The Children's Hospital of Pennsylvania has received $1.8 million from the Newborn Screening Translational Research Network to build a long-term repository of clinical records for infants diagnosed with genetic disorders through newborn screening tests.
The database, dubbed the Long-Term Follow-Up Data Collection Tool, is part of an ongoing five-year project led by NBSTRN to create an infrastructure that supports research associated with newborn screening tests performed on all babies born in the US.
NBSTRN was created in 2008 with funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development. It is coordinated by the American College of Medical Genetics.
Over the next three years, CHOP will build a system to collect, manage, and disseminate clinical data collected by participating centers. The repository will store de-identified clinical records of newborns who test positive for any of more than 80 currently tested genetic disorders and whose positive diagnosis has been corroborated by confirmatory tests.
It will also include information from the results of follow-up tests, disease complications, medications, and treatment records collected by large clinical centers that will follow these patients until they reach adulthood, CHOP said.
The repository will store the genetic data related to the newborn screening tests and will also store a limited amount of genomic information— for example, data on pathogenic variants — but not much since most newborn screening tests are not genomic-based, Peter White, director of CHOP's Center for Biomedical Informatics, told BioInform.
He added, however, that the genomic component of the data collected is likely to increase in the next few years as next-generation sequencing technologies evolve and many of these disorders are linked to genetic variants.
The goal of collecting this information is to provide clinical data from many patients accumulated over several years that can be used by researchers seeking to improve current tests and treatments as well as develop better ones, CHOP said.
At present, "newborn screening programs are primarily limited to a short-term focus" and test for "disorders in which early intervention is possible," White noted in a statement.
"If we can broaden the data capture to follow up children over a longer term, we can tap the potential to develop new medical tests and interventions for diseases that are not currently detectable or treatable," he said.
This could include interventions for diseases such as spinal muscular atrophy, a neurological disorder that can be detected at birth although it isn't currently on the list of diseases that are screened in newborns, he said.
Michael Bennett, director of CHOP's metabolic disease laboratory, added that the database could help clear up uncertainty about what happens "after the newborn period" to children who are diagnosed with genetic disorders, since these diseases "don't go away."
Collecting data on these children will allow researchers to "monitor long-term outcomes and develop better strategies to diagnose and treat diseases," he said in a statement.
For example, a centralized database that contains clinical records from several patients would make it easier for scientists to detect patterns in data that would not be obvious in the small number of cases that a single hospital would have access to, CHOP explained.
These data patterns could help researchers determine which treatments would lead to better outcomes or could help them tailor treatments — such as gene therapies, enzyme replacements, or drugs that act on particular metabolic pathways — to work for particular genetic profiles, Bennett said.
Additionally, “there is substantial variability in both the amount and types of information that clinical centers collect about patients, both in their initial visits and for long-term follow-up,” White explained to BioInform.
As such, creating a resource that gathers information on patients and makes it available in a "uniform fashion" would help "alleviate concerns about data quality and differences in modes of collection,” he said.
CHOP isn't ramping up its informatics capabilities significantly in preparation for hosting the resource because it isn't expecting to handle large quantities of data. The repository will only collect information on patients that test positive for genetic disorders, White told BioInform.
The team is using the Research Electronic Data Capture, or RedCap, program to build electronic data capture forms to collect information on patients from participating centers, he said. The hospital is currently testing the system in a pilot project focused on about four disorders.
The pertinent issue for CHOP and NBSTRN is settling on what information to collect, "because we are not always certain what research questions are going to be asked," White said.
"Our challenge is to make sure that we are inclusive enough in collecting data not just on the disorder itself but the ancillary data that may or may not be important for specific research questions, but not to be so inclusive that it precludes anyone from ever contributing data," he said.
In addition, CHOP and NBSTRN are working on setting up best practices for data workflow, standards, and security, as well as creating policies for obtaining patient consent and rules that will govern researchers' access to data in a manner that complies with federal regulations, among other "bioethics issues" that need to be worked through, he said.
"Those practical challenges will require resolution to ensure that the project moves forward," he said.
CHOP intends to collect data on patients until they reach adulthood — 16 or 18 years of age, depending on the state, White said. Individual hospitals that intend to submit data will be required to obtain the necessary consent from their patients in keeping with the dictates of their respective institutions. When the children reach adulthood, they would be required to re-consent to continue in the program, he said.
CHOP has also set up a data governance group that is working on a standard set of data elements that would be required for submissions related to any disorders, which includes things like patients' demographics or lab results. Depending on the type of disorder, data providers would be required to add disease-specific elements, White said.
The hospital is currently working with two state labs to collect data that meets these requirements as part of a pilot run of the system.
CHOP plans to launch the system this summer, pending approval from its institutional review board, with information related to only one disorder, and then it will be expanded to include additional ones later on, he said.
He added that the IRB's approval is crucial because CHOP will need to store the data it collects from all the centers with identifiers — although it will not share these with researchers — in order to be able to capture data "longitudinally."
"The reason that we will do this is because any center contributing data at multiple timepoints for a patient will need to be able to add to an existing record for that patient, and the only convenient way to do this is if we track identifiers for those patients centrally," he explained in an email. "This is especially true if there are multiple centers involved in a patient’s care over time."
Since it will store identifiers locally for patient tracking purposes, CHOP will work to ensure that its system complies with the Federal Information Security Management Act (FISMA) security requirements, White said.
Besides CHOP's database, the NBSTRN is building another resource, called the Virtual Repository of Dried Blood Spots, that will contain a record of blood samples that are collected by participating states at time of birth along with information about what samples were positive for any of the disorders that are screened routinely, White told BioInform.
The virtual repository will provide NBSTRN members with the tools to review and manage sample requests, configure approved orders, track shipments, and view researcher responses to the samples they have received.
It will also include information about the kinds of data that states collect as well as each state's requirements for sharing patient data.
This way, researchers that are interested in studying specific conditions would be able to query the database for states that have samples available for use, White explained.
NBSTRN intends to roll out the blood spot virtual resource sometime next year, he said.
Have topics you'd like to see covered in BioInform? Contact the editor at uthomas [at] genomeweb [.] com.