Engineers from Centro Universitário FEI and pediatricians from Unifesp (Federal University of São Paulo) have collaborated to create an AI tool designed to assess the pain levels of newborns in neonatal intensive care units (ICUs). This innovative technology leverages multimodal models that combine both visual and textual data, enabling the interpretation of the facial expressions of infants with greater precision and less subjectivity. This is particularly critical as newborns cannot verbally communicate their discomfort and rely solely on the observations of caregivers.

Ruth Guinsburg, a neonatal pediatrics professor at Unifesp and head of the Neonatal ICU at Hospital São Paulo, highlights the limitations of current pain assessment protocols in neonatal settings, which tend to rely heavily on observational pain scales that can be subjective. The perception of pain can vary widely, depending on the emotional state of the observer—be it a doctor, nurse, or a distressed parent—leading to inconsistencies in care. The newly developed AI tool aims to standardize these observations, minimizing the impact of personal interpretations on crucial medical decisions.

The introduction of such technology in neonatal ICUs could have far-reaching implications for the treatment of the most vulnerable patients. By enabling more accurate assessments of pain, the AI can assist healthcare providers in making better-informed medical choices, ultimately improving the quality of care and potentially enhancing patient outcomes. This advancement reflects a broader trend toward integrating artificial intelligence into healthcare, offering promising new avenues for pediatric care and pain management.