Related Experiment Video
Updated: Jan 18, 2026
05:35
Author Spotlight: Developing a Point-of-Care Hemoglobin Estimation Method for Anemia Management
Published on: January 19, 2024
1.3K
Diagnostic and Screening AI Tools in Brazil's Resource-Limited Settings: Systematic Review.
Leticia Medeiros Mancini1, Luiz Eduardo Vanderlei Torres1, Jorge Artur P de M Coelho1
1Faculty of Medicine, Universidade Federal de Alagoas, Av. Lourival Melo Mota, S/n - Tabuleiro do Martins, Maceió, 57072-900, Brazil, 558232141461.
JMIR AI
|September 10, 2025
Summary
Artificial intelligence (AI) offers significant potential in Brazilian healthcare, particularly for diagnosis and screening in resource-limited settings. This systematic review found AI tools effective, with machine learning being the most common algorithm.
Area of Science:
- Health Informatics
- Medical Artificial Intelligence
- Public Health Technology
Background:
- Artificial intelligence (AI) presents transformative potential for global healthcare systems.
- Brazil is actively exploring AI applications, especially in diagnostic and screening tools for healthcare.
- Resource-constrained environments within Brazil are a key focus for AI implementation in health.
Purpose of the Study:
- To systematically review AI applications in Brazilian healthcare.
- To specifically analyze AI tool usage in resource-limited Brazilian healthcare settings.
- To evaluate the effectiveness and implementation considerations of AI in Brazil.
Main Methods:
- A systematic review of 714 papers from 1993-2023 across six databases (PubMed, Cochrane, Embase, Web of Science, LILACS, SciELO).
- 25 papers were selected for final analysis, with meta-analysis of area under the receiver operating characteristic curve, sensitivity, and specificity.
- A random effects model was used to address study variability in performance metrics.
Main Results:
- Ophthalmology and infectious disease are key specialties for AI tools, with a concentration of studies in São Paulo (52%).
- Machine learning was the most utilized AI algorithm (44%). Combined sensitivity was 0.81, specificity 0.74, and AUC 0.83 (P<.001).
- Most initiatives were public (68%), often requiring standard computer hardware and the Windows OS, highlighting infrastructure considerations.
Conclusions:
- AI diagnostic and screening tools are balanced and widely applied across Brazil.
- The effectiveness of AI tools is demonstrated, but the need for secondary testing indicates areas for future research.
- Implementation of AI in Brazilian healthcare requires careful consideration of technological infrastructure and financial limitations.
