Skip to main content
Analysis of Nurse’s Notes to Provide Insights into Racial Inequality and Patient Outcomes in Brazilian Hospitals
SC&I Associate Professor Charles Senteio and his research collaborators based in Brazil seek to use Natural Language Processing (NLP) to examine words and phrases used in medical records to identify biases which can help inform interventions.
SC&I Associate Professor Charles Senteio and his research collaborators based in Brazil seek to use Natural Language Processing (NLP) to examine words and phrases used in medical records to identify biases which can help inform interventions.

A new study seeking to identify and describe the factors which increase the likelihood that racial minority patients experience different treatment and health outcomes than nonracial minority patients across large hospital systems in Brazil is being undertaken by Associate Professor of Library and Information Science Charles Senteio and his co-researchers.

The research, the first study of its kind in the U.S. or Brazil that Senteio and his collaborators are aware of, examines specific word use and their association with the race of the patient, and aims to better understand and address persistent racial inequity of health outcomes and determine the association between nursing notes and patient deterioration and mortality.

“Our long-term goals,” Senteio said, “include informing clinical decision support systems to alert clinicians to the clinical circumstances and decisions which are at risk of racial/ethnic bias, both in the U.S. and in Brazil.”

“Our central hypothesis is that nursing notes will be less frequent for patients from racial/ethnic minority groups and will be qualitatively different in nature (i.e., contain more stigmatizing language) compared to patients from nonminority groups,” Senteio said. “As in the U.S., patients in Brazilian hospitals who are from racial/ethnic minority groups endure worse health outcomes than white patients for a myriad of diseases,” Senteio said.

Senteio’s research collaborators include SC&I Assistant Professor of Library and Information Science Tawfiq Ammari, and two faculty members at the Universidade Federal do Rio de Janeiro (UFRJ) Anna Nery School of Nursing: Adjunct Professor and International Relations Coordinator Priscila Brigolini Porfirio Ferreira Ph.D., and Vice Dean and Full Professor Silvia Teresa Carvalho de Araujo Ph.D.

The project, Senteio said, has three central aims: to use Natural Language Processing (NLP) to determine associations between the frequency of nurses’ notes and patient’s race/ethnicity; identify differences between phrases and words used in nursing notes based on patient’s race/ethnicity; and to evaluate differences in nursing notes based on the experience of nurses working for hospital systems with patient populations that are more racially and ethnically diverse. They will examine nurse’s progress notes because nurses, compared to physicians, provide more frequent contemporaneous notes.

“Prior research in the U.S. has shown that the use of stigmatizing words and phrases are indicators of cognitive biases in healthcare delivery shown to impact quality of care,” Senteio said. “The harmful effects of the use of stigmatizing language (e.g., ‘addict,’ ‘non-compliant,’’ ‘drug seeker’) has been described as perpetuating negative stereotypes and bias across various nursing specialties and it can influence clinicians’ attitudes and beliefs towards patients. This in turn can impact the care delivery, specifically clinical decision making such as prescribing medication.”

Other U.S. studies, Senteio said, have shown that Black and Hispanic patients are 36% and 30% less likely than white patients to receive analgesia for acute pain in Emergency Departments; that patients from racial/ethnic minority groups are more likely to receive care from multiple providers than patients from nonminority groups, pointing to poor continuity of care for minority patients; and Black patients are more often seen by clinicians with less clinical training and typically treated at hospitals with lower quality outcomes.

“Prior research in the U.S. has shown that the use of stigmatizing words and phrases are indicators of cognitive biases in healthcare delivery shown to impact quality of care,” Senteio said.

Comparatively fewer such investigations have been done in Brazil, Senteio said, and in addition, “Brazil has striking racial inequalities which stem from its history of colonization which is rooted in legitimizing racism in political and social ideology.”

The team will begin their research in Brazil with a pilot study, Senteio said, and next steps include conducting a randomized controlled trial to test the efficacy of helping clinicians identify potential areas of subconscious racial/ethnic bias and emphasizing sensitivity to how social factors (e.g., culture, level and type of social support) influence health behaviors which impact outcomes.

“Our long-term goals,” Senteio said, “include informing clinical decision support systems to alert clinicians to the clinical circumstances and decisions which are at risk of racial/ethnic bias, both in the U.S. and in Brazil.”

This project is being supported by funds from a 2022 Global Health Seed Grant from  the Rutgers Global Health Institute.

Pictured, left to right:  Silvia Teresa Carvalho de Araújo, Charles Senteio, and Priscila Brigolini Porfirio Ferreira.

Learn more about the Library and Information Science Department on the Rutgers School of Communication and Information website.

 

Back to top