51˶

'Big Data' No Small Challenge for Health eHeart Study

MedpageToday
image

This article is a collaboration between 51˶ and:

When Kristen Downing signed up for the University of California San Francisco's ambitious after a heart checkup there, she was excited for what she thought was an opportunity to help others with adult congenital heart disease like her.

But then Downing, 45, started to realize what she had enrolled in wasn't what she thought.

"I was so excited about doing it but unfortunately as I got into it I realized it's highly geared toward acquired heart disease, which is a completely different animal," she told 51˶.

Another participant -- who, along with several others asked that her full name not be used -- came in through the study website where most of the planned 1 million participants will join and have their only contact with the study, but then accidentally indicated she had coronary heart disease instead of electrical problems with her heart in the online questionnaires.

Donald Bauer read the website but got confused as to just who could enter.

Yet another made it through online enrollment without incident but didn't hear anything further for so long she forgot she was in the study.

Only one of the five participants reached by 51˶ reported smooth sailing so far in what is projected to be a 10-plus year study.

It's so big, so digital, and so ambitious that it's no surprise Health eHeart is running into the same old research problems faced by other studies based on the FitBit and other digital health devices (See Digital Health Devices Running on Questions) but on an unprecedented scale and a few new ones too, experts suggested.

What's the "Big" Deal?

The study aims to stream in dense biometric data to be combined with periodic surveys, electronic medical records, and, for a subset of participants, genetics in a research-grade Internet database with the overarching goal of discovering new ways to tackle heart disease.

Participants will variously be armed with a range of smartphone-connected health devices and apps to track physical activity, sleep, blood pressure, heart rhythm and rate, pulse oximetry, weight, and 6-minute walk distance on an up to daily basis.

"I've been doing clinical trial research for over 3 decades, and I've never seen a study as innovative and potentially transformative as this," , of Brigham and Women's Hospital in Boston and president-elect of the American Heart Association, told 51˶.

, director of the digital medicine program at Scripps Health in La Jolla, Calif., agreed, calling Health eHeart a "modern era ."

There have been studies that are as large as Healthy eHeart hopes to be -- namely the Million Women and studies -- but none have such frequent measurements or have collected such a wide array of data.

The "nimble" adaptive design that will allow both observational and randomized substudies to be embedded at will is also unprecedented on this scale, explained Antman, who has been a liaison in the .

But breaking new ground adds uncertainty as to how well it can be done, acknowledged , one of the trial's primary investigators and director of clinical research at UCSF.

Wait, What's the Study About Again?

One unique challenge is how to make sure participants like Downing and Bauer actually understand what they're getting into.

Most Health eHeart participants won't ever meet with study staff to ensure informed consent. They must rely for information on a website geared to the broadest possible audience, as the aim is to enroll anyone 18 and older with an Internet connection regardless of heart condition status.

"If no one actually talks to the patient and the patient enrolls online, then it is hard for anyone to be sure that they really have understood the study (including its goal, risks and benefits)," cautioned , a research scholar at the Hastings Center, an independent bioethics research institute in Garrison, NY.

Even with the best informed consent process and patient materials, misunderstanding is endemic in medical research, Johnston noted.

"We might optimally think we would like people to totally understand the study ... but we've got decades of research data showing us people generally don't understand everything about the study they're enrolled in," she told 51˶.

That ranges from not understanding risk to not even recalling being involved in a research study, added , also a research scholar at the Hastings Center and editor of its journal .

Misperceptions about risk have typically been found in over half of patients in clinical trials. Assumptions about benefit may be even more pervasive, as in one phase I oncology study in which .

Those kinds of problems don't invalidate informed consent as long as participants are provided a , with materials at an appropriate reading level and so on, Johnston noted.

But she questioned whether the information given prior to joining the Health eHeart study is sufficient.

You Want Me To Do What?

The only outlines a basic requirement to answer health-related questions, check-in regularly, and stay in the study for a decade or more before participants click "I understand. I'm in."

After that, patients provide name, email, and birth date and are taken to the formal informed consent document, explained , another of the PIs at UCSF.

