Machine Learning Based Hypotension Prediction

Hypotension during surgery and in the intensive care units are associated with increased rates of complications such as acute kidney injury and myocardial infarction, and the risk of serious complications increases with the duration and depth of hypotension.  

Advance warning that hypotension is imminent could facilitate diagnostic and therapeutic measures to lessen the clinical impact of hypotension.  The prodromal stage of hemodynamic instability is characterized by subtle and complex changes in different physiologic variables. These changes reflect altered compensatory mechanisms and result in unique dynamic signatures in arterial waveforms that could be used to predict hypotension. 

Here we describe a machine learning based algorithm, the Hypotension Prediction Index, to predict hypotension before it occurs using high fidelity arterial pressure waveform recordings.  The algorithm was developed to observe subtle signs that could predict the onset of hypotension in surgical and intensive care unit patients, and the performance of the algorithm was validated in various unique data sets.

以機器學習預測低血壓

手術期間和在重症加護病房中，發生低血壓與併發症的發生率增加有關，例如急性腎臟損傷和心肌梗塞，而其風險隨著低血壓持續時間和嚴重度而增加。