Spinal Muscular Atrophy: Creating Structure from Unstructured EHR Data to Study Rare Diseases in Neurology

While there is no cure for Spinal Muscular Atrophy (SMA), several new treatments for this genetic disease have entered the market in the past five years, providing the possibility of a higher quality of life for patients experiencing this degenerative neurological condition. 

SMA is a group of hereditary diseases that progressively destroys nerve cells in the brainstem and spinal cord that control activities such as speaking, walking, breathing, and swallowing. SMA is a rare neuromuscular disease that leads to muscle weakness and atrophy, and is often diagnosed during childhood, sometimes on newborn screening tests and other times when children do not achieve developmental milestones, such as standing and walking at typical ages. SMA may not be diagnosed until adulthood for people who are more mildly affected.

The most common kind of SMA is diagnosed with a blood test that detects abnormalities in the survival motor neuron 1 (SMN1) gene. Because there was no U.S. Food and Drug Administration-approved (FDA) treatment of SMA until recently, the best options for patients experiencing SMA consisted of managing the symptoms, supporting function, and preventing complications.

The outlook for patients with SMA changed in 2016 with the FDA’s approval of Biogen’s Spinraza. The first drug approved to treat children and adults with SMA, Spinraza is administered by injection into the fluid surrounding the spinal cord. That was followed by FDA approval in 2019 of Novartis’ Zolgensma, a gene therapy for children less than 2 years old who have infantile-onset SMA. Additionally, in 2020, the FDA approved Roche’s orally administered drug Evrysdi to treat patients two months and older with SMA.

Unlocking SMA Data Throughout the Patient Journey
While the new therapies offer hope to patients with SMA and the physicians who treat them, for life sciences companies, real-world data can provide another perspective outside of clinical trials in order to provide a holistic view of the disease and treatment of patients. 

Patients being treated by community-based clinicians have often not been accounted for in the medical literature, leaving a gaping hole in our understanding of disease diagnosis and treatment.

Traditionally, gathering real-world data on patients with SMA has been done in referral medical centers through detailed prospective data collection. However, this is resource intensive and excludes patients and families who aren’t able to participate in the data collection due to either time commitment constraints or not receiving care at these specialized centers. Patients being treated by community-based clinicians have often not been accounted for in the medical literature, leaving a gaping hole in our understanding of disease diagnosis and treatment.

Another challenge stems from the fact that some of the most valuable data points for neurological patients are often documented in their electronic health records (EHRs) as unstructured, free-form data that is not easily extracted by conventional analytic methods. An estimated 80% of medical data is unstructured, often containing some of the most critical information to understanding neurologic diseases and treatment outcomes, including drug safety and tolerance.

An estimated 80% of medical data is unstructured, often containing some of the most critical information to understanding neurologic diseases and treatment outcomes, including drug safety and tolerance.

Often, this unstructured data appears in the notes sections of EHRs and includes valuable contextual information, such as insights into clinicians’ thoughts, concerns, and the reasons why genetic testing and/or a certain treatment is or is not pursued—a level of granularity far beyond structured EHR data and claims data. As it relates to SMA, this text-based documentation may include information pertaining to ambulatory status (Is the patient walking on their own? Using crutches? Using a walker?) and respiratory status (Is the patient breathing on their own? Using a respiratory assistance device?) and how their status changes over time. 

Because real-world data on the broad population of patients with SMA (including their treatment outcomes) are difficult to gather, life sciences companies and physicians lack insight across healthcare settings. Some of the current gaps remain in understanding why patients do not pursue genetic testing to confirm their diagnosis, which patients are being treated at community practices, and what barriers for treatment exist.

American Academy of Neurology’s Axon Registry Offers Real-World Data

Verana Health has partnered with the American Academy of Neurology to maintain, curate, and analyze data from the Axon Registry^®, one of the largest real-world clinical data registries for neurology. More than 1,300 neurologists—with nearly 40 different EHR systems—contribute their structured and unstructured EHR data to the Axon Registry, which contains more than 6 years of longitudinal patient records, including 17 million patient encounters involving 3 million patients.

Qdata SMA is the first Neurology Qdata module for life sciences analysis

Verana Health takes Axon Registry data a step further to inform research, transforming it—through the VeraQ^®population health data engine—into high-quality, disease-specific real-world de-identified data modules, known as Qdata^®. Qdata SMA is the first Neurology Qdata module for life sciences analysis, and I’m proud to have been part of the team that helped develop and validate it. 

With detailed de-identified clinical data pulled from patient records, Qdata SMA goes beyond the diagnosis and treatment codes contained in claims data. It helps surface deep information on the genetic basis of the diagnosis, progression of the disease, real-world outcomes, and the decision to use various treatments. With a more holistic view of the patient journey represented in unstructured notes, Verana Health’s Qdata SMA serves as a quality real-world data blueprint to generate real-world evidence for rare disease research.  

For life sciences companies, Qdata SMA can help:

• Identify potential patients for clinical trials within community-based care settings
• Understand the patient journey and the natural history of disease for late onset SMA patients 

By unlocking data across the patient journey, Verana Health brings structure to the unstructured data that is of critical importance to studying rare diseases such as SMA

By unlocking data across the patient journey, Verana Health brings structure to the unstructured data that is of critical importance to studying rare diseases such as SMA. As the field of neuroscience changes with new advances in treating rare genetic diseases, Verana Health and the American Academy of Neurology are uniquely poised to enable research that delivers quality insights from quality data.To learn more about Qdata SMA or to meet with Verana Health at the AAN 2022 Annual Meeting, click here.