Additional informed consent modules are then introduced as participants are invited to different substudies or embedded interventions, he added.

"It's not all upfront," he told 51˶. "Maybe that's another possibility where some confusion might come in."

Still, a lower level of comprehension of the study, requirements, and risks might be acceptable given the low level of risk for the observational aspects of the research.

"Maybe it's okay if it's not as clearly explained as in a drug development study," Johnston acknowledged.

Most studies to enroll patients remotely through the mail, a kiosk, or other means have been this kind of low-risk study, Maschke pointed out.

Keeping it simple actually has been a top priority for Health eHeart, Marcus noted. It's hoped that will boost enrollment and keep people in the study for the long term.

Still, he acknowledged the balancing act required.

"It's an iterative process," he said, noting that the study itself is an experiment that will determine whether its research methods work.

Help, I'm Stuck!

Even for Downing, who got the scoop on the study directly at UCSF, what she described as a simple enrollment process through the website didn't go smoothly.

After logging in, answering the intake questionnaires was convenient and simple, up until she started feeling the questions were leading her into the wrong category, she said.

"I feel a little stuck in the process because the questions went to coronary artery disease and I don't have that," she told 51˶. "I don't know how to go further and I don't know how to back it up."

Each participant gets assigned to a research coordinator at UCSF they can contact with questions and problems, but Johnston noted that doesn't ensure understanding among others who haven't been proactive.

A web-based study is fraught with other pitfalls too, noted , director of the Jacksonville Center for Clinical Research and a past president of the Academy of Physicians in Clinical Research.

"Certainly computerization doesn't solve a lot of the problems of research, and, in fact, it introduces a whole new set of problems," he told 51˶.

Prior research in which participants have been sent devices -- as will be the case for some patients in Health eHeart -- has shown people don't always use them as intended or at all, he noted.

And in a study where there is no direct relationship with most participants, how will the researchers know if that's the case? he wonders.

With regard to verifiability of answers provided online, the plan is to validate those of a subset of participants against electronic medical records and in-person measurements, Marcus explained.

Who Are You, Again?

Another unique challenge is whether participants will stay 10 years without the kind of personal connection and investment that comes from meeting with investigators periodically at office visits.

"We fully recognize that there is this theoretical risk," Marcus noted. "It is a challenge unique, I think, to us."

Assigned research coordinators should help, he said, as should engagement through newsletters, social media, and "game-ification."

"We're taking a cue form other nonmedical Internet-based projects like in gaming, where people clearly keep coming back," he said. "We're going to have points and badges and statuses."

Money Matters

Despite the challenges, adaptive designs that allow for observational studies and clinical trials to be embedded like that in Health eHeart are likely to be used more and more because the phase III trial format is so expensive, averaging a little over $45,000 per participant, noted the AHA's Antman.

"Start doing the math and you see it is not a sustainable format for us to actually answer the clinical questions we have," he told 51˶.

The per patient cost of Health eHeart is expected to end up being an order of magnitude or two less, noted Pletcher. But with a million participants, those costs will still add up, although the organizers won't give an exact figure.

"In terms of cost, we are not trying to be evasive. We just don't know yet," another PI , chief of cardiology at UCSF, explained. "We are still in the development phase and continue to add and develop technology in the study. So the total cost of the study isn't really known."

While some industry partners donated a large number of digital health devices for use in Health eHeart, those companies haven't funded the work itself.

So far, the study is running on over $2.8 million grants and fundraising from , a pledged roughly $1 million contract from the , and "the donated time of all of us," Marcus noted.

"Anything can tank at any time, that's not unique to us," but the funding has been sufficient for what's been done to date, he added.

Most of the expense was in setting up the patient portal website and other infrastructure, such that each additional patient adds little to the margin, Pletcher explained in an interview.

"That's why we can afford to just open our doors and recruit," he said.

Participants like Kathi Sigona who have had a good experience in the trial are also recruiting their friends, family members, and social networks in hopes of adding to the "big data" picture on heart disease.

"I don't live in the lab, I don't think anybody does," she said. "So the more real-life experiences that can be recorded, the better [clinical] decisions can be made on that data."

From the American Heart Association: