Key result
AI in non-communicable diseases predominantly targets cancer, while widely reported neural networks underperform.
Why the study?
Advances in AI offer potential to leverage electronic medical data and continuous physiological monitoring to detect complications and support care for costly, high-mortality non-communicable diseases.
Systematic Review
This systematic review highlights that while artificial neural networks are the most frequently reported AI systems for non-communicable disease monitoring, they do not necessarily yield the best performance, and cancer remains the most analyzed disease area.
Answers are grounded in this paper and cite their sources.
Prioritize validated top-performing AI over prevalent neural networks for NCD monitoring; challenges assumptions of superiority for most-reported models while confirming cancer dominance.
The wide availability of electronic medical data from multiple sources, such as clinical settings and wearable devices, evidences an important opportunity for the analysis of non-communicable diseases with high mortality rates globally and costly management expenditures, like cancer, diabetes, cardiovascular or chronic respiratory diseases. Advances in artificial intelligence (AI) have a great potential to use the information obtained from the constant monitoring of physiological parameters to create models that can serve as support in the warning of possible health complications, enabling medical professionals to provide anticipated care, thus benefiting national economies through reduction of medical expenses. The present study is a systematic review of different scientific databases taking two time periods, from 1956 to 2021 and from 2011 to 2021, to show the trends in the use of artificial intelligence algorithms and the performances obtained in the monitoring and detection of non-communicable diseases using electronic data. It is found that the analysis of cancer data predominates over that of other non-communicable diseases; also, the best performance of AI techniques was not achieved by artificial neural networks despite being the most reported computing systems in the time period considered.
No takes yet. Share an insight, caveat, or question.
Ramirez-Bautista et al. (2023) conducted a systematic review in Non-communicable diseases. Artificial intelligence algorithms was evaluated on Trends in the use of artificial intelligence algorithms and performances obtained in the monitoring and detection of non-communicable diseases. A systematic review of AI in non-communicable diseases found that cancer data analysis predominates, and artificial neural networks did not achieve the best performance despite being the most reported.
Synapse has enriched 10 related papers. Consider them for comparative context: