IJPDS International Journal of Population Data Science

A new study from Brazilian researchers, published in the International Journal of Population Data Science (IJPDS), shows a significant advancement in the use of hospital data by developing an algorithm that connects fragmented patient data.  

Publicly available Brazilian health databases lack unique patient identifiers which, for researchers, is problematic. The new computer program developed by Prof. Kenneth Camargo however, enables such data to be processed by tracking individual patients across multiple hospital visits, refining assessments of case severity.

To test the utility of the computer program, a de-identified national database of over 16 million hospitalisations of women of reproductive age (10-49) between 2017 and 2020 was used to link related hospitalisations for the same patient. The program identified nearly 5 million obstetric hospital episodes from 2018-2019.

The analysis showed that multiple-record patient episodes, all of the services accessed by patients to treat a clinical condition or procedure, were associated with longer hospital stays, higher costs, and a lower chance of being discharged alive.

Crucially, the study found a significant difference in how case severity is measured when considering linked records. Using the new method, a 13.15% increase in recorded hospital deaths compared to separately analysing each hospitalisation, was identified. The computer program also reduced the number of records with unclear discharge reasons from 2.29% to 0.73%, improving data accuracy.

Lead author Professor Claudia Coeli said, “This research demonstrates the value of using computer algorithms to connect fragmented patient data. By accurately tracking hospital episodes, health officials can gain a more precise understanding of patient outcomes and case severity, leading to better healthcare planning and resource allocation.”

 

Click here to read the full article

Professor Kenneth Camargo, Universidade do Estado do Rio de Janeiro, Instituto de Medicina Social, and Professor Claudia Medina Coeli, Universidade Federal do Rio de Janeiro, Instituto de Estudos em Saúde Coletiva, Brazil

Medina Coeli, C., Soares Madeira Domingues, R. M., Meijinhos, L., Medina Coeli Bastos, D., Sobrino Pinheiro, R., Saraceni , V., Bastos Dias, M. A., Santana Paiva, N. and Camargo Jr, K. (2025) “Using a deterministic matching computer routine to identify hospital episodes in a Brazilian de-identified administrative database for the analysis of obstetrics hospitalisations”, International Journal of Population Data Science, 10(1). doi: 10.23889/ijpds.v10i1.2467.