental scan of apps on the topic of interest to find out what is already out there. If it is possible to build on existing work, many resources will be saved and the science can advance faster. An environmental scan can be done via a systematic search in app stores and the scientific literature. Sometimes apps are given names that are very different than the technical names we use to describe behavioral interventions, which can make finding the right search terms challenging. For example, an app for cognitive behavioral therapy for insomnia might be called Snooze instead of CBTi, since the former certainly has a better ring to it. To insure your search is thorough, reach out to 3-4 experts on the topic to turn up apps that may not yet be in the app stores or in a published study.
  2. Who is on your team? App development is not a solo endeavor. The right team should reflect the necessary expertise which may include clinicians, computer scientists, engineers, user experience experts, and mathematicians. The team should be convened at the earliest conceptual stages of the project. Half way through a grant application is way too late. All team members should be intellectually engaged in all phases of the project-- from concept to final product. One mistake I sometimes see behavioral scientists and clinicians make is treating computer scientists or engineers as technicians (“hey can you build this app for me?”) or computer scientists or engineers treating behavioral scientists or clinicians as the final stamp of approval (“hey we developed this great app can you give it your stamp of approval?”). Neither approach is creative nor collaborative.
  3. Start with a problem, not a solution. Once you convene your team, present the problem you are trying to solve, not the solution. Often I see folks come up with an idea for an app and then convene a team to develop and test it. This misses the opportunity of leveraging your team’s input on the best ways to use technology to solve the problem. For example, a problem might be that many of your patient’s won’t use a commercial weight loss calorie tracking app for more than a couple of weeks. Present this to the group, allowing everyone to ask questions and brainstorm ideas for how a new app could solve this problem.
  4. Are scope and resources aligned? Before diving in, do some homework on the steps of the app development process and how much each costs in time and resources. Ask several investigators who have done this work about the time and costs required. Then ask yourself, what do I have the time and resources to accomplish? If for example, you have a one-year $10K seed grant, how much can you realistically accomplish? I frequently see investigators with plans that exceed their resources and then they burn through their funds before they have any meaningful progress, or they skip essential developmental steps to save money.
  5. Do patients/consumers want this app? Your friends, family, and coworkers might love your idea for a health app, but will the users you are designing it for? If so, which ones? All of them or just ones with certain characteristics? Engage the target population from the very beginning of the process. The key is to set the stage to make them comfortable giving you candid feedback. This may mean having someone other than yourself moderating focus groups and usability sessions. People will detect your excitement and tell you what you want to hear. Tell them you are looking for constructive criticism, you value their honesty, and will take all feedback to heart to make this the best product it can be.
  6. What do relevant healthcare professionals think? Regardless of whether your app targets healthcare professionals, getting their perspective is essential especially if you do not deliver patient care. For example, would docs feel comfortable recommending the app to their patients? Do they think patients will use it? What do they see as the barriers to their patients using the app?
  7. What if your app works? Will you commercialize? Make it freely available? How will it reach people?Plans for implementation and dissemination are just as important for apps as they are for any other behavioral interventions.
  8. Consider studying a commercial app or device instead. The commercial marketplace contains so many products that have potential but haven’t been subject to empirical testing. Instead of inventing a new app, design a study to test something that already exists. If you can’t find a commercial app that is comprehensive enough in terms of behavior change strategies, try building it into a more comprehensive program.
  9. Keep it open source. This is science after all. Make other investigators aware of the app and invite them to build upon it and apply it in new ways. This keeps the science nimble and efficient.
  10. Don’t forget Jony. Jony Ive is the genius behind the slick visual and tactile appeal of the iPhone, an aspect that is widely lauded as the reason iPhone has the competitive edge. A problem in academia-produced apps is finding the money for aesthetics since it can be costly and grant reviewers may not view it as essential to the science. Engaging user interface (UI) and user experience (UX) experts on the team will insure that users find the app appealing and actually enjoy using it.