In March 2015, Apple announced the launch of ResearchKit, an open source software framework designed to help doctors and scientists gather more data more accurately through iPhone apps.
Not only does this technology have the potential to significantly increase the volume of patient data available to researchers (as apps built on the platform can access data generated using iPhones’ built-in sensors), it could also make clinical trial recruitment hugely more efficient and cost-effective by widening researchers’ reach and through the use of ResearchKit’s participant consent module, as well as encouraging other consumer companies to follow Apple’s altruistic lead.
Or so biomedical researcher, professor and TED fellow Max Little, one of the collaborators behind a Parkinson’s app developed on the ResearchKit framework, believes. "Probably the most exciting thing for me as a researcher [about ResearchKit] is that Apple has committed to release this platform open source, free for the community to use as they feel fit," he says. "It’s an unprecedented move by the world’s most valuable company which will probably set a strong, positive example for others to follow."
Collecting patient data through smartphones
Little is no stranger to using smartphones for collecting patient data. Since 2012, before getting involved with the ResearchKit platform, he has been working with collaborators in the US on developing apps for Android devices to collect behavioural data from Parkinson’s sufferers, to already incredible effect.
"I was dissatisfied with having to collect objective behavioural data on Parkinson’s using in-person lab appointments and the only alternative up until that point was using the internet and unreliable voice recording methods," he recalls. "Since smartphones had ‘matured’ by , we were able to get cheap devices, which looked like an excellent technology to do remote monitoring, because the sensors built into these devices are numerous and very high quality."
laboratory at Rockefeller University is working to identify new, clinically relevant molecules from soil bacteria.
He and his colleagues have already managed to collect a vast amount of data across several studies. "For example, we have continuous ambulatory and structured test data across many different ‘modalities’ such as gait, dexterity and reaction times, voice, balance, postural and rest tremor, social interaction and mobility," Little notes. "One study set up last year has this kind of data from several hundred participants over about six months each, while another, based at Oxford University, has structured test data from nearly 500 participants, which is paired with ‘deep phenotyping’ data such as full exome sequencing, blood analysis, fMRI, cerebrospinal fluid samples and sleep disturbance.
"It is very comprehensive and we are able to combine the behavioural data collected using the smartphones and this phenotype/genotype biological data."
iPhone apps: the next logical step
Expanding into the iPhone arena was simply the next logical step for Little, and it was one he decided to embark on with Sage Bionetworks, a non-profit biomedical research organisation he’d been collaborating with since 2013.
And so the mPower app, which was set up to accurately measure the symptoms of Parkinson’s using only the iPhone’s built-in sensors, was born. "These sensors can potentially capture symptoms like tremor, voice problems, problems with finger movement and the like, and the app then sends this data to a server where algorithms provided by medical researchers can analyze the sensor data and turn it into meaningful information," Little explains. "This can be used to further improve our scientific understanding of the disease."
Reaching the ‘other half’ of the consumer market
Now Little has access to the iPhone half of the consumer market, as well as Android users, the possibilities for his research have expanded significantly. "iPhone is the ‘other half’ of the consumer market so you simply have to cater for it if you want to reach out to the most number of participants," he says. "Moreover, there may also be subtle confounding factors about the demographics of Android versus iPhone users and we want as diverse samples of behaviour as possible."
And already, just a few months on from mPower’s release, ResearchKit’s incredible reach has been clearly demonstrated. "mPower and the other four apps using ResearchKit launched to massive appeal and in total, as I understand it, the downloads are nearly in the 100,000 participant range already," Little remarks.
"That is truly fantastic because it demonstrates the interest and willingness of ordinary users to participate in this kind of research without any obvious reward to themselves. It really does show that if you bring the trial to the participant, you can achieve massive scale at next to no cost."
Over the next five years, Little also hopes to prove beyond doubt that technology such as ResearchKit has real value for his particular field of research, Parkinson’s. "One thing I’m hoping to see is a proof-of-concept of the use of smartphones as population-wide screening tools for the early symptoms of Parkinson’s, which I think is an enormous unmet need," he notes.
The future of data collection
More generally, the ResearchKit framework, which has also been used to develop apps for studies on asthma, breast cancer, cardiovascular disease and diabetes, really could be the future of patient data collection, according to Little. "It lays a concrete foundation for a highly accessible, low-cost, objective clinical behavioural data collection platform," he says.
"Granted, we can’t yet collect biological samples this way, so it certainly won’t displace lab or clinic visits – yet; I would bet on technology like this being closer than we think. But, there are vast ‘oceans’ of patient research that just doesn’t happen because traditionally, it is hugely expensive and perhaps impossible to do, for example, collecting high-frequency information about fluctuations in motor symptom severity, or linking patient movements to precision environmental databases, or objectively quantifying the social interactions of participants on a day-by-day basis.
"All of these aspects are quite easy to do with platforms like ResearchKit and the data we will be able to collect this way could provide the answer to some long-standing scientific questions in drug efficacy and disease progression modelling, just to take two examples."