Advanced search
Start date
Betweenand

System for predicting foot complications in diabetic patients using matching learning

Grant number: 20/09430-1
Support type:Research Grants - Innovative Research in Small Business - PIPE
Duration: April 01, 2021 - December 31, 2021
Field of knowledge:Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
Principal researcher:Leissi Margarita Castañeda Leon
Grantee:Leissi Margarita Castañeda Leon
Company:BIOO Artificial Intelligence em Tecnologia da Informação Ltda
CNAE: Desenvolvimento e licenciamento de programas de computador customizáveis
Tratamento de dados, provedores de serviços de aplicação e serviços de hospedagem na internet
City: São Paulo
Principal researchers:Bruno Sergio Ferreira Massa
Assoc. researchers: Fernando Antonios Maman ; Sergio Henrique Garbe Orestes
Associated scholarship(s):21/05089-6 - System for predicting foot complications in diabetic patients using matching learning, BP.TT
21/04457-1 - System for predicting foot complications in diabetic patients using matching learning, BP.TT
21/03980-2 - System for predicting foot complications in diabetic patients using matching learning, BP.PIPE

Abstract

In the world there are already 500M of people with diabetes and 30% of them will suffer from a physiological disaster called "diabetic foot", which generates $ 350B of cost dollars a year. The mortality rate can be compared to more aggressive cancers. Foot complications are present in about 19 to 34% of diabetic patients. It is estimated that 9 to 26 million people worldwide are diagnosed with diabetic foot. Mortality in this patient group is 5 times higher than in diabetic patients without foot complications. Therefore, identifying patients who are at risk of developing foot complications and taking early action is essential for the good evolution of cases. The prediction of complications in patients with chronic diseases has been the subject of studies involving the application of concepts of artificial intelligence and machine learning. The results are very promising with an accuracy of 85.3% in predicting the readmission of patients within 30 days after discharge and 89% in predicting the development of diabetes for example. Thus, this project deals with research for the creation of algorithms that allow predicting the appearance of foot complications as lesions in diabetic patients, exploring machine learning based on retrospective structured (exams) and unstructured (free text) data from the electronic medical record (clinical consultation tracking software used by doctors). The research will produce a proof of concept with data from the Hospital da Clínicas of USP and has already been approved by the Scientific Committee of the Instituto de Ortopedia (protocol number 1408). In parallel, a partnership with InovaHC is undergoing contractual analysis to apply the algorithm and incubate the company in this ecosystem with the objective of reducing the occurrence of "diabetic foot" in the HC complex by up to 85%. The proof of concept represents the first step towards the future development of a system for predicting foot complications in diabetic patients, acting in clinics, hospitals and public and private health systems. In addition to the public health benefits, the future system has great economic potential. For example, treating patients with diabetic foot represents an expense of $ 8.78 billion annually in the United States. This is the market niche that the company intends to operate. Thus, complications of the diabetic foot are responsible for great impact on the individual and society. Technologies like the proposed aim to improve the identification of patients at risk, enabling more assertive and individualized actions with better efficiency, decreasing the total cost and improving the application of health resources. (AU)