Prediction of hypotension during haemodialysis through data analytics and machine learning

What is it about?

This study has demonstrated the use of machine learning (ML) techniques in predicting intradialytic hypotension (IDH). The analysis has used only observations that are routinely collected during clinical care, and the key variables among them have been identified. ML has been used to build a predictor that could be useful in preventing IDH events from occurring. Of the five different ML algorithms tested, the Random Forest model had the highest overall predictive accuracy (75.5%), while the Bidirectional Long Short-Term Memory model achieved the highest sensitivity (78.5%). When only pre-dialysis data were used as inputs, the prediction performance decreased but nevertheless remained clinically useful. The inclusion of real-time data obtained during dialysis would be expected to improve the performance of this algorithm back toward that of the original model.

Featured Image

Photo by TopSphere Media on Unsplash

Photo by TopSphere Media on Unsplash

Why is it important?

Patients undergoing haemodialysis at home or in dialysis centres are at risk of IDH, which is one of the most common complications of this treatment. IDH is unpleasant for patients, increases the need for additional costly medical intervention, and is associated with increased mortality. Prevention of IDH would clearly be beneficial to dialysis patients, but until now there has been no reliable way to predict if it will occur during a dialysis session. The proposed prediction model using ML algorithms offers great promise as a tool for pre-emptively identifying patients at risk of IDH, so that preventative measures can be taken to avoid the harm that it causes.

Perspectives