Big Data Analytics for Early Diagnosis of ALS

Date: 07/12/2019

Poster presented at: ISPOR 2017

Location and date: Boston, MA; May 20-24, 2017

Authors: Tara Grabowsky, Oodaye Shukla, Manjula Kasoji, Charlotte Merrill, Wendy Agnese, Nazem Atassi


  • To identify early predictors of ALS by applying big data analytics to a large claims database.


  • Using multiple analytic methods, this study has identified features in ALS patients’ claims histories that differentiate them from the general population before initial ALS diagnosis.
  • ALS patients may present with clinically relevant symptoms suggestive of connective tissue disorders, skin disorders, and nonspecific neurological complaints 5 years before ALS is diagnosed.
  • Medically significant predictors seen in patients who were eventually diagnosed with ALS included, but were not limited to, nervous system disorders, hereditary and degenerative nervous system conditions, connective tissue disease, skin disorders, lower respiratory disease, gastrointestinal disorders, neurologist visits, orthopedic surgeon visits, gastroenterologist visits, nontraumatic joint disorders, otolaryngologist visits, and the use of riluzole prior to diagnosis.
  • Analysis of a cohort that has 5 continuous years of history showed that the frequency of ALS patient features increase over time, which perhaps suggests yet another method for differentiating ALS patients prior to diagnosis.
  • Findings from the national dataset may help increase and improve early screening for ALS in an appropriate patient population and provide evidence supporting the use of this robust methodology in other therapeutic areas.
  • Early diagnosis and ALS management may lead to reduced utilization of healthcare resources and costs; however, future investigations are needed to confirm the heath economics outcomes and benefits associated with early diagnosis of ALS patients.

Big Data Analytics for Early Diagnosis of ALS