Affordable Care Act premiums rise another 13% as sign-ups open

• FYIMoney Oct 30, 2026
  Affordable Care Act benchmark premiums rise 13% for 2027 after a 26% jump this year, as enrollment opens Nov. 1
  Primary carePractices
  Average benchmark silver premiums on the Affordable Care Act marketplaces are rising 13% for 2027 in finalized rates, 11% in HealthCare.gov states and 16% in state-run marketplaces, KFF reported Oct. 30. Premiums rose 26% for 2026, the year enhanced premium tax credits expired, and people above 400% of the federal poverty level no longer qualify for any credit, according to KFF. Paid enrollment in February 2026 was 19.2 million, down 12% from a year earlier, and 11 insurers had announced exits from at least one state as of Oct. 29, including Cigna from all 11 states where it sells individual plans, KFF said. Open enrollment on HealthCare.gov runs from Nov. 1 to Jan. 15, 2027, the Centers for Medicare & Medicaid Services said.
  Sources: KFF, CMS fact sheet, plan year 2027 marketplace plans and prices
• FYIEvidence Oct 29, 2026
  General-purpose AI models gave seriously wrong drug doses in about 2 of every 100 answers, Doximity analysis reports
  All physicians
  About 19.5 of every 1,000 answers from widely used AI models contained a dosing error rated serious or potentially lethal, according to a Doximity analysis of more than 8,500 clinical dosing questions covering more than 2,200 FDA-approved drugs, Healthcare Dive reported Oct. 29. The most common error was failing to adjust a dose for age, kidney function or drug interactions even when the question supplied that information, and newly approved drugs had five times the odds of an error, the company said. Doximity released the findings with its own dose-verification tool, DoseCheck, which it said was correct 99.9% of the time. The figures are the company's own, Healthcare Dive's report did not name the models tested, and no peer-reviewed results have been published.
  Source: Healthcare Dive
• FYIEvidence Oct 30, 2026
  AI reading 25,000 primary care notes a week found about 7 times as many missed specialist referrals at Singapore clinics
  Primary care
  An AI system that scans consultation notes for referrals that were never made raised weekly missed-referral detection about sevenfold at seven National University Polyclinics clinics in Singapore, according to a quasi-experimental study published Oct. 30 in npj Digital Medicine. The system flagged 428 cases, of which 225 were verified by clinicians and 31 more were reported by patients, with 99.9% specificity and 48.8% sensitivity, the authors reported. In a blinded one-week comparison of 83 flagged cases, its positive predictive value was 61.8% against a 51.8% confirmation rate among seven physician reviewers. The study reported about 41 fewer projected staff-hours and no patient outcomes.
  Source: npj Digital Medicine
• FYIMoney Oct 28, 2026
  Employer coverage of GLP-1 weight-loss drugs falls to 53%, and 30% of employers still paying are weighing dropping it
  Primary careEndocrinology
  The share of U.S. employers covering GLP-1 drugs for weight loss fell to 53% from 57% a year earlier, the first decline in three years, according to WTW's 2026 Best Practices in Healthcare Survey of 471 employers with at least 100 workers, released Oct. 28. Of employers that still cover the drugs, 30% are considering ending coverage, and 91% of those that dropped or are weighing dropping it cited cost, WTW said. Employers project median health costs to rise 9.7% in 2027 after plan changes, the highest rate since the early 2000s, the survey found. "Pharmacy has moved from a line item to the center of the healthcare cost conversation," said Jeff Levin-Scherz, WTW's population health leader for North America.
  Sources: WTW (GlobeNewswire), Medical Daily
• FYIEvidence Oct 30, 2026
  Penn Medicine AI trained on 400,000 CT images screens for hundreds of conditions a scan was not ordered to find
  Radiology
  Penn Medicine researchers built an AI model, Percival, that runs in the background on chest and abdominal CT scans and looks for disease and future risk beyond the reason each scan was ordered, Becker's reported Oct. 30 on a study in npj Digital Medicine. The model was trained on more than 400,000 CT images from nearly 52,000 Penn Medicine BioBank patients and outperformed two other AI imaging models on 279 chest CT conditions and 230 abdomen and pelvis conditions, with the widest margins for lung cancer and acute lymphoid leukemia, according to the report, which gave no accuracy figures. The model is in preclinical testing, its broader findings have not been validated outside Penn Medicine, and developers hope to begin testing in early 2027. "We're hoping to elevate 'opportunistic screening,'" said senior author Walter Witschey, PhD.
  Source: Becker's Hospital Review