Table 1

PICO/PECO research question development

PopulationInfants experiencing pain.
The intervention/exposure will be pain assessment using automatic facial recognition ML algorithms.
ControlThe study control will be indicator-based pain assessment gold-standard (eg, pain scale, pain score and category).
OutcomePrimary outcome:
Model accuracy by predicted assessment measures type:
  1. Numeric score: mean SE or equivalence;

  2. Categorical pain degree (yes/no; no/moderate/severe):

    Concordance statistic (AUC ROC) or equivalence.

Secondary outcomes:
  1. Generalisability;

  2. Interpretability;

  3. Computational efficiency and related costs.

  • Definition of infants: infants will be defined as young children no more than 12 months, including newborn or neonate, full term, premature and postmature infants.

  • AUC ROC, area under the curve for receiver operating characteristic curve.