What Can Patient Registries Tell Clinicians About Possible Treatments?

Leaders in Genetic Autism Patient Registries Discuss Common Goals at Workshop

In the genetic autism community, there are dozens of patient registries that have been set up over the last decade. Each of these online databases contains a goldmine of information from thousands, or even tens of thousands, of patients. In the rare disease space, where so little is known or understood, patient registries may be the key to answering some of the most pressing questions about the genes that cause certain forms of autism and the behaviors that characterize such disorders.

Many of these patient registries were first set up by non-profit foundations and advocacy groups, such as the Dup15q Alliance or Phelan-McDermid Syndrome Foundation (PMSF), as a way to connect families and help them better understand the characteristics, or phenotypes, of a specific form of autism. Other databases were developed to accelerate scientific discovery and facilitate new treatments by research groups and government agencies like the National Institutes of Health, the Broad Institute, and Simons Foundation.

All of these registries have the same end goal: to benefit patients and their families. But is there a way these separate registries can help each other—and would it be beneficial to share knowledge by linking them together? Those were the questions foundation leaders, scientists, pharmaceutical representatives, and government officials came together to discuss on October 24th at a workshop hosted by the Alliance for Genetic Etiologies in Neurodevelopmental Disorders and Autism (AGENDA) and the Simons Foundation Autism Research Initiative (SFARI) in New York City.

“The goal today was to gather subject matter experts to help advise these groups in designing the next generation of registries to make them useful for both the patients and researchers,” says Jennifer Tjernagel, MS, senior project manager for SFARI and the Simons Variation in Individuals Project (Simons VIP). “We need to spend time defining the purpose of these registries—answer the what and why rather than jumping into the how.”

“We also want to get feedback from external users in order to shape the kind of data that is needed by stakeholders,” she added. “Then we will be well-positioned to start discussing technical considerations for designing these platforms.”

There is a need to engage more individuals with autism in research. Nearly one in 68 children has autism spectrum disorder (ASD) in the United States. ASD can be difficult to diagnose because it is associated with a wide range of cognitive and behavioral phenotypes. It is estimated that hundreds of genes contribute to ASD, and only about 20 percent of them have been discovered. Furthermore, there are no drugs currently on the market to treat the core symptoms of the disorder.

“We want to help each other and share the wealth,” says Russell Bromley, founder of Translational Research Acceleration Consulting. “We have common issues and differentiating issues. How can we go about integrating data as it relates to patient registries? Wouldn’t it be nice if we had an interoperable database where we could collect and deposit data that was relevant to all of us in a standardized way?”

What is a Registry?

First, the working group had to take a step back and define the purpose of patient registries. They seemed to fall into three basic groups: foundation registries (including contact registries and research databases), clinician-researcher databases (including natural history studies), and industry-initiated programs. As a result, the goals of these patient registries were identified as threefold: to accelerate scientific research, inform clinical trial design for new therapies, and connect patients and families who live with rare diseases.

“If we build something that is open source and flexible, and works for those three groups, it would be transformative,” says Stephan Sanders, PhD, a pediatrician and geneticist at University of California, San Francisco.

For end users in the research and drug development setting, patient registries offer access to large swaths of data and enable detailed analysis. They also inform clinical trial design and help with patient recruitment. Each entry contains a general description of the unidentified patient. If the subject fits into their inclusion criteria, the study coordinator can then contact the patient.

“A patient registry is an epidemiology tool that can be leveraged through the patient organizations. We can contact the patients who are eligible for a targeted trial and say ‘this is available for you,’” says Michelle Krishnan, PhD, translational medicine leader in rare diseases at Roche who points to Simons Simplex Collection (SSC), Simons VIP, and SPARK for Autism as examples of registries with clean, accessible and usable information for pharma.

Michelle Krishnan, PhD, of Roche speaking.

“If pharma is one of our principal users of a registry, we need to help them find the right patients for trials of new treatments because ultimately we want them to develop therapies,” adds Mr. Bromley.

On the patient advocacy side, registries offer difficult to obtain statistics about rare disorders. For many families, these data are their first window into understanding co-morbidities, the effectiveness of treatments, and changes that commonly develop over time. For example, the Dup15q registry offers data points on how many children with ASD also have epilepsy. In addition, the registry allows families to connect with each other for support and participate in educational events like webinars.

“We know we have things in our database the FDA will never be interested in, but they are going to help families go to school, design their IEPs, and understand their child’s prognosis in the coming years,” says Liz Berry-Kravis, MD, PhD, professor at Rush University Medical Center.

What Kind of Information Should a Registry Have?

Most registries contain hundreds of questions that are completed online by parents or adult patients. Others include clinical information that is entered by study coordinators when patients visit a research site location. Each family is assigned a subject ID number to protect their confidentiality. But what information is most pertinent to the end user? What questions would be included in a core data set?

It all comes back to the end users and the goal of the registries. “Does it help with school? Does it help with research? Does it help the development of a new drug? If it doesn’t meet those criteria, we shouldn’t collect it,” said Dr. Sanders.

Genetic Data: A Must Have

Most participants agree that the genetic report is a critical component in these registries. This genetic information might include entire genome sequencing, if it is available, or a simple cellphone snapshot of a genetic report that is uploaded to the registry web site or an app that would be developed.

To make a real difference, Dr. Sanders says he would need genetic information on 200,000 patients. Some registries, like Simons VIP Connect, have an associated biorepository with whole blood DNA, saliva DNA, and some fibroblasts. The Broad Institute is linked to labs that process and ship sample collection kits for genetic sequencing.

