One day in the future, all human data will be connected. But in the current parallel universes of healthcare and research, those seeking to share data face daunting challenges related to access, security and interoperability. To make matters worse, beyond traditional types of clinical and associated genomic data, there is no home for personal health data, such as behavioral, psychological and physiological data, that could be gathered through wireless devices and wearable sensors. And even if there were a place to collect and pool such data, the current healthcare system provides people with few incentives to donate them.

The digitization of clinical information has proven painfully slow. Today, over 80% of US hospitals and physician offices have an electronic medical record (EMR) system, but it has taken decades and billions of dollars to get to even this rudimentary stage. Similar amounts have been spent by national healthcare systems in Europe to develop EMR systems with better interconnectivity, but these too have low patient engagement; in Sweden, for example, only 400 of the country's 9.6 million inhabitants have accessed their EMRs in the past two years.

In the United States, the EMR universe remains an unstandardized patchwork of systems designed not for patients but to facilitate provider-oriented workflows and reimbursement. Platforms are built around documentation rather than collaboration with outside clinical institutions or researchers. Patients are also an afterthought, with clunky, provider-created 'patient portals' seemingly designed to discourage engagement.

In the research setting, few projects exploit EMRs because of the thicket of red tape surrounding patient consent, different data formats and lack of standardization. For the majority of human data gathered from clinical trials, results are the property of commercial sponsors, and patients hardly ever find out what happens to their data. For publicly funded studies, molecular and phenotypic data are lodged in a patchwork of repositories, many of which are structured and annotated in different ways, with complex legal and consent restrictions on how data are shared. And yet if human population research is to succeed, researchers will need to query data from many institutions, derived from millions, not thousands, of patients.

It is not all gloom and doom, however.

Late last year, the not-for-profit HL7 Argonaut Project was launched to create standardized services to facilitate sharing of health data, including EMRs. In the research setting, the Global Alliance for Genomics and Health has been working on its own interoperability standards and interfaces. One pilot initiative, Beacons (http://ga4gh.org/#/beacon), provides a single point of access to multiple genomic repositories at 20 different institutions. By assigning researchers different tiers of access (public, registered and controlled), it balances the desire for data sharing and the need for data protection.

These initiatives facilitate sharing and access to traditional types of clinical and genomic data. But what of efforts to share other types of patient-generated data?

One example is Sage Bionetworks' collaboration with Apple on the ResearchKit, a set of apps dedicated to recruiting individuals into clinical research. Sage provided not only input on the patient interface but also the computing environment that hosts patient-derived data. For the Parkinson's disease (mPower) app, it took just three months to enroll 11,360 patients—the largest Parkinson's trial ever assembled.

Similar to online offerings in other areas of human activity, convenience was one factor that drove rapid adoption of the ResearchKit. The apps are available on a ubiquitous platform (iOS devices) and provide interfaces that are accessible, understandable and easy to use—all attributes largely an anathema to existing offerings from healthcare providers.

Another key attribute of the ResearchKit is its utility. It enables a whole range of patients, who previously could not easily access clinical trials, to do so for the first time—clearly answering a pent-up, unmet need. Again, for most existing provider-created patient portals and open EMR systems (which many patients find impenetrable), utility is less obvious.

Perhaps the most important attribute of all is health data security. Data breaches are unhappily a fact of life in the online world. For personal and financial information they are serious enough, but for personal genetic information disclosure raises a whole other level of concerns for individuals and their relatives, affecting employment, health insurance and even paternity issues. Thus far, Apple has maintained its reputation for keeping ahead of data breaches. The same cannot be said of many healthcare providers, which in recent years have been subject to a lengthening list of security breaches. This year alone, nearly 90 million EMRs have been compromised, including identity, clinical and financial information.

In the face of such lapses, are we likely to entrust healthcare providers with more of our personal health data? A Commentary on p. 921 proposes that ownership be transferred back to individuals. It also outlines a new initiative, Unpatient (http://unpatient.org/), which borrows concepts (block chain and cryptographic hashing) from bitcoin, that may provide greater protection from hacks.

The question is, will the ability to take ownership of our health data also provide each of us sufficient empowerment and incentive to take greater control of our health? If we are able to control the types of health data we release, to determine who can use them and for what purpose—and equally have the power to rescind our permission—then are we more likely to embrace personal health data donation?

In this journal's view, placing patients at the center of the healthcare model would facilitate that empowerment. Whether people can be convinced that personal health data services are sufficiently convenient, useful and secure will likely determine uptake and engagement. Until such questions are addressed, the dream of large-scale human population research beyond biobanks will largely remain just that—a dream.