Roche subsidiary Genentech is no stranger to online patient communities -- the pharma company signed a five-year data access deal with online patient community company PatientsLikeMe in April. At the Health 2.0 conference in Santa Clara this week, Genentech product development leader Joling Mew spoke about the ways that Genentech has worked with PatientsLikeMe, and other online patient communities, to address persistent problems in its clinical trials.

"When it comes to designing a clinical trial, how do you actually put the patient first?" she asked. "If you work in any other kind of consumer-facing product, the answer is fairly obvious: You ask your customer, [through] surveys, focus groups, or observing their behavior. We need to find a way to better understand our patients, as people; their friends and family, their hobbies and interests, and not just as individuals defined by their disease. "

Mew spoke about how consulting patient communities like PatientsLikeMe, Crohnology, SmartPatients, Inspire, and MediGuard helped the company to crack some hard clinical trial cases, and ultimately increase enrollment and retainment.

In one study tied to prescription flu prevention and flu treatment medication Tamiflu, Genentech researchers were attempting to learn about the effectiveness of the drug on patients with weakened immune systems because of transplant surgery or some other cause. But in five years of attempting to enroll the study they only had about 25 patients. They hypothesized that immunocompromised patients were simply too cautious to get sick in large enough numbers to enroll a sufficient sample size, even at multiple sites.

That's when Genentech reached out to MediGuard to survey 150 immunocompromised patients. They found that 60 percent of immunocompromised patients had flu like symptoms at some point. The reason they weren't encountering those patients at the enrollment sites is that they weren't going to the hospital with those symptoms -- they were either going to a primary care facility or straight to the pharmacy.

The team implemented several new interventions based on that information: they set up a virtual central site using telemedicine, recruited patients via a pharmacy outreach service, and had healthy immunocompromised patients preregister so they could flow into the system if they later got sick. These measures were able to increase enrollment from 25 to 88.

Another trial had to do with a rare disease called ideopathic pulmonary fibrosis. This disease had no existing treatment options in the United States, so Genentech expected patients would be more enthusiastic about clinical trials; instead, once again, enrollment was low. In this case, the company consulted with PatientsLikeMe and Inspire.

They found that there was a lack of education leading many patients to not consider a clinical trial to be a treatment option on the same level as an organ transplant, even though both options were high risk approaches. Additionally, patients were very concerned about getting their hopes up in a clinical trial and ending up in the placebo group.

"These are very sick patients," Mew said. "The thought of not being able to be on medication is going to be a strong deterrent. [Trial designers] couldn’t get rid of their placebo arm, but what they could do was change the ratio of drug to placebo to make it a little bit more appealing."

A final trial, of patients with inflammatory bowel disease, was challenging for the opposite reasons: the field was very competitive, and patients had a lot of treatment options. What a patient survey, conducted through the networks Crohnology and MediGuard, revealed to Genentech was that fear of embarrassment was a major demotivator for signing up for an experimental treatment. Researchers were able to consider telemedicine, remote monitoring, and home nurse visits in the trial design to help mitigate those fears.

Mew said that Genentech works with a number of different online patient communities because different communities offer access to different patient populations, and it's essential to hear from the right group.

"It really needs to be done per patient population," she said. "Some of the things we’ve learned in a particular patient population, if we took that and assumed it to be the same for all other patient populations, we would be grossly mistaken."