As Internet-based research evolves, IRBs and PIs need updated guidance

Data security, collection, IC, recruitment are focus

As Internet use has exploded worldwide, so has Web-based research. Between 2004 and 2009, the number of web-based research studies published in the American Psychological Association's (APA) Journal of Personality and Social Psychology rose by more than 500%.1

Even within Web-based research, the methods and strategies have been continually and rapidly evolving. This has made it challenging for IRBs when reviewing these studies for issues important to research participants, including privacy and confidentiality, experts say.

"One of the things that's easy to overlook in Internet-based research is this expectation of privacy versus information that is placed online for the public," says Vern Paxson, PhD, a member of the social-behavioral research committee at the University of California–Berkeley.

"There's a tendency among researchers to think if they can grab it over the Internet that it's fair game," Paxson adds.

The office for the protection of human subjects at the University of California–Berkeley recently completed new guidance on conducting Internet-based research. The eight-page document, which is available on the office's website, focuses specifically on recruitment, informed consent, data collection, and data security.

"We review many social-behavioral applications here, and we were beginning to see a trend with applications involving Internet-based research. So we saw a need to develop guidance for our investigators," explains Adrienne Tanner, CIP, IRB coordinator in the office for the protection of human subjects.

The IRB office spent about one year writing and reviewing the guidance. The development time was long enough that the types of Web-based research reviewed by the IRB evolved and required some updates in the guidance, Tanner notes.

"Internet research is constantly changing, so we'll probably make more revisions in the future," she adds.

The guidance addresses specific issues investigators encounter when planning a study that involves some Web-based research. For example, one part discusses how it's challenging to properly identify and qualify subjects.

"Without face-to-face or voice-to-voice interaction, it's difficult for investigators to be sure that participants are not misrepresenting themselves," the guidance reads. "In certain situations investigators should discuss measures taken to authenticate subjects."

Interactions with subjects via the Web is more complicated from a research ethics perspective, Tanner says.

"When contacting someone through Facebook, is that an acceptable practice?" she asks. "The person using their social network working site might not feel that way and might have the expectation of privacy."

When developing guidance involving Web-based research, it's important to use the flexibility inherent in human subjects protection regulations, suggests Rebecca D. Armstrong, DVM, PhD, director, research subject protection in the office for the protection of human subjects and responsible conduct of research coordinator at the University of California–Berkeley.

"Our IRBs at Berkeley are very good at utilizing the flexibility that is in 45 CFR 46," Armstrong says. "This enables us to alter consent forms or waive documented informed consent to facilitate research without increasing risks to subjects; it's a balancing act."

The new guidance includes seven main points and examples about conducting informed consent online.

"I recollect a protocol that came to a full committee review that had a researcher who wanted to download an individual's Facebook page at one point in time," Armstrong recalls. "There was a lot of discussion because of the privacy issues involved."

The purpose of the research was to examine how individuals signal social status and express their individual characteristics through online social networking sites. This particular protocol raised issues of secondary subjects, such as the people whose identifiable information appeared on the main research subject's Facebook page, she said.

"In a Facebook profile you have photos, names and identifying information of yourself and your friends, and your friends have not consented to have this information made available to researchers even if you have consented to it," Paxson explains. "It's a complex determination to decide how to protect those subjects or even to identify whether you have that problem."

The study involved looking at Facebook profiles to analyze pictures, status updates, posted links and "check-ins." The IRB had the investigators remove all identifying information pertaining to secondary subjects before coding the data.

"Our IRBs tend to be creative when coming up with solutions," Armstrong says.

Another common issue with Web-based research involves determining the level of review needed, Tanner says.

"For example, if an investigator wants to use publicly available, existing data for their research, it would not need to go through a review," she says. "But it's difficult to know if data published online is publicly available or not, and it depends on the intention of the person posting that information."

The guidance addresses that issue by focusing on the use of a login or registration: "Data only accessible through special permission or registration/login (with username and password) are generally not considered public. When determining whether or not data are public, the investigator must decide if there exists an expectation of privacy."

This nuance might occur when someone posts online personally identifiable information about someone else without asking the person for permission, Tanner says.

"We wanted to let investigators know these are factors to consider," she adds.

One of the goals for the Web-based research guidance is to prod investigators into thinking about human subjects protection issues that they might otherwise overlook, Armstrong says.

"That's how we frame our guidance, and we give them lots of examples to make it real for them," she explains.

From the IRB's perspective, the guidance is very useful, Paxson says.

"It's easy for investigators to assume that Internet communication is private, but that's not a sound assumption," he says. "While a best effort is made for confidentiality, it is not a guarantee."

The IRB asks investigators to include in their informed consent documents a sentence explaining how there is no way to guarantee confidentiality, Tanner says.

"We want to make sure participants are not under the assumption that their information is 100% secure," she adds.

At UC–Berkeley, investigators can use template language developed by the human subjects protection office. These templates are available online.

"We also have developed an online consent builder that walks investigators through the process by answering questions," Armstrong says.

Reference:

  1. Denissen JJA, Neumann L, van Zalk M. How the internet is changing the implementation of traditional research methods, people's daily lives, and the way in which developmental scientists conduct research. Intl J Behav Dev 2010; 34(6):564–575.