Apple iPhone Power: Thousands Sign up for its Health Tool Amid Bias Criticism

Bloomberg – Stanford University researchers were stunned when they woke less than 24 hours after the launch of  Apple’s iPhone tool, the ResearchKit,  last week to find that 11,000 people had used it to sign up for a cardiovascular study .

“To get 10,000 people enrolled in a medical study normally, it would take a year and 50 medical centres around the country,” said Dr Alan Yeung, medical director of Stanford Cardiovascular Health.

“That’s the power of the phone.”

With ResearchKit, Apple has created a pool of hundreds of millions of iPhone owners worldwide, letting doctors find trial participants at unprecedented rates. Already five academic centres have developed apps that use the iPhone’s accelerometers, gyroscopes and GPS sensors to track the progression of chronic conditions like Parkinson’s disease and asthma.

At the same time, other researchers caution that potential flaws in the information gathered through ResearchKit may make the data less useful. The software’s consent forms may not be clear enough, or its applications won’t capture data fully and accurately and protect the privacy of participants, they say.

“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,” said Dr Lisa Schwartz, professor at the Dartmouth Institute for Health Policy and Clinical Practice, in an e-mail.

“Bias times a million is still bias.”

Misleading data?

For starters, the average iPhone user is more likely to have graduate and doctoral degrees than the average Android user, and has a higher income as well, according to polling company CivicScience Inc. Those sort of demographic differences could skew the findings from a study.

Misleading data can also come from a user accidentally hitting a button or giving the phone to someone else, said Dr Michael Gibson, a professor at Harvard Medical School and an interventional cardiologist.

And apps on a phone may be more restricted in the types of questions they can ask than standard trials, which allow researchers to ask open-ended questions in face-to-face encounters. Asking about specific side effects – “Mrs. Jones, are your teeth itching?” – may prompt false memories and make people more apt to report them, a problem that an open-ended question wouldn’t have triggered, Gibson said.

False reporting

Yet the iPhone also helps address a problem that standard trials often encounter: People enrolled in studies often falsely report their activity to researchers. By using its internal components or secondary devices connected wirelessly via Bluetooth, the iPhone can silently measure users’ behaviour, without relying on them to keep track or be honest about what they’re doing.

“People don’t want to say they did zero exercise – they want to say they did something.” Stanford’s Yeung said. “They don’t really tell us the truth.”

The US Food and Drug Administration supports tools that make data collection during clinical studies easier and more efficient, Andrea Fischer, a spokeswoman at the agency, said in an e-mail.

“Mobile platforms and other tools could potentially contribute to efficient data collection, which could mean getting medical products to patients faster,” Fischer said.

ResearchKit isn’t considered a mobile medical app, she said. Mobile medical apps that could be used to diagnose or treat a disease require FDA approval.

Modifying behaviour

Stanford researchers are using their ResearchKit app to study ways to encourage people to modify their behavior to improve heart health. Their app aims to automate as much data collection as possible, Yeung said. Participants will be asked to keep their phone on them as much as possible for a week, letting the GPS and accelerometer track their activity.

At the end of the week, users will be asked to do a “poor man’s stress test” by walking as fast as they can for six minutes. Then participants will be randomized to different types of coaching, through games or with more basic reminders, and three months later will be asked to repeat the one-week intensive tracking and stress test. The results may show what types of coaching are most effective in improving fitness.

Accurate tracking

Other researchers are also looking for ways to use the iPhone to more accurately track behaviour. A team at the Icahn School of Medicine at Mount Sinai, working with digital health company LifeMap Solutions Inc., is studying whether having an iPhone app that educates asthma patients and reminds them to use their inhaler can improve symptoms and reduce doctor visits.

While the app currently requires patients to manually enter information, the team is working on an add-on to a standard inhaler that will send a Bluetooth signal to a patient’s phone, logging use automatically, said Corey Bridges, LifeMap’s chief executive officer.

As of the launch, more than 2,500 people had enrolled and consented to participate in the asthma study, according to LifeMap.

The more that data-gathering can be automated, the more the app may be able to reduce the risk of dropouts and patients lost to follow-up, a group dreaded by investigators because it is so difficult to know how they fared after participating in a trial, said Gibson. ResearchKit will at least let researchers know if their patients are still alive, he said.

Parkinson’s app

The data may not be perfect, but many concerns about ResearchKit – such as whether the patient sample is representative – are issues with traditional clinical trials as well, said Todd Sherer, CEO of the Michael  . Fox Foundation for Parkinson’s Research, which has collaborated with nonprofit group Sage Bionetworks on one of the apps.

The Parkinson’s app had 5,589 consenting users by Tuesday morning, according to Sage. Sherer said he didn’t know the cost of developing the app, but the foundation’s biomarker study, a traditional trial with almost 800 participants over five years, has cost about $60 million.

“I don’t think we want to give the perception that this type of research will replace the more standard, physician- based, direct interaction with the patient” in traditional trials, Sherer said. “But I do think this provides a complementary type of research in a different way. Any kind of tool that will make it easier to engage more people in research is really important.” – Bloomberg