To Device or Not to Device?
At Moving Analytics, we are often asked questions about the use of connected devices like activity trackers or weight scales. The promise of connected devices is clear - being able to seamlessly and objectively collect patient behavioral and clinical data points and analyzing them could lead to better insights and treatment plans. In an effort to improve our patient experiences, we have spent the better part of a year testing well over 50 devices such as activity trackers, weight scales,BP cuffs etc, not to mention several service aggregators (in an effort to remain neutral and keep this article short, I won't be naming everything we tested).
A Case for Usability
The questions we often get asked center around two themes -
1. Can you really trust patients not to lie about their data? How do you know that they are being truthful?
2. How do you get objective data on which to base clinical decisions?
In our experience 1.is less of a problem - in our case, a patient might literally have had their chest opened and usually has very high motivation to not let that happen again.They understand that by providing accurate information, their care manager can help make sure that doesn't happen. 2. is not usually a problem if the patient actually uses the device. A more likely scenario is that patients do not remember to log their information, they simply forget. The device is meant to solve the ease of use of logging objective data, not the problem of objective data.
In reality, the problems we have found to be more prevalent are from a user experience and logistics perspective:
1. How does one reduce the cognitive burden of setting up, reducing and maintaining the device?
2. How does one create habits for the patient to use and derive value from the device?
3. How does one process data from a device(or family of devices) to provide succinct, actionable insights that allow a clinician to do their job?
The Device Wish List
As you can see, the problems are not with patients or clinicians, they stem more from usability. So what have learned on that front? In our view, a connected device should have the following characteristics -
1. Should be relevant to the problem at hand, if the goal is to reduce weight, then use a weight scale, if the goal is to get patients to walk more, use a pedometer. But one should not overburden the patient with unnecessary devices. That increases the cognitive burden.
2. Should have validated measurements that are within the limits needed to make clinical judgement. We have had experiences with heart rate-based activity trackers that are sometimes 20 beats per minute off from the true value. In such scenarios, it might be easier to simply teach the patient to measure their pulse manually and enter the information.
3. Should be passive and non-intrusive, with very little setup time and minimal recharge burden. Anytime you add steps in the setup, connection, data pulling etc of devices, that's one less reason for the patient to use the device. In particular, for wrist-worn devices such as activity trackers, issues like wearability, allergies come in to play. An ideal experience for us is to simply turn on the device, have a one-step linking to a server and then start receiving data.
4. Should have easy, anywhere-anytime connectivity (cellular preferred) that is not dependent on smartphones.Connectivity is key to ensuring that patients do not have the cognitive burden of debugging and transferring data from these devices. I cannot stress enough how devices with cellular connectivity have a much superior experience compared to those that implement Bluetooth connectivity. It is my hope that with the advent of 5G, this modality of data transfer becomes more common.
5. Should have easy to use open, cross-platform APIs (with triggers for data availability) that allow developers to pull and utilize information for their needs. Not being able to work across platforms (Android, iOS, web) grossly limits their usability as well. Being able to read and manipulate data easily is key in getting these devices adopted quicker since it will help generate insights that will bring clinicians onboard.
6. Should be affordable and have a justifiable return on investment. Most patients can't afford a $500 connected device. IMHO, the true payers for these devices have to be insurance companies, the challenge will lie in showing why spending $500 will result in at least$500 in savings because the user adopts healthy behaviors that can reduce downstream costs.
Equally important is to pair the device itself with a solid on-boarding and training program for the end user that is virtually "novice-proof."
It was surprising to us the kind of issues we ran into, for example some patients are not permitted by their doctors to lift more than 10 pounds of weight. Now if a device-kit you provide them weighs more, how do you get it home? Another example - if a health system buys 500 BP cuffs or activity trackers, where exactly will their store them? Who's in-charge of putting in the batteries or charging them?
Lastly, we have found that sometimes, it's better to not provide a device at all. Some patients are perfectly willing and able to manually enter their information and have high compliance. For such patients, manual entry is simply replacing a diary entry.Thus a one-size-fits-all approach is not recommended. Forcing someone to use a device when they don't want to and are already compliant is creating a problem where none exists. The questions to ask here include - is the patient adopting healthy behaviors? If they are not, is it because they are having a hard time tracking their behaviors? Are they willing to then use these devices to help ease their burden of tracking. If yes, then go ahead. It is important to start with the intervention first - why are you providing the device in the first place and is it helping you?
As providers or payers think about incorporating connected devices into their treatment modalities, we strongly urge them to start with the user experience first and see how devices can match or execute against that user experience.Think about costs, budgeting, logistics and whether it makes sense to do it internally or work with external companies that can manage those operations for you. Think not about the device itself but about what behavioral interventions you want to achieve in your population and how you see devices helping you achieve that goal and make your day easier. Focus on the result and use devices as a means to a goal. And remember - less is more.