HVH Presents Big Data Analysis on Early Diagnosis of ALS at ISPOR 2017
HVH Precision Analytics presented findings of its big data analysis for earlier diagnosis of ALS (amyotrophic lateral sclerosis) to help clinicians treat patients sooner and allow patients to possibly enroll in clinical trials.
HVH’s poster presentation, titled “Big Data Analytics for Early Diagnosis of Amyotrophic Lateral Sclerosis (ALS),” took place at ISPOR 2017 — the 22nd Annual International Meeting of the International Society for Pharmacoeconomics and Outcomes Research — held May 20-24 in Boston.
The average time from the onset of symptoms to an ALS diagnosis is about one year. This results in a detrimental delay in the start of treatment, potentially precluding patients from enrolling in clinical trials.
For the big data analysis, HVH used its three-stage methodology to collect, aggregate, and assess the journeys of a large U.S. patient claims database, called the Truven MarketScan Database, of more than 170 million people.
A total of 12,332 people with ALS (2010-2014) were identified in the dataset with an average of 4.4 years of claim history. The average age of patients was 60, and 58 percent were male. About 25 percent had a prescription claim for Rilutek (riluzole).
Patients from California, Florida, Massachusetts, and New York were analyzed for the first time for features including their diagnosis and procedure codes, medications, standard provider types, and types of standard care facilities. These features were then ranked by frequency and classified according to a mutual information method.
The goal was to find medically significant predictors of ALS in patients who were later diagnosed with the disease. Connective tissue and skin diseases, lower respiratory diseases, and gastrointestinal disorders were found to be among the most frequent features of the analysis.
Findings revealed that certain signals in patients’ claims histories seemed to differentiate themselves from other patients up to five years before their initial ALS diagnosis.
“Tremendous progress has been made in the field of data analytics and predictive modeling to help better understand undiagnosed patients,” Tara Grabowsky, MD, chief medical officer of HVH and author of the published findings, said in a press release.
“This is an example of how the intensive analysis capabilities of the HVH platform can potentially identify early signals in ALS that could ultimately be used to develop earlier diagnostic tools. Given the rapid and debilitating nature of this horrible disease, this is an important step forward.”