EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

1817. Data envelopment analysis with machine learning in healthcare efficiency: a prematurity study

Invited abstract in session TD-28: Advancements of OR-analytics in statistics, machine learning and data science 7, stream Advancements of OR-analytics in statistics, machine learning and data science.

Tuesday, 14:30-16:00
Room: 065 (building: 208)

Authors (first author is the speaker)

1. Fernando Marins
Production, UNESP - São Paulo State University
2. Elias Carlos Aguirre Rodríguez
São Paulo State University
3. Elen Yanina Aguirre Rodríguez
São Paulo State University
4. Diego Eduardo Quagliato Scarelli Cava
São Paulo State University
5. Luiz Fernando Costa Nascimento
São Paulo State University
6. Aneirson Francisco da Silva
São Paulo State University
7. Anderson Rodrigo de Queiroz
North Carolina State University

Abstract

Prematurity (birth before the 37th week of gestation) is a leading cause of neonatal mortality and heightened complications in newborns. In Brazil, the annual cost of premature births is approximately R$8 billion, with newborns spending an average of 51 days in intensive care, totalling a yearly cost exceeding R$15 billion, so improving the health system for mothers and premature births is essential. Data Envelopment Analysis (DEA) is a vital tool to evaluate healthcare system efficiency across territories, but measuring efficiency over time requires re-implementing DEA models. Consequently, Machine Learning (ML) emerges as a viable solution for predicting efficiency scores through supervised learning. Therefore, this study integrates DEA, specifically the Constant Returns to Scale (CRS) model with input orientation, with ML techniques to develop a predictive model for healthcare system efficiency based on prematurity and incorporating indicators such as the number of physicians, beds, healthcare establishments, and per capita health expenditure (R$) in Brazil's microregions. The DEA analysis assesses each microregion's readiness to care for premature newborns, providing inverted efficiency indicators where Decision-Making Units (DMUs) with values near or equal to one suggest lower performance. This approach enables determining efficiency for new DMUs through ML, based on DEA indicators and results, while identifying sectors needing improvement in the healthcare system.

Keywords

Status: accepted

Back to the list of papers