Artificial Intelligence in Healthcare: Systematic Review of Diagnostic and Screening Tools in Brazil's Resource-Limited Settings (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Artificial Intelligence (AI) has the potential to transform global healthcare, with extensive application in Brazil, particularly for diagnosis (D) and screening (S). </sec> <sec> <title>OBJECTIVE</title> This study aimed to conduct a systematic review to understand AI applications in Brazilian healthcare, especially focusing on the resource-constrained environments. </sec> <sec> <title>METHODS</title> A systematic review was performed. The search strategy included the following databases: PubMed, Cochrane Library, Embase, Web of Science, LILACS, and SciELO. The search covered articles from 1993 to November 2023, with 25 articles selected for the final sample. Meta-analysis data were evaluated based on three main metrics: ROC curve, sensitivity, and specificity. A random effects model was applied for each metric to address study variability. </sec> <sec> <title>RESULTS</title> Key specialties for AI tools include ophthalmology and infectious disease, with a significant concentration of studies conducted in São Paulo state (52%). All articles included testing to evaluate and validate the tools; however, only two conducted secondary testing with a different population. In terms of risk of bias, 10 articles (40%) had medium risk, 8 articles (32%) had low risk, and 7 articles (28%) had high risk. Most studies were public initiatives, totaling 17 (68%), while 5 (20%) were private. In developing countries like Brazil, minimum technological requirements for implementing AI in healthcare must be carefully considered due to financial limitations and often insufficient technological infrastructure. Of the articles reviewed, 76% used computers, and 72% required the Windows operating system. The most used AI algorithm was Machine Learning (44%). The combined sensitivity was 0.8113, the combined specificity was 0.7417, and the combined area under the ROC curve (AUC) was 0.8308. </sec> <sec> <title>CONCLUSIONS</title> There is a relative balance in the use of both diagnostic and screening tools, with widespread application across Brazil in varied contexts. The need for secondary testing highlights opportunities for future research. </sec>

Ähnliche Arbeiten