Evaluation of Pain in Infants of the First 100 Days of Age With Artificial Intelligence Techniques

Pain assessment in infants is difficult due to the lack of verbal communication and the subjective nature of existing methods. Current tools require evaluating multiple indicators separately, which can be time-consuming and variable between observers. Artificial intelligence can enable rapid and standardised direct pain scoring, improving accuracy and clinical efficiency. The limited number of comprehensive studies in this area highlights an important gap in the literature.

This research was conducted to develop an application that will evaluate pain in infants in the first 100 days of life using artificial intelligence techniques.

This study is an artificial intelligence-based research conducted with 1000 newborns hospitalised in a Neonatal Intensive Care Unit. A data collection form and the Neonatal Pain, Agitation, and Sedation Scale were used as data collection tools. Data analysis was performed using IBM SPSS Statistics 21.0, with descriptive statistics and Cohen's Kappa test applied. A p value of < 0.05 was considered statistically significant. Image and audio recordings from all 1000 newborns were independently labelled by the researchers, and these labelled data were used to develop the artificial intelligence model. The labelled image and audio recordings were utilised for model training (80%), validation (10%), and testing (10%).

The mean gestational age of infants was determined to be 38.52 ± 1.07 weeks, and the postnatal age was determined to be 2.90 ± 0.77 days. It was determined that there was no difference between the pain scale scores labelled by the researchers (p > 0.05). It was determined that the success rate of the created artificial intelligence model in correctly predicting the presence of pain in newborns was 82%.

It was determined that the developed artificial intelligence model was successful in predicting the presence of pain in newborns in the Neonatal Intensive Care Unit.