In what boils down to crowd-sourcing medical information, Apple’s ResearchKit promises to turn the iPhone into a powerful tool for medical research. But will it live up to that promise?

Apple is putting the power of clinical trials in our pockets with ResearchKit, the open-source software framework designed for medical and health research. It will help doctors and scientists gather data more frequently and more accurately from clinical research participants using iPhone apps, enabling faster insights at lower cost.

ResearchKit leverages the sensors and other capabilities of the iPhone to track movement, take measurements and record data. When granted permission from the user, ResearchKit can access data from Apple’s HealthKit app such as weight, blood pressure and glucose levels, which are measured by third-party devices and apps. ResearchKit can also request access from the user to access the accelerometer, microphone, gyroscope and GPS sensors to gain insight into a user’s gait, motor impairment, fitness, speech and memory.

Several world-class research institutions have already developed apps with ResearchKit for studies on asthma, breast cancer, cardiovascular disease, diabetes and Parkinson’s disease. Using the built-in templates for informed consent, users decide if they want to participate in a study and how their data—and which parts of their data—is shared. Participants can perform activities and generate data wherever they are, providing more objective information than simply filling out forms for their activities.

More data will be generated through these apps for researchers to analyze than ever before. For example, just four days after its release, Stanford University School of Medicine’s MyHeart Counts app was downloaded 52,900 times, with over 22,000 users consenting to the study. But more data isn’t necessarily better data.

Interested in beyond-the-pill pacts? More info here.  

On the surface, ResearchKit sounds like the long-awaited answer to ongoing issues in traditional clinical trial processes, including limited participation due to proximity to institutions running trials, frequent data entry and the integrity of that data and limited data collection.

Apple has created three customizable modules to address the most common elements across different types of clinical studies: surveys, informed consent and active tasks. Programmers can use these modules as they are, build upon them or even create new modules of their own.

ResearchKit initially includes five active task modules that invite users to perform activities under semi-controlled conditions, while iPhone sensors actively collect data. The tasks can be a simple ordered sequence of steps or dynamic, with previous results informing what is presented. In this way, researchers and programmers can create custom apps for their relevant disease states. These modules simply record the data and pass it on to the researchers; Apple does not store it or track it in any way.

Since ResearchKit resides on the iPhone, it will be easier to recruit participants for large-scale studies, accessing a broad cross-section of the population. The data that it collects mostly comes from sensors and other apps; there is little chance of error in the measurements as compared to patients recording their data in paper-based diaries. Even the data that patients will enter themselves into ResearchKit apps will be more accurate: programmers can put limits on that data so that it fits within proper parameters.

Although ResearchKit solves many issues with clinical trials, it also creates some of its own.

Patient population

Apple promises access to a diverse, global population through ResearchKit, but that population might not represent the population as a whole.

IPhone users are more wealthy and educated than the general population, and minority groups are underrepresented in its user base. Additionally, ResearchKit is only available on iPhone 5 and newer models and the latest generation of iPod touch, which excludes a large segment of iPhone users.

On top of that, the patient populations for ResearchKit apps will be largely self-selected:  those using the apps are already likely to be interested in their own health. So can the results generated from this narrowly defined population be extrapolated to the population as a whole?

Another point to consider with the self-selected patient population is that app desertion rate can be high, so researchers won’t have complete data from those who don’t finish the trial. This will also bias the data toward better outcomes since those who actually finish the trials are more motivated to see a positive outcome.

Data Validation

There is no validation that participants have a specific condition before they can enter a trial. This lack of verification can further skew the results of the trials. Going forward, tighter controls on who can enroll in each trial by verifying their basic information will lead to better qualified participants and more robust trial data.
 
Secure Communication

Verifying participants’ information might be hampered by the current lack of secure communication mechanisms between ResearchKit apps and their researchers’ servers.
 
This is up to the app developers to implement, as is HIPAA compliance and compliance with international research regulations. Even if secure communications are implemented properly by app developers, sharing personal medical information is a sensitive subject—especially with current data breaches. There will likely always be privacy concerns, especially in participants who don’t fully understand how their health data will be used.
 
Big Data

ResearchKit trials could potentially have hundreds of thousands of participants, each one with the potential to have inaccurate data. How will researchers separate the signal from the noise with such large amounts of data?
 
Cleaning that data will be a huge job, and further making inferences from that data to the general population could be difficult. Building trust in the trial results in light of the challenges listed here could be an uphill battle with the general public. More thought needed here.

Going forward, simple improvements such as data validation will go a long way toward more qualified patient populations and more robust trial outcomes. But how can ResearchKit be made available to a more representative patient population?
 
The answer could lie in the open source framework of ResearchKit. Researchers will have the ability to contribute to specific activity modules in the framework, like memory or gait testing, and share them with the global research community to further advance what we know about disease. And since it’s open source, there is the opportunity to expand into the Android realm as well.

On a global scale, Android is the far more popular operating system, and its user base is more representative of the population as a whole. It would benefit these clinical studies if users across platforms could use these apps.

That said, Android has a fragmented operating system with disparate hardware platforms that have differences in their sensors (accelerometers, GPS, gyroscopes), and even in chipsets from device to device. Researchers would have to account for all of these differences and build and test apps across platforms, which is nearly impossible on their limited budgets.

While ResearchKit is not the perfect solution for clinical trials research, it is a good first step, especially when it comes to to clinical trial recruitment, which has been the bane of the healthcare industry for far too long. Results of the pioneering ResearchKit apps—for asthma, breast cancer, cardiovascular disease, diabetes and Parkinson’s disease—will reveal the true utility of such a mobile, global medical research solution.

Martha Walz is a senior content strategist at Ogilvy CommonHealth Worldwide.