November, 21-22nd
Auditorium MalrauxCampus Manufacture des Tabacs, Lyon
November, 25-27th
Amphithéatre BuffonParis

FAPESP Week France


Back to news   |   02/12/2019 17:50

Tool for big data analysis could help in decision-making in the medical field

Group at USP are developing computer systems to treat and extract information from large volumes of data derived from public hospitals; the aim is to create a database doctors can consult to support diagnoses and prescribed treatments

By Heitor Shimizu, from Paris  |  Agência FAPESP – Sophisticated computer systems, capable of storing, indexing, analyzing, and making sense of large datasets that cannot be processed by traditional software could become essential tools for supporting decision-making in the medical field.

Research focused on that aim is being conducted by the Databases and Images Group (GBDI) at the Institute of Mathematical and Computing Sciences of the University of São Paulo (ICMC-USP) in São Carlos. The topic was addressed by Professor Agma Traina (photo), in a lecture given at FAPESP Week France.

“One of the main challenges in the Computing Science field is to integrate, organize, and harness large volumes of multimodal data from different platforms in order to drive decision-making processes; that is, to enable the use of data from various sources, such as patient tests, monitoring, and treatment, to collect information on similar cases and build a better understanding about a particular case,” said Traina.

The studies conducted at the GBDI laboratory handle large quantities of complex data, derived from public hospitals in the State of São Paulo. The group mainly works with images and videos capable of providing medical information about similar cases treated in the past.

“When a specialist analyzes, for example, a patient’s chest X-ray, they might remember having seen a similar result in the past, but they’re unlikely to know when or where it was and with which patient. But, if they can instantly search a database for similar cases, tests, results, and treatments indicated in the past, they’ll be able to make decisions with less effort and more confidently,” she told Agência FAPESP.

Part of the research is supported by FAPESP through a Thematic Project coordinated by Traina. She says that the project involves organizing databases, metric access methods (employed to speed up evaluating similar consultations), and processing and visualizing images, which enable tools, algorithms, and methods to be offered to specialist doctors for combining and accessing highly valuable information on old and current cases.

“For that we need to bring together professionals in machine learning, databases, data lineage [concerning the origin of data], image visualization and processing. In our group, we have computer scientists, doctors, mathematicians, and other researchers who work together in order to solve the proposed problems,” said the researcher, who is also a member of the Coordination for the Area of Science and Computer Engineering of FAPESP.

Traina highlights that the size and complexity of databases of electronic patient records present considerable processing challenges, both in terms of developing and applying analysis and knowledge extraction techniques and in supporting the development of practical tools for clinical use.

“However, they also involve infinite opportunities for creating algorithms and methods capable of displaying relevant information related to a particular patient or group of patients, which would usually be hidden by the large volume of data,” she said.

“Moreover, efficient manipulation of that data helps make electronic patient records a more useful platform for supporting health professionals, handling fast demand medical applications, as well as strategic government decisions on health,” she said.

The FAPESP Week France symposium was held between November 21st and 27th, thanks to a partnership between FAPESP and the universities of Lyon and Paris, both in France. 

Photo credit: Heitor Shimizu / Agência FAPESP