The mHealth Summit is one of the premier mobile health industry conferences which does a great job at bringing together key stakeholders from government, industry (large and small businesses), and academia to address the state of mHealth and how to utilize this field to effectively address some of our nation’s greatest health challenges.
The Summit includes over 4,500 mHealth leaders from 50 countries. This year the Summit seeks to highlight emerging best practices for mHealth implementation which can guide the creation of better health interventions and sustainable mHealth business models. The Summit will also provide a unique focus on mHealth research – a key concern of iMedicalApps – which addresses how to develop and commercialize evidence based mHealth solutions for certain diseases.
This article focuses on a session presented by the National Institutes of Health entitled “State of the Science in Research on Mobile Health Technologies.” The session was attended by about 300 people. The speakers included Joseph Cafazzo (University Health Network) and Bonnie Spring (Northwestern University Feinberg School of Medicine). The panel was moderated by Roderic Pettigrew from the National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health.
Are Randomized Controlled Trials necessary for mHealth?
One key issue discussed by the speakers which resonated with audience members asking questions was the degree of evidence needed for mHealth. This is a fundamental concern for research on mHealth technology.
Cafazzo addressed this issue head on after indicating the his center (Centre for Global eHealth Innovation) uses a user-centered approach to developing apps and engaging in research to understand the impact of apps. He noted that randomized controlled trials present a number of challenges for mHealth innovators.
First, they are expensive which can prohibit smaller mHealth innovators from performing them. Cafazzo also stated that mobile apps are not drugs and should not be treated as such with regard to the research performed on them. Second, they take years to complete and mHealth technologies evolve so quickly that this research methodology cannot keep up with the technology it is being used to study. Third, some trials are performed without earlier stages of innovation being done well.
According to Cafazzo, the earlier stages of innovation should include a user centered design with feedback mechanisms throughout the process. The concept for the mobile app should arise from a need which users have. The design of the app should be assessed by users. Once the app is created, user testing should be done to work out all the problems with the app. After this testing, a pilot test should be done to determine the potential impact of the app. Cafazzo gave examples of successful pilots which provided useful information regarding the impact of apps. Randomized controlled trials performed after these pilots were completed simply reiterated what was already known by the pilots.
However, the trials were completed years after the pilot. Cafazzo stated that this inefficiency could be eliminated by researchers becoming more flexible with regard to the level of evidence required for mobile app effectiveness.
Spring presented a contrasting, albeit not opposing, viewpoint regarding this issue. Using the example of hormone replacement therapy research as a cautionary tale, Spring indicated that there have been instances in history where researchers thought randomized controlled trials were not warranted and the results were patients who died as a result of lack of quality research data. Indeed, the notion of harm is an important aspect of this discussion.
However, a more important discussion which the researchers did not clearly address is where does the notion that a randomized controlled trial is necessary for mobile apps come from? There are multiple devices that are marketed (with FDA approval) which have no trials of any sort to support their efficacy. Why? Because they are unlikely to harm individuals. FDA, as the nation’s overseer of safety and effectiveness of drugs and devices, determines (with guidance from Congress) what medical products are safe for the public and what is not.
In the mobile health context, the FDA’s regulations are extremely limited to date and do not cover most mobile apps. So, perhaps researchers in the mHealth sector should first ask what types of apps are regulated by FDA. If an app is not regulated by the FDA, then the incentive to develop evidence for an app’s effectiveness and perhaps safety is much lower. This is why so few app innovators have any effectiveness data – just as most software companies do not engage in research on effectiveness.
There are other incentives for producing data beyond the FDA. The main incentive is competition. If competitors have research data which impacts the sales of their mobile apps, then research may become a standard for a particular type of app. Even in this context, the research threshold may be low.
As Cafazzo indicated, a pilot study may be adequate to support efficacy. In fact, spending any more time on research beyond a pilot might detrimental to the success of the app since the amount of time that the app is viewed as a current solution may be 18 months to two years.