Duke Medicine is one of the organizations that recently launched a ResearchKit app in the second wave of iPhone-enabled clinical trials. At the HIMSS Connected Health Conference, Dr. Ricky Bloomfield, Director of Mobile Technology Strategy and Assistant Professor in Internal Medicine and Pediatrics at Duke University, gave some advice to other organizations that might be considering a ResearchKit study of their own.
He said that Duke's newest study, Autism and Beyond, had 180 people sign up in the first weekend, more than twice the signups that a comparable non-connected study was able to achieve in eight months.
"The initial excitement [around ResearchKit] has died down, and our engagement has been fantastic," he said. "Almost every user is doing three video surveys. If they weren’t doing this for the right reasons we wouldn’t expect them to devote that time to it."
Bloomfield said that not every study is a good fit for ResearchKit in its current implementation. When researchers are thinking about whether or not a study might be good for ResearchKit, there are a couple of things to consider. One, does the study make use of some aspect of the technology on an iPhone or Apple Watch -- either the touch screen, the built-in sensors or trackers, or, as in the case of Autism and Beyond, the phone's camera? Two, does the study really need the scale that ResearchKit offers?
"There are certain groups that might have a certain chronic disease or a rare condition, and it can be very difficult to do studies on that population because they’re spread out all over the world and there aren’t many of them," he said. "If they can download an app wherever they’re at, it’s much, much easier to try to scale."
Once researchers decide to do a ResearchKit study, they have to decide whether to build their own back-end infrastructure for the data or work with a third party vendor. Most ResearchKit apps in operation now are working with a third party such as Sage Bionetworks, but Duke chose to go in-house for its study.
"Because of the nature of the data in our study – we’re actually creating videos of children – we felt that was very sensitive," he said. "In order to keep that data secure we wanted to store it locally at Duke. So we chose not to take the path of using a third party provider for our infrastructure."
He said that building their own infrastructure was expensive and time-consuming, and recognized that a lot of organizations would choose to outsource. He cautioned those that did to make sure they vetted that third party carefully for HIPAA compliance and data security.
Another area Bloomfield cited as a challenge for ResearchKit is localization. In order to really capitalize on the potential for scale, researchers might want to make their app available in other countries (Autism and Beyond will soon be released in China, for instance). But not only does this require hiring a competent translator ("Google Translate is not ideal for medical questions," he cautioned), it also requires getting the study reviewed by an additional institutional review board for each localization.
Correction: A previous version of this article said the app was already available in China.
Interestingly, the aspect of ResearchKit that Bloomfield was most excited about was informed consent, because it diverges so much from what he sees as a deeply flawed status quo.
"The current state is a piece of paper with thousands of words on it, it’s very hard to understand, and you hand it to someone and ask them to read it and to sign it and you might ask if they have any questions. I don’t think we intend that to be coercive, but you can imagine a study coordinator standing over you, waiting for you-- it can be intimidating, it can be hard to understand, and many patients might be embarrassed to ask about it. ... The ResearchKit informed consent process I think helps to solve a number of those problems because this process can be done at the user’s home, at their own pace, and most importantly, it takes this very convoluted legalese process and simplifies it to have several different tiers of language."
In the ResearchKit app, patients get a broad summary of the document and then can click down levels to explain it in a little more detail including, all the way down, the thousand-word document. But they can consent to the study based on much easier-to-understand criteria.
"In my mind, the single biggest innovation to ResearchKit is being able to consent someone in this way," he said.