The National Institutes of Health’s Precision Medicine Initiative (PMI) Working Groupconvened at Intel’s corporate headquarters in the heart of Silicon Valley late last month for its fourth and final public workshop before issuing recommendations to National Institutes of Health (NIH) Director Francis Collins in September. The agenda focused on how data from mobile health (mHealth) and other personal technologies can be combined with genomic and medical data to enhance the reach, effectiveness, and conclusions of the PMI’s million-person research cohort. Given PMI’s focus on inclusiveness and diversity, it is important to note that 90 percent of all Americans have a cell phone, and 64 percent have smartphones. This high degree of prevalence holds across gender, race/ethnicity, age, and social-economic status, meaning that mHealth tools will enhance participation, not exclude large swaths of Americans.
Collecting participant data
One set of discussions addressed the potential benefits and barriers of employing mobile technologies to collect participant data in a cohort study of this size, such as:
- Designing studies to be mobile from the start: Leaders from MD2K and the Health eHeart and Parkinson’s mPower studies described the challenges and solutions they unearthed when designing mobile research studies, such as the danger of a device manufacturer pushing out a software update in the middle of a study that unexpectedly changes data formatting. To maximize the value of mobile platforms, designers needed to shift their thinking from using highly precise data that could only be collected from small groups (e.g., through clinic visits) to using mobile-derived data that may be less precise but come from vast, geographically diverse populations.
- Integrating data from consumer devices: Data scientists from Open mHealth and Evidation Health provided suggestions on how to manage the quality, integration, and validation of data streaming from a diverse array of devices (e.g., smartphones, activity trackers, connected scales) – most of which were designed for consumer use and not medical research. Each manufacturer typically uses different data formats and proprietary algorithms for calculating steps, distance, etc., which creates a challenging environment for researchers looking to draw meaningful conclusions.
There are activity trackers designed to meet the needs of researchers (for example, the Framingham Heart Study is using the customizable Shimmer platform). However, several workshop participants noted that in spite of the challenges, they’ve started using commercial devices like those from Fitbit because their low cost and independent validation studies (like these) have facilitated data collection and interpretation.
There didn’t appear to be any representatives from companies like Fitbit or Apple at the workshop, but the head of Samsung’s Simband program talked about the development of its new mobile device designed to facilitate research via open access to all of its sensors and data streams.
Maintaining long-term engagement of cohort participants
In addition to collecting data on participants, the workshop explored how mobile technologies can be used to maintain their long-term engagement. Given the NIH’s goal of recruiting both healthy individuals as well as those living with disease, the question of motivation, and how to sustain it across audiences, emerged as a central theme:
- Demonstrating value in real time: Panelists urged the PMI to find ways to provide value back to all the participants, perhaps via useful health tips or other feedback on their phone. Several panelists advocated for the ability for a participant to do “n of 1” experiments using their own data to give them insight into their health status.
- Building a foundation of trust: As with any discussion of mobile technologies, security and privacy were examined at length. While data de-identification and other best practices would be standard operating procedure, it is likely that the PMI will suffer a data breach at some point. The degree to which the initiative has built a solid foundation of trust with its participants will enable it to recover (or not) from that breach. One very interesting exchange occurred when a panelist was describing how the initiative could maximize security by encrypting all data at the point of collection, and only enable decryption at central enclaves controlled by PMI. Collins spoke up to caution everyone that one of the goals of the initiative is to empower a wide array of people to innovate using the data that are collected. While it may be straightforward for academic researchers to make the connections and get the credentials to enter an enclave, it would undoubtedly leave many non-traditional innovators out in the cold.
With the fourth and final workshop completed, the PMI working group will now turn its focus to synthesizing all that they have learned over the past few months. While it has never been far from their minds, this final workshop saw several working group members ask pointed questions on the “must-haves” versus the “nice-to-haves.” Studying the health of 1 million Americans over decades offers many scientific opportunities and to address all of them would likely result in an unwieldy study design (akin to what doomed the National Children’s Study). Additionally, the Obama Administration has requested $130 million in fiscal year 2016 for the million-person cohort. The working group members have done the math and want to provide constructive advice on how NIH can most effectively use the $130 per participant during the effort’s first year (assuming that Congress appropriates the full amount requested).
In some respects, personalized medicine is already here (for example, matching therapies to patients who have specific genetic defects is common in oncology). PMI’s goal is to move beyond those islands of precision and develop a systematic research plan and scientific resource that can yield insight into all illness. FasterCures is committed to fostering exactly this sort of enterprise-wide transformation of research and medicine. We look forward to the working group’s recommendations and the opportunity to help the million-person research cohort launch and thrive for years to come.