Presented By O’Reilly and Cloudera
Make Data Work
September 11, 2018: Training & Tutorials
September 12–13, 2018: Keynotes & Sessions
New York, NY

Breaking the rules: End-stage renal disease prediction

Olga Cuznetova (Optum), Manna Chang (Optum)
11:20am–12:00pm Wednesday, 09/12/2018
Data science and machine learning
Location: 1A 12/14 Level: Beginner
Secondary topics:  Health and Medicine
Average rating: ***..
(3.33, 3 ratings)

Who is this presentation for?

  • Data scientists, healthcare professionals, clinicians, and business executives

What you'll learn

  • Explore an applied machine learning approach that leverages both supervised and unsupervised learning
  • Learn how to use nonclinical data to predict disease progression


Traditionally, nephrologists use laboratory data to diagnose and identify end-stage renal disease (ESRD) patients. However, identification often occurs too late in the disease’s progression to place the patient on a kidney transplant list to avoid dialysis. Identifying patients earlier along the chronic kidney disease (CKD) progression is essential for better patient outcomes for the 30M Americans living with CKD today. Currently, less than a third of ESRD patients are able to get a timely kidney transplant, a procedure that raises their five-year survival rate from 35% to 80%.

Olga Cuznetova and Manna Chang share an approach for predicting CKD patients with a high likelihood of progressing to ESRD in advance by applying supervised learning, allowing care managers to take measures to slow disease progression and better manage a kidney transplant program. They also share an unsupervised approach that looks at the classification of patients that tend to develop this disease faster than others, focusing on how the two methods work with claims data and how they complement each other. By using a machine learning approach, more CKD patients will be able to get proper treatment in timely manner.

Photo of Olga Cuznetova

Olga Cuznetova


Olga Cuznetova is a data science team lead at Optum Enterprise Analytics, where she guides junior team members on their projects and helps implement data science solutions that address healthcare business needs. Currently, her projects focus mostly on building disease progression and clinical operations models. A few examples include predicting high-cost diabetic patients, predicting the progression tofo end-stage renal disease, implementing a substance abuse disorder model using external client data, and predicting medical prior authorization outcomes. Previously, Olga completed a one-year technology development program with a focus on the development of essential technical skills, healthcare business acumen, and an analytical skill set, which led her to choose a data science career path. Olga holds a BS in finance from Central Connecticut State University. When she has a spare moment, you can find her traveling both in the United States and abroad.

Photo of Manna Chang

Manna Chang


Manna Chang is senior data scientist at Optum Enterprise Analytics, where she plays a leading role in providing and developing innovative technologies and methods to meet customer needs and answer healthcare-related challenges. Her experience includes applying machine learning techniques in drug discovery and genomic outcome studies. Manna holds a PhD in biochemistry and an MS in statistics. She loves sci-fi movies and enjoys hiking.