Autor:innen:
L. Reinhart (Düsseldorf, DE)
A. Bischops (Düsseldorf, DE)
J. Kerth (Düsseldorf, DE)
M. Hagemeister (Düsseldorf, DE)
B. Heinrichs (Jülich, DE)
J. Dukart (Jülich, DE)
S. Eickhoff (Jülich, DE)
E. Mayatepek (Düsseldorf, DE)
T. Meissner (Düsseldorf, DE)
Objectives
Recent advances in Artificial Intelligence (A) have led to its increasing use in child and adolescent healthcare. Early identification of developmental concerns is indispensable for timely intervention, however developmental assessments are resource intensive. AI carries large potential as a valuable tool in early detection of developmental issues. In this systematic review, we aim to synthesize and evaluate the current literature on the use of AI in monitoring child development, including possible outcomes, acceptability, and utilization of such technologies by parents, caregivers and healthcare workers.
Material and methods
The systematic review is based on an extensive literature search comprising the databases PubMed, The Cochrane Library, Scopus, Web of Science, Science Direkt, PsycInfo, ACM and Google Scholar (1996-2022). All articles addressing the use of AI in monitoring child development or describing respective outcomes and opinions were included. Child development was categorized based on the WHO definition of cognitive, physical, language, socioemotional, motor and sensory awareness development. Two authors independently extracted data and selected the studies. Of 2813 identified articles 71 were included.
Results
Currently, only a few well-proven AI applications in monitoring child development are available. Most of them use machine learning methods to detect abnormalities in child development. Overall, the sub-areas of cognitive and language development are particularly well mapped comprising 16 and 18 studies of the included articles, respectively. Besides, 26 articles focused on early detection of autism and included more than one sub-area of child development.
So far, there is only one article that explicitly addresses advantages and disadvantages of AI applications in child development monitoring. The article reports moderate acceptance among the parental population towards AI in pediatric medicine and concerns with quality, privacy, shared decision making, cost, convenience, human element of care and social justice. Many publications show prospects of possible AI applications in preventive pediatric and adolescent medicine without providing concrete evidence.
Conclusion
Numerous studies depict the use of AI in pediatric medicine and monitoring of child development. However, no long-term outcomes are yet established due to the recent introduction of AI. Further studies are necessary to evaluate whether and to what extent the use of AI in pediatric medicine is desired and accepted by parents, children and health care workers.