However, obtaining the genetic information isn’t the only roadblock. Wendy Chung, MD, PhD, a clinical geneticist at Columbia University, advises that the reports also need to be curated by trained genetic professionals. For example, PMSF uploads genetic data to an i2b2/TranSMART database where geneticists at Harvard interpret and verify the genetic information. One suggestion was to have a genetics core use resources like ClinVar verify the diagnosis. Once the genetic information is validated, the family would be contacted to complete further questions.

But the devil is always in the details. It is difficult to find genetic professionals who understand the nuances of the disease and there is a risk of oversimplifying the registries. Some disorders overlap, while others are loosely defined because only 10 cases have been identified thus far. Despite these challenges, working together would reduce the cost of mining through cumbersome amounts of genetic data.

Core Data Set Components

When it is accessible and easy to use, registry data yields the potential for amazing results. In his presentation, Dr. Sanders discussed how cross-disorder analysis is essential to understanding the neurobiology of genetic autisms.

For example, registry information could be used to develop behavioral “growth charts” specific to rare genetic disorders, similar to those that pediatricians use to track height and weight. By understanding the genetics of a single gene and correlating it with rich phenotype data, such as IQ and age of walking, researchers can draw correlations that are beneficial to the patient and could potentially predict who will have a genetic mutation.

“This type of resolution excites me. It is achieved by having a rich database and controls. If we can get our data to a place where it is accessible, we can begin to find genetic correlations that relate to the masses,” explains Dr. Sanders.

However, IQ and intellectual disability are also the toughest measurements to obtain. While scientists agree it is one of the biggest research needs, advocacy groups feel that the exact IQ can be a difficult question for parents to answer. Dr. Chung said a functional assessment might be more appropriate.

Other key components that were identified included contact and demographic information, developmental history, and functional adaptive behaviors. The importance of natural history study information was also debated. Dr. Krishnan argued that natural histories are particularly useful in rare diseases.

However, the data collected also needs to be longitudinal so the registry remains evergreen. A child with one of these genetic differences was likely very different five years ago than they are today. “Life is a movie, not a picture,” says Liz Horn, PhD, MBI, a principal investigator at the Bay Area Lyme Foundation and PMSF.

What Type of Features Do Registries Need to Have?

It is hard to argue that an open-source, centralized registry that includes different rare autism genetic disorders would not be beneficial for all stakeholders. But there are numerous technical, logistical, ethical, and monetary concerns that come into play.

Randall Carpenter, MD, of the Rett Syndrome Research Trust speaking.

If a common registry were developed, it would have characteristics of both a patient registry and a researcher database. “You want to develop something that is FAIR—findable, accessible, interoperable, and reusable,” says Alexa McCray, PhD, professor of medicine at Harvard Medical School. “We need to foster a culture in which data sharing and open access are the norm.”

In order for the registry to be effective, it would have to be secure but operate with nimble governance, allowing scientists from all over to mine basic levels of data. Tiers of access were suggested. While anyone might be able to view the basic aggregated data, researchers and clinical trial recruiters who want to take a deeper dive could obtain an additional level of clearance. Each patient would also need to be contactable.

The data would also need to be harmonized to a common standard. In the past, time and resources have been drained cleaning up bad data. A common data dictionary would need to be developed for all terms. There might be ten different ways users write the term “years old.” A list of options to choose from, or binary data, is more usable than open-ended responses.

Registries must be patient-centric and easy for families to use. Those registries that are burdensome for parents to complete need to be pared down to include only the most critical questions for end users. For example, families in the PMSF registry are asked 300 different questions that yield 1,700 different data points.

“Our families will do anything the research community asks them to do, but we don’t want to ask our parents to fill out information that we won’t use,” says Megan O’Boyle, principal investigator of the PMSF Data Network.

Features that allow the user to save their work and return later increase usability. Automated reminders should also be sent out to parents who need to update information. Additionally, registries need to give information back to the patients through progress updates, annual reports, and survey results that allow the families to compare their responses to others in their community.

Next Steps

With these challenges and benefits in mind, the participants discussed how patient registry data might be integrated. Two options were identified: building an entirely new platform housed in an open-data community system like i2b2 that each organization would encourage patients to enter data into, or making all of the preexisting patient registries speak to each other.

Steve Roberds, PhD, chief scientific officer from the Tuberous Sclerosis Alliance, favored the interoperability mode because diagnostic labels are not always fully understood. “People with Tuberous Sclerosis Complex and many of these other diseases are often diagnosed with that disease before they are diagnosed with autism, so they aren’t going to be in an autism registry. There are a lot of overlapping phenotypes. We share autism with everyone in this room, epilepsy with a few people in the room, and tumor growth with others,” he said.

One idea was to aggregate all the genomic information that is currently available across every registry platform into a single database. The profiles would link back to the organization that supplied the information if a researcher or clinician wanted to dive deeper or contact the patient. Ideally, the registry would also collect information on family members, including biological parents and siblings, which may be used as control groups.

Most participants agreed it would be a shame to waste the time and resources that have already been spent setting up the existing registries. In fact, some registries are already interoperable and share certain information with the National Database for Autism Research (NDAR), for example.

“There is great value in what exists and teasing out these genetic causes of autism, but there also needs to be some commonality,” says Ms. Horn. “Is that best achieved through a new way of getting the data or in taking the data that exists and transforming it?”

The ultimate goal of any registry is to develop treatments and cures that benefit patients and their families. Rare diseases are a promising field where patient registries can have a transformational impact. With stakeholders from research, pharma, and patient groups working together, scientists are confident that they can piece together a clearer picture of what causes, characterizes, and transpires across the varying forms of autism.