Back to Bigger Picture

BIGGER PICTURE Better care

A warning before blood pressure falls, with a plan for what to do

An operating-room monitor normally shows what is happening now. The HYPE trial tested whether a learned warning from the arterial pressure waveform could help anaesthesia teams act before prolonged low pressure developed.

AI-generated illustration: A pressure-wave pattern offers an early warning before a drop.
A pressure-wave pattern offers an early warning before a drop.AI-generated illustration of the idea, not a photograph of the actual people, equipment or event.
Trial
68 randomized; 60 completed
Median low-pressure time
8.0 versus 32.7 minutes
Setting
Elective noncardiac surgery in Amsterdam

How the idea works

  1. Sample arterial pulses

    An arterial line supplies the pressure waveform.

  2. Predict a pressure drop

    Learned waveform patterns trigger an early warning.

  3. Investigate and respond

    The team follows diagnostic guidance and a treatment protocol.

More information than a single pressure number

An arterial line provides a continuous pressure waveform. The Hypotension Prediction Index looks for patterns associated with an approaching pressure drop, updating its warning value every twenty seconds in this study. The purpose is to create a window for assessment before the usual threshold is crossed.

Sources: [1][2]

Prediction was paired with a clinical response

The intervention did not merely display a score. Anaesthesia staff also had diagnostic guidance and a treatment protocol to investigate the cause and respond. Both groups aimed to maintain mean arterial pressure at least 65 millimetres of mercury; the difference was anticipatory information and its associated workflow.

Sources: [2]

Fewer minutes below the threshold

Of 68 randomized patients, sixty completed the trial. Median time with low blood pressure was eight minutes with the warning-and-response system versus 32.7 minutes with standard care. The primary measure, which combined the depth and duration of hypotension, also improved.

Sources: [1]

The tested unit was the whole response loop

This preliminary, unblinded study came from one centre and selected operations. It did not establish reduced postoperative organ injury or mortality. Its useful design lesson is narrower and concrete: a forecast can change a physiological outcome when it arrives with a workable route to action.

Sources: [1]

The original sources

  1. JAMA / University of Twente repositoryThe HYPE Randomized Clinical Trial: machine-learning warning for intraoperative hypotension
  2. TrialsThe use of a machine-learning algorithm that predicts hypotension during surgery in combination with personalized treatment guidance: study protocol for a randomized clinical trial

What to explore next

Can a draft note give a doctor some attention back?Seeing hunger grow before the next full survey
Explore all 99 applications