Earlier this week, Apple gave us wrist computers and took away almost all of the ports in its notebook computers, and also announced something that gadget fans may not have expected: a set of apps called ResearchKit designed to help medical researchers collect data from ordinary citizens for their research. Tens of thousands of people have already signed up for studies, which is potentially great for science. Is it good for us, the potential research subjects, though?
The Stanford heart disease study, for example, uses a series of screens that users must scroll through to explain the responsibilities of study participants and its possible effect on our lives. They’re logical and easy to understand, but do they tell the entire story?
People seem very thrilled with the potential of HealthKit, but what about its potential to violate our privacy and enable our inner hypochondriac? I checked with Nicholas Evans, a postdoctoral fellow in advanced biomedical ethics at the University of Pennsylvania. (Disclaimer: Dr. Evans is married to one of my friends.) He downloaded the asthma study app and found numerous problems with how it explains the study and the app to participants. “All in all, this appears to be a solid attempt to design an informed consent document for the digital age, but it reads more like an End-User License Agreement with some extra bells and whistles,” he observed after pretending to sign up for the study. There’s a big difference between an e-mail app and a program meant to detect hand tremors or determine your risk of heart disease.
ANONYMITY ISN’T GUARANTEED
One key promise that Apple made when announcing ResearchKit was that data will never pass through Apple’s hands while being submitted. Stanford says that survey data will be anonymized by a nonprofit third party, Sage Bionetwork, with participant names stripped out and numbers subbed in. In a smaller and more specific study, it may be possible to identify participants by looking at their demographic data.
ResearchKit apps also allow participants to share their data with other “qualified researchers,” without making it clear who those researchers are (nobody knows yet) and what they could be using the data for. While the screen to opt in or out of data sharing explains this, there’s no room to explain the complex relationships between universities, hospitals, and private companies. “[The data-sharing screen] elides what they are going to do in terms of other researchers, who may have commercial interests or ties to third party commercial entities,” observes Penn’s Evans. Yes.
These disclosure screens may be part of other research study apps, and they simplify (arguably, they over-simplify) the possible risks and benefits of taking part in a study.
Current practices for recruiting human subjects in an experiment presume that researchers and subjects are interacting in person in some way.
ResearchKit allows a study to gather huge numbers of people. These people might get bored or drop out, or they could make up their data points or lie. Anyone who thinks that someone wouldn’t invest significant time and energy in pretending to have a serious illness hasn’t spent much time in the darker corners of the Internet. There’s a difference between an app-based data collection effort and full-blown Munchausen’s by Internet, but there will always be the risk that motivated trolls and liars can sign up.
It might not do any good to gather huge amounts of data from people all over the country to begin with. The obvious problem is that people who own smartphones tend to be wealthier than people who don’t. Another professor at Dartmouth explained to Bloomberg News that all of this data collection could be a waste of time that doesn’t advance science at all. “Just collecting lots of information about people — who may or may not have a particular disease, and may or may not represent the typical patient — could just add noise and distraction,” Lisa Schwartz of the Dartmouth Institute for Health Policy and Clinical Practice explained in an e-mail.
Apple’s marketing of ResearchKit plays up its promise, but maybe all we are doing is collecting servers full of anonymized noise. Taking part in medical research doesn’t always benefit you, personally: studies that follow people with particular risk factors or that test treatments that are ineffective don’t directly help anyone.
Should you take part in research through HealthKit, if you own an iPhone? That’s up to you. Carefully read the disclosures, and be cautious about what personal information that you share. Research with human subjects that tries new methods can have unforeseen consequences.