Videos,

Ashis Kumar Das Shares Insight on ADVI’s Recent Study Analyzing Early Onset Colorectal Cancer

Recently, ADVI Health’s strategic analytics, value and economics (SAVEs) solution team conducted a study analyzing early onset colorectal cancer (EOCRC) in younger demographics in the US. The team worked to identify early risk predictors of EOCRC through machine learning models.

Hear more on the study’s findings from Ashis Kumar Das, PhD, MD, MPH, director, SAVEs:

ADVI Health’s Ashis Kumar Das breaks down the SAVEs team’s recent study on early onset colorectal cancer (EOCRC) in younger demographics.

Panelists

Ashis Kumar Das, PhD, MD, MPH

Director, Strategic Analytics, Value and Economics (SAVEs)

Ashis serves as a director of ADVI’s strategic analytics, value and economics, (SAVEs) solution team, where he leverages real-world evidence (RWE) expertise to help ADVI’s clients generate compelling evidence for their products.

Get in Touch

Full Transcript

Ashis Kumar Das: Colorectal cancer is the second leading cause of cancer related mortality in the US. Although its incidence is stabilizing, the incidence of early onset colorectal cancer, defined as persons developing colorectal cancer younger than 50 years, is increasing. EOCRC diagnosis is often delayed, and current screening processes lead to many false positives and are invasive. Previous studies predicting EOCRC risk with machine learning models, have primarily used electronic health records data involving smaller centers or small geographic areas.

In this study, we try to identify the predictors of EOCRC using machine learning and a nationally representative American data sample. We used the National Health interview survey from 2019 to 2023. We applied several machine learning models to see which one is able to identify the predictors in a precise manner. We did several adjustments as well, which are called hyper parametric tuning. After doing the adjustments, we found the balanced random forest model with balanced bagging classifier, was the top performer. The top five predictors were hypertension, smoking, age, hyperlipidemia, and gender.

This study demonstrates the potential of using machine learning techniques to predict EOCRC risks, using publicly available data in large American populations. Early identification of patients at risk for developing EOCRC could facilitate improved screening methods and address the current challenge of delayed diagnosis and treatment.

If you want to learn more about this study, please get in touch with the ADVI SAVEs team.