HSS - Hospital for Special Surgery

06/14/2025 | Press release | Distributed by Public on 06/14/2025 06:53

HSS Study at EULAR 2025 Congress Uses an AI Model to Predict Readmissions of Pregnant Women with Lupus Based on Social Determinants of Health

Hospital for Special Surgery (HSS) presented a research study at the European Alliance of Associations for Rheumatology (EULAR) Annual Meeting showing that an artificial intelligence (AI)-based model can predict readmissions of pregnant women with lupus by looking at patients' social factors and clinical comorbidities.

Systemic lupus erythematosus (SLE), commonly known as lupus, is a chronic autoimmune disease in which the immune system attacks the individual's own healthy tissue, causing pain, inflammation, and eventually damage to various organs. It commonly affects young women, and pregnancy is a particularly vulnerable period for this high-risk population. Although lupus-related medical complications affect pregnancy outcomes, social determinants and economic conditions, known as social determinants of health (SDOH), may also be important in shaping maternal health outcomes.

"Pregnant women with lupus have five times higher maternal mortality compared to those without lupus," said Sandhya Shri Kannayiram, MBBS, MD, rheumatology fellow at HSS and principal author of the study. "There is little data on how social factors, including Income, Insurance, housing stability, access to transportation, availability of utilities, and literacy, along with clinical comorbidities, affect pregnancy outcomes in individuals with lupus."

To better understand how SDOH impact readmissions for pregnant patients with SLE, the team used ten-year data (from 2011 to 2021) from the National Readmissions Database of the US, encompassing approximately 66,000 hospitalizations in the United States, and counted the number of readmissions within 30 days of discharge during pregnancy and delivery. "We found that approximately 2,500 [pregnant women with lupus] were readmitted to the hospital within 30 days," said Dr. Kannayiram.

"We used a supervised 'glass box' machine learning model called the Explainable Boosting Machine (EBM), which helped predict who is likely to return to the hospital within 30 days," said Dr. Kannayiram. She explained that EBM is a form of machine learning with a type of AI that learns from data to make predictions. "It's an AI model that can be trained on the dataset to predict outcomes with transparency and accuracy compared to traditional statistical models," she added.

The team specifically examined the major social determinants from the database that may influence readmissions during pregnancy in SLE, including Income and insurance status (public, private, or self-pay). Dr. Kannayiram explained that the results obtained showed patients living in lower-income neighborhoods were nearly twice as likely to be readmitted compared to those from wealthier areas, and those with Medicaid or Medicare insurance were more likely to be readmitted within 30 days. Patients discharged to locations other than home, such as rehabilitation facilities or nursing homes, were also more likely to return. However, they also found unexpected correlations. "We found that the size of hospitals was highly related to readmissions, rather than the location, such as cities or rural areas," said Dr. Kannayiram, with large hospitals being the ones with more readmissions. Additionally, unlike most pregnancy studies, where older mothers are usually at higher risk, this study found that younger women with lupus were more likely to be readmitted. Dr. Kannayiram said that this result may be related to severe lupus in younger individuals since about half of the readmitted patients had Medicare coverage, which may suggest a high prevalence of disability or kidney disease within that younger population.

By using the predictive AI model, the team also discovered that income and insurance status were among the top five predictors of 30-day readmission during pregnancy and delivery-related hospitalizations, explained Bella Mehta, MBBS, MS, MD, rheumatologist at HSS and lead author of the study. She noted that these results could inform targeted interventions, such as policy actions to advocate for improved Medicaid and Medicare coverage, as well as post-discharge care coordination. "Integrating SDOH screening into prenatal care for SLE patients, and incorporating a multidisciplinary team of social workers, could reduce preventable readmissions and improve maternal outcomes," added Dr. Mehta.

According to Dr. Mehta, in future studies, the team plans to analyze in detail how individual-level social factors, such as housing instability, food insecurity, education, mental health, transportation access, and patient demographics, interact with clinical comorbidities to determine pregnancy outcomes for women with SLE. "This study underscores the vital intersection between clinical care and structural inequality," said Dr. Mehta. "The lupus community and broader maternal health initiatives must not only consider disease management but also address the social factors that influence outcomes."

Poster details

Title: Predictors of 30-Day Readmissions in Pregnant Patients with Systemic Lupus Erythematosus: The Role of Social Determinants of Health
Authors: Sandhya Shri Kannayiram, Yiyuan Wu, Lisa Sammaritano, Michael Lockshin, Rich Caruna, D. Ware Branch, Jane E. Salmon, Bella Mehta.
Abstract n°: 2325
Presentation: June 14th, 2025. 10:00 AM CEST

HSS - Hospital for Special Surgery published this content on June 14, 2025, and is solely responsible for the information contained herein. Distributed via Public Technologies (PUBT), unedited and unaltered, on June 14, 2025 at 12:53 UTC. If you believe the information included in the content is inaccurate or outdated and requires editing or removal, please contact us at support@pubt.io