Life Sciences Blog

Predicting Recalls Model and a Real-World Example

by | Aug 5, 2021

Recent events are providing the opportunity to test out the concept of ‘predicting recalls’ using publicly available data. Reed Tech has been conducting a long-term project with a ‘Recall Prediction Model’. The model is being updated frequently with new medical device safety and quality data as reported to US FDA. With recent recalls in the news, we can share some interesting results on how the model is performing.

On August 3, 2021, the FDA identified a Class I recall for Philips Respironics V60 and V60 Plus ventilators upgraded to enable High Flow Therapy (software version 3.00 and 3.10). More than 16,500 devices have been recalled, covering devices distributed between May 1, 2009 and June 2, 2021.

With this announcement, the Reed Tech Team reviewed the data for the recall and compared how the Recall Prediction Model presented specific findings for the Philips products. The model accurately predicted the likelihood of this recall. The products recalled were the V60 ventilators.

In our database, we can see that these devices have product IDs of 74890 and 71222:

If we look back at our prediction file from 2020 (Batch 85 = May 28th, 2020),  we can see that the model gave the devices a prediction score of 1 which is the highest risk of having a recall.

In this real-world example, we show the power of predictive analytics, harnessing the depth of regulatory source data available in Reed Tech Navigator™ for Medical Devices.  

Over time, we are seeing amazing results from the Recall Prediction Model. There are possibilities for solving many use cases with the cleansed and normalized source database, Navigator™ for Medical Devices. For hands-on, filterable product search and a unified view of products, Navigator™ places analysis at your fingertips.

 Learn more about Navigator™ or contact us to request a Predictive Recalls Custom Analysis, see a demo or just ask questions:

 [email protected]


Questions? We are here to help. Visit our Knowledge Center or Contact Us.