Algorithm may help identify Fabry disease based on insurance data
Model could lead to earlier diagnosis, timely treatment, study suggests
An international team of scientists has developed a novel algorithm that may help identify people with Fabry disease based on insurance claims data.
The model showed particular promise for detecting potential Fabry patients who are young and/or female, according to the study “Unveiling the untreated: development of a database algorithm to identify potential Fabry disease patients in Germany,” which was published in the Orphanet Journal of Rare Diseases. The work was funded by Sanofi, which is developing a potential Fabry treatment called Venglustat.
Diagnosing Fabry disease can be difficult, particularly since the disorder can manifest quite differently from person to person. It’s not uncommon for people who are experiencing Fabry symptoms to go years or even decades before they get the correct Fabry disease diagnosis and start treatment.
To help facilitate faster and more accurate diagnoses, a team of scientists at Sanofi and other institutions conducted an analysis of insurance data from a healthcare database in Germany, covering more than 5 million people followed from 2010 to 2017.
Fabry patients typically experience long journey to diagnosis, treatment
“Patients with FD [Fabry disease] experience long odysseys in their diagnostic pathway from symptom presentation to diagnosis and eventual treatment,” the researchers wrote. “This model provides useful insights to physicians on the identification of potential FD patients using available clinical information.”
From the database, the scientists identified 46 people who had a definitive Fabry disease diagnosis and were on treatment for the disorder. They also identified a control group of 460 individuals who did not have the disease.
The team then constructed a mathematical model that basically looked for patterns of insurance claims which were more frequently seen in the Fabry patients than in the controls. Based on the available data, the model accurately labeled about 4 out of 5 patients in either group.
Using this model, the researchers identified 288 patients diagnosed with other, more broader conditions who didn’t have a recorded Fabry disease diagnosis, but were judged to have a high likelihood of potentially having Fabry.
The researchers noted that, compared with the group of patients with definitive Fabry, this group of potential Fabry patients tended to be much younger. More than half of the potential Fabry patients were younger than 30, whereas the vast majority of patients with confirmed Fabry were older than 30. This supports the notion that the algorithm may help identify Fabry disease earlier, the researchers wrote.
Algorithm may ‘facilitate timely diagnosis and early treatment’
“These findings support the relevance and clinical value of the algorithm, which may facilitate timely diagnosis and early treatment of patients with FD,” the team wrote.
The potential Fabry group also had markedly more female patients than the definitive Fabry group. Since Fabry is caused by mutations in a gene located on the X chromosome (one of the two sex-determining chromosomes), it tends to be more severe, and therefore more obvious, in male patients. Consequently, female patients are more likely to go undiagnosed. These data further support the algorithm in helping to identify Fabry patients who would otherwise be overlooked, the researchers wrote.
A notable limitation of this study is that, since they were working with anonymous insurance data, the researchers weren’t able to follow up with any of the potential Fabry patients to see if they actually did have the disease. The scientists stressed more studies will be needed to validate the utility of the algorithm.
About the Author
