Medumo is Leveraging Machine Learning to Predict and Prevent No-Show

Healthcare is embracing the tremendous potential of machine learning, and for good reason. The ability to predict individual-level behaviors and events before they occur is changing health and healthcare as we know it. Recently, MIT researchers announced their work on applying machine learning in the ICU to decide on best treatments and to make real-time predictions by using previous data, symptoms and learning from ICU cases. In the chronic disease realm, IBM just joined forces with the JDRF to analyze global type 1 diabetes data to identify factors leading to disease onset. With millions of dollars pouring into machine learning and artificial intelligence, we’re excited to already see, with a comparatively modest amount of data, the power and stability of our own predictive model.

For one of our biggest hospital clients, Medumo saw a 61% drop in the patient no-show rate, which is a major pain point in healthcare. On average, our hospital clients are seeing a 20X Return on Investment for every dollar spent using Medumo CareTours to help patients stay meaningfully engaged throughout their colonoscopy preparation. Our predictive model (accuracy ratio 62%) was consistently able to identify the patients at highest risk of defaulting, defined as late cancellations, no shows or inadequate colonoscopy preparations. Two days before the procedure, our model flags as high risk 75% of the patients who will ultimately default, allowing time for healthcare providers to intervene.

As we grow our users and more data becomes available, we will be able to add strength and stability to our model for a more sophisticated predictive tool to help both patients and providers enormously. Stay tuned and to learn more about Medumo and recent updates, visit or follow us on social: LinkedIn, Twitter.

#Predictivemodel #MachineLearning #Colonocopy #Noshowrate #MIT #Healthcare #Digitalhealth

Recent Posts

See All