A Growth Artificial Intelligence Algorithm for leNgth and Weight Study

Condition(s) studied

Study summary

This is a data collection and machine learning accuracy testing project that aims to a) collect training data to enhance, by machine learning, an artificial intelligence (AI) algorithm for measuring length in infants and young children and b) test the accuracy of the AI algorithm by comparing the AI predicted length with the gold standard measured length. Images and videos will be collected by care givers and healthcare professionals, together with physical length measurements. These data will be used to train the AI algorithm and to explore potential improvements. Other data to be collected is user experience feedback.

Eligibility

Primary outcome measure(s)

• Accuracy length AI — Date of enrolment, at baseline
  Accuracy of the Length AI vs length gold standard (WHO methodology with length board in cm) assessed using several different parameters: the bias (cm), agreement and reliability measures, mean absolute error (cm), mean absolute percentage error (%), percentiles of the absolute error (cm), and root mean square error (cm).
• Accuracy caregiver length — Date of enrolment, at baseline
  Accuracy of the caregiver measured length (own preferred methodology in cm) vs gold standard measured length (WHO methodology with length board in cm), assessed by same parameters as mentioned in the first primary outcome measure.
• Accuracy caregiver vs AI length — Date of enrolment, at baseline
  Accuracy of the caregiver measurements (self preferred methodology in cm) vs Length AI by same parameters as mentioned in the first primary outcome measure.

Trial sites (4)