arrow-leftarrowbookmarkcategorychatcheckclosedots-with-opacitydotsdown-chevronfacebookhero-lefthero-righthvh-logo-whitelinkedinplaceholderrelated-findingssearchtwitter
Customer Story

Using Knowledge Transfer Techniques to Uncover the Clinical Definition and Journey for ​nr-AxSpA​

Industry: Healthcare

The Challenge

  • An ICD-10 diagnosis code does not exist for Non Radiographic Axial Spondyloarthritis (nr-AxSpA)
  • To overcome this, HVH leveraged multiple nationwide databases, including one that contained detailed, qualitative clinical notes to find a nr-AxSpA patient population

Our Approach

  • HVH applied Natural Language Processing (NLP) and Semi-Supervised Labeling against transcribed clinical notes to categorize patients, extract features, and learn machine learning models that can identify nr-AxSpA patients within the axial spondylopathy cohort in EMR data
  • The models learnt were carefully designed to be readily transferable to national level claims data, and were used to identify nr-AxSpA patients within claims data. This helped uncover the real world epidemiology and patient journey associated with the target patient group​

Our Outcome

  • HVH delivered detailed incidence and prevalence of nr-AxSpA
  • Using Machine Learning, HVH delivered an analysis of prominent events and features of diagnosed patients, including timing of their episodes of care along their journeys​
  • HVH also identified characteristics pre-diagnosis of Ankylosing Spondylitis (AS) that are considered predictive of AS​
  • HVH then provided target lists of physicians for these patients

Results

The client used this information to inform targeting, messaging, incidence and prevalence and sales force planning. The information was also used to develop a HEOR study on cost of care differentials.