Challenge

How do you build patient segments in a largely underdiagnosed population?

 

Download

 

See how one brand team used a novel way to quantify and understand meaningful patient segmentation.

The NASH market is at a critical point in its history. Multiple companies are neck and neck in the race to bring the first approved treatment to market, with others working to fill additional unmet needs such as less invasive diagnostics. Our client developing a novel treatment for NASH came to us with one of the biggest challenges facing this market: “How do you build patient segments in a largely under-diagnosed population?”

They wanted to create granular segments for their patient profiling and planning, but were struggling even to size the market overall, due to a variety of roadblocks:

  1. Diagnosis is difficult due to lack of symptoms and condition awareness
  2. Many patients forego invasive liver biopsies, the primary method of reliable diagnosis
  3. Multiple causes and comorbidities

 

Solution

Machine learning and integrated data shines light on true NASH population and meaningful segments

 

Phase 1
Disease understanding and method design

DRG's data science team recommended a machine language algorithm to expand the NASH population beyond diagnosed patients,
learning through key predictors of the disease (i.e., co-morbidities, test results), before they created and prioritized
meaningful segments.

The client team worked with DRG's disease experts (NASH, diabetes, and obesity) and data scientists to reach a common understanding
of the progression to NASH and key characteristics of potential patients.

Phase 2
Machine learning modeling
  • DRG's data science team matched EHR records and claims data to build a profile of known NASH patients.
  • The team used machine learning modeling to answer the question: "Which test results and vitals inform the diagnosis of NASH?"
    The algorithm flagged potentially undiagnosed NASH patients based on key characteristics identified in the machine learning exercise
    (i.e., biopsy, elevated liver function tests, comorbidities, obesity levels).
  • Through this process, the team identified a much larger population of undiagnosed patients likely to have NASH.
  • Then, the data science team was able to create meaningful sub-segments based on test results and vitals measurements (i.e., "diabetic seniors") for analysis and comparison.
Phase 3
Sub-segment delivery and recommendations
  • DRG provided the launch team with 12 meaningful, prioritized sub-segments and recommendations for planning and strategy.

 

Results

As a result of the analysis, the client was well-positioned to:

  • Improve forecasting accuracy and prioritize investments
  • Employ better-targeted customer engagement efforts based on unique characteristics of subsegments
  • Shift strategy, as a result of real-world data upending previous assumptions

 

DRG Differentiators

  • Largest overlapping claims and EHR dataset in the industry
  • Commercial and therapeutic experts with deep market knowledge, ideas and answers from the project outset – limited time needed to “get up to speed”

 

Get a more complete view of your patients

When you are faced with tough business questions, DRG analytics specialists are here to help you define the best path forward.

Visit: decisionresourcesgroup.com/analytics

Contact: questions@teamdrg.com

Read our latest executive briefing: Patient Segmentation Strategies for Today's Personalized Markets

Synergy Between AI and Digital Platforms for PBMs; Driving Value-Based Programs

View Now