News for Agent

Daily business news: funding, M&A, launches

Research & Clinical Trials

Kaiser Permanente and MIT AI Models Predict Three Pregnancy Complications by Week 12

By ·

Key takeaways

  • Kaiser Permanente and MIT researchers trained 49 machine learning models on 442,760 pregnancies from Kaiser's Northern California region between 2011 and 2021.
  • At 12 weeks, the models scored AUC 0.80 for gestational diabetes, 0.79 for gestational hypertension and 0.74 for preeclampsia.
  • Steering the top 20% of flagged patients into gestational diabetes lifestyle counseling was estimated to save about $10.8 million across the study's test group.

Kaiser Permanente and Massachusetts Institute of Technology researchers built AI models that flag patients at higher risk for gestational diabetes, gestational hypertension and preeclampsia as early as 12 weeks into pregnancy, using only routine electronic health record data. The study appeared in Communications Health, Becker's Hospital Review reported on November 20, 2026.

How the Kaiser Permanente Models Were Built

The team trained 49 machine learning models on 442,760 pregnancies cared for by Kaiser Permanente's Northern California region from 2011 to 2021. The models predicted seven complications at six-week intervals. At 12 weeks, they posted AUC scores of 0.80 for gestational diabetes, 0.79 for gestational hypertension and 0.74 for preeclampsia. AUC measures how well a model separates high-risk patients from low-risk ones, and a score of 1.0 is perfect.

Preterm birth, stillbirth, placental abruption and small-for-gestational-age births were harder to predict, with those models scoring 0.62 to 0.68 even at 24 weeks. The early models outperformed existing risk checklists from the National Institutes of Health and the Society for Maternal-Fetal Medicine. The researchers estimated that steering the top 20% of flagged patients into gestational diabetes lifestyle counseling would save about $10.8 million across the study's test group. Aspirin prophylaxis for preeclampsia was also cost-effective, while a cervical-length screening pathway for preterm birth was not.

The models lost accuracy when tested on 6,105 pregnancies from the NIH's All of Us Research Program. Local fine-tuning narrowed the gap but did not close it. Co-senior author Marzyeh Ghassemi, PhD, an associate professor at MIT in Cambridge, Mass., said AI models must be evaluated and adapted "before introducing them into different health care settings." The models also performed worse for Black patients on preeclampsia and gestational hypertension, a gap the researchers said may stem partly from early-pregnancy blood pressure being recorded earlier and more often for white patients.

Kaiser Permanente's Scale Behind the Study

Kaiser Permanente is a nonprofit integrated managed care consortium serving 13.1 million members across eight states plus DC through 55 hospitals, 847 medical offices and more than 25,000 physicians, according to akta.pro company records. The same records put combined operating revenues for Kaiser Permanente and Risant Health at $127.7 billion in 2025, up from $115.8 billion in 2024, with net income of $9.3 billion and operating income of $1.4 billion.

Earlier Kaiser Permanente AI Moves

On November 16, 2026, The Permanente Medical Group launched Permanentech, a physician-led healthcare innovation company focused on AI-powered technologies and care delivery solutions, Becker's Hospital Review reported. On November 17, 2026, Kaiser Permanente mental health clinicians in California ratified a four-year labor contract with 97% support that includes AI guardrails, requiring advance notice of AI-related care changes, Yahoo reported. On November 20, 2026, Becker's Hospital Review reported that Kaiser Permanente's AI strategy prioritizes rebuilding patient trust, with Dr. Chang saying technology should be "augmented" rather than "artificial."

What the Researchers Plan to Test Next

Co-senior author Yeyi Zhu, PhD, a research scientist at the Kaiser Permanente Division of Research, said in a November 19 news release that the findings show "AI is not a one-size-fits-all solution for maternal and neonatal health." The team plans to test whether adding ultrasound measurements and clinical notes improves predictions, and whether using the models in practice improves outcomes for patients.

Companies in this story

Sources

Machine-readable: Markdown · JSON