mHealth: prove your value and embrace the clinical trial

mHealth can make a lot of sense to a lot of people. The promise of mHealth has inspired people to develop thousands of health apps and more than $4 billion of investors’ money has poured into the sector.  But is all that effort and money justified?  Because to date, mHealth hasn't made much of an impact in the day to day business of healthcare. The problem is that healthcare providers have serious questions about validity, reliability, interoperability and reimbursement.  

Let's say you want to develop a mobile monitoring solution (sensors in wristwatch + mobile app) for heart patients that tracks heart rate and physical activity, and that lets patients report symptoms and communicate with their doctor. What a life saver! Patients will be tracked 24/7 with the result that doctors can intervene at the slightest hint something goes wrong. Patients will be reminded about their medication, improving adherence. And cardiology clinics can improve their productivity by monitoring large patient populations on a remote, semi-automated basis. Great!

But when you put your solution to the test you may find that patients don't like your solution (uncomfortable, ugly, not desirable, too complex, difficult to use, short battery life, etc) and hence quickly stop using the device. Moreover, your device could fail miserably as an accurate and reliable measure of heart rate (your device's readings are erratic compared to a 'gold standard' measure) and physical activity (a beer swigging patient registers several steps for every swig). And to top it off, doctors (and their fragile telemonitoring dashboards) are overwhelmed with calls from anxious patients who can't make sense of the scary readings they see on their phone.  Wisely, the regulator rejects your solution and payers decide that it isn't worth reimbursing. Obviously that's a sad story if it's a product that you developed, but at least it gives doctors, patients and payers some clarity and holdfast on what (not) to prescribe. But the reality today is that the vast, overwhelming majority of mHealth apps and tools have not been subjected to rigorous evaluation. 

If mHealth is to gain traction in healthcare, then it is imperative that we evaluate its performance. 'Performance' can and should be broadly defined. Questions for evaluation could include:

• Do patients accept the mHealth solution? Can it be seamlessly integrated into people's everyday lives? Do they actually use it?
• If the tool claims to measure or monitor a health-related parameter, does it do so accurately and reliably?
• Does it really "engage" patients and change behaviour?
• Does it advance integrated care processes, by improving communication and collaboration among healthcare professionals?
• Does it increase the productivity of healthcare providers and have economic benefits?
• And perhaps most crucially, does it improve the quality of care and improve health outcomes?

If you ignore some of these questions in your evaluation, then you may scupper the study - and create further resistance to mHealth.  For example, the Scripps Translational Science Institute recently conducted a randomized control trial (RCT) comparing chronic care using connected devices against standard disease management models. To the horror of mHealth evangelists, the study concluded that the devices had no discernible impact on healthcare utilization, health self-management and health outcomes. However, an analyst at Chilmark Research highlighted a number of flaws in the study that could explain the poor result. For example, he pointed out that the use of the devices was not “stitched seamlessly into the fabric of patients’ everyday lives - put differently, this is a workflow problem exacerbated by software.” Key issues in this regard were the provision of study phones (as opposed to letting patients use their own phones) and demanding that patients log into a third-party portal using that third-party phone. “A smartphone is not the same as YOUR smartphone. Our phones are an extension of ourselves, inextricably linked with our behaviour, moods, and decisions for hours and hours every day. The baseline for impact on 'real' patient behaviour is already off.”

The above case illustrates the importance of understanding the end-user's perspective and experience, to make sure that the solution is used properly in the first place and really does "engage" and "empower" patients. Ultimately, however, there is no avoiding the Randomised Controlled Trial, the gold standard in scientific method.  

Some argue that Randomised Controlled Trials are too expensive and time consuming to have much applicability in the fast moving world of technology. Indeed, by the time you publish the results of your trial, the technology tested may be obsolete and irrelevant. Nevertheless, in a healthcare context it is imperative that products are proven effective, otherwise they will not be prescribed and reimbursed. Some digital health companies are embracing this challenge and succeeding as a result. For example, companies such as AirStrip Technologies, WellDoc and Corventis (now Medtronic) have all received FDA approval for some of their products, which is facilitating their adoption in clinical practice.  

In preparation for the mHealth Hackathon in Brussels, 18-20 March 2016, we asked a panel of experts to brainstorm a number of challenges under the topic "mHealth Performance Metrics".  They came up with two challenges that address two urgent and pivotal tasks in this domain. First, if mHealth is to take off, then a lot more mHealth clinical trials need to be done. Hence, any services that make that task a little easier, faster and cheaper will be very welcome. Numerous cloud-based services exist (e.g. Medidata, Science37, Exco InTouch, ProofPilot) that help researchers organise and manage clinical trials, but there probably is potential to create new tools, functionality and methodology specifically for evaluating mHealth.  For example, Evidation Health, a company recently launched by GE Ventures and Stanford Health Care, is focused specifically on quantifying the clinical and economic outcomes of digital health solutions.

The second challenge tries to make life easier for physicians and patients who need to choose among the thousands of apps on the market. Here the idea is to create a curation system that makes it easier to find the right mHealth solution for a specific context or need. Several initiatives already exist (or existed) to tackle this challenge and some hospitals have created their own “app stores” but it is a challenge that is still far from solved.

If you’re up for the challenge (or have your own idea) then consider taking part in the mHealth Hackathon, there’s still time. All details at mhealth.be.