INFANT® (INtelligent Fetal AssessmeNT) is software to interpret fetal heart rate monitoring that took us over 27 years to develop.
It has taken this time because correct CTG interpretation is difficult. On the one hand INFANT® must be good at detecting problems within all manner of patterns and on the other hand it must not be over sensitive so as to raise concern unnecessarily. Achieving this balance is very hard, for human and computer.
INFANT® analyses the FHR and contraction pattern and alerts clinicians to potential problems in a timely way. INFANT® has been validated on databases of several thousand interesting, abnormal and challenging traces and has been found to perform at the level of experts whose ability we have measured.
Watch the INFANT® Presentation
How is INFANT® deployed?
INFANT® is a software module that runs within our K2MS Guardian™ Electronic Patient Records computer system for labour and delivery. INFANT® runs in real time at the bedside using a direct feed from the fetal monitor.
The Guardian™ system communicates INFANT® all around the hospital network so that all monitoring can be viewed at central stations, offices or remotely from home to ensure that all CTGs of concern are flagged to senior staff wherever they may be.
INFANT®’s algorithms have been modelled on human expert interpretation and has been measured and assessed in a paper that compared its performance to 17 human experts.
INFANT® analyses the same features of a CTG trace that an expert would do, Baseline, Variability, Accelerations and Decelerations (type and timing, depth, duration and area), Signal Quality, Contractions (location and frequency, and whether they are pushing), Sinusoidal and Bimodal patterns; all the features that clinicians are familiar with.
INFANT® also has algorithms for the detection of maternal heart rate patterns to avoid cases where the maternal heart rate has been inadvertently recorded instead of the fetal rate. This also works for twins.
When the fetal heart rate monitor is switched on, INFANT® begins to identify every feature on the CTG and then builds a picture of what the trace shows.
Every feature adds to a ‘belief’ that a certain classification has been reached; INFANT® analyses these beliefs and begins to draw inferences.
INFANT® is highly dynamic and applies the same approach to its analysis of the features of the CTG that a human expert would do. It shows its current level of belief via an ever changing histogram, one per feature.
INFANT® will draw attention to changing patterns in the CTG.
For Yellow and Red levels of concern, the clinician must acknowledge they have seen and are responding by applying their thumbprint. If this is not done within a few minutes an alert tone will sound. If this is also ignored, or if there is no clinician in the room, then a recorded human voice will issue a warning that will suggest senior clinical staff are summoned.
Our research shows that there is no clinician in a delivery room typically 27% of the time, so using a human voice to raise alert will empower the birth partner to summon help.
INFANT® is not about de-skilling clinicians or devolving CTG interpretation to a machine, quite the opposite.
INFANT® has been designed as a tool to support and inform clinical decision making and leaves the clinician firmly in charge. INFANT® does, however, contain a lot of teaching and educational material in order to provoke understanding and raise standards in CTG interpretation.
INFANT® is a support system to inform, raise awareness and re-enforce education at the point of care to raise skill and avoid human errors.
INFANT® was granted regulatory approval to be released in to clinical practice in 2009, however, K2 is a strong believer in appropriate healthcare evaluation and so the decision was taken to release INFANT® as part of a controlled study, The INFANT Study.
The INFANT Study, the largest maternity study undertaken in the UK, was funded by the Department of Health's Health Technology Assessment Programme (HTA) at a cost of £6m. It ran for 3 years and recruited 47,000 women from 24 hospitals across the UK and Ireland.
More information about The INFANT® Study see the link below (04 - The INFANT® Study - A Multicentre Randomised Controlled Trial of 47,000 Births)