Work

The role of Open Access Data in democratizing healthcare AI: A pathway to research enhancement, patient well-being and treatment equity in Andalusia, Spain

Year: 2024

Type: article

Abstract: Artificial Intelligence (AI), a transformative technology with vast potential in the field of healthcare, presents an array of opportunities for innovation, with the ability to transform medical care from diagnosis to treatment and patient monitoring [1,2].However, one of the main concerns about AI is the issue of data bias, which refers to the distortion or unfairness that can arise from the data used to train or evaluate AI algorithms.Data bias can affect the accuracy, validity and reliability of algorithms, and can lead to discriminatory or harmful outcomes for certain groups of people [3].Traditionally, data-driven initiatives have primarily focused on building models and optimizing accuracy, often overlooking the fundamental issue of data bias.This oversight has the potential to propagate algorithmic bias, reinforcing stereotypes and structural inequities, particularly when deployed in real-world scenarios.Thus, it is essential to recognize that many of the models published so far may inadvertently perpetuate disparities present in data sources and patient populations [4].In the medical field, AI's most significant contribution lies in illuminating the critical role of data in facilitating more objective, consistent, and immediate decision-making, while promoting health equity [5].However, healthcare data is powered by the medical knowledge system, which includes influential stakeholders like research funders, universities, researchers, and academic journals.These entities play pivotal roles in determining research priorities, defining patient cohorts, providing platforms for research, and disseminating research findings, whose collective influence significantly shapes the trajectory of medical research and our understanding of health and diseases, thus contributing to the persistence of health disparities

Source: PLOS Digital Health

DOI: 10.1371/journal.pdig.0000599