|
3, 2, 1: Health AI Brief
Every Friday
July 10, 2026
|
|
AI is reshaping healthcare fast. Below are 3 key AI developments, 2 studies, and 1 takeaway for this week to help you better lead with AI. Target read time: 5 minutes. |
|
3
Market Signals
Kaiser Permanente call‑center nurses told CalMatters that software tracks their call lengths and an AI model flags likely daily "underperformance," with some nurses reporting scrutiny for calls that run past 15 minutes. Kaiser piloted AI grading of nurse empathy and tone in 2024 (the pilot ended that November) and says it does not use handle time to assess performance. The California Nurses Association, representing 25,000 nurses, opened contract talks this month with AI oversight as a central demand, while state lawmakers advance bills protecting nurses who override automated recommendations. My Take
Most of the AI workforce debate is about whether jobs disappear. This story starts one step earlier: AI is already changing what the job is, and the frontline team of nurses is pushing back before any jobs are lost. Incredible Health's 2026 State of Nursing report, released July 7 from a survey of 2,240 US nurses, found 44% now use AI at work, up from 15% a year ago. Nearly half of those users say the technology has saved them little or no time, 46% received no AI training in the past 12 months, and only 8% say their employer has clearly communicated an AI strategy. My Take
Despite the 3x jump in use, nearly half of nurses see little value. The likely culprits: almost half got no training, and, more striking, just 8% were in on the AI strategy. The CMS Innovation Center's ACCESS model went live July 5: a 10‑year, voluntary Original Medicare model that pays fixed annual amounts for tech‑supported management of chronic conditions across 4 tracks: early and advanced cardio‑kidney‑metabolic conditions, chronic musculoskeletal pain, and depression or anxiety. Half the payment arrives in quarterly installments; the other half is reconciled against clinical outcomes like blood pressure control and patient‑reported pain improvement. My Take
After 60+ years of paying for activity, Medicare is now holding back half the payment until outcomes land. Vendors carry the risk, and this 10‑year saga is just getting started. |
|
2
Research Studies
A 16‑month evaluation across Spain's Quirónsalud hospital network tracked 2.34 million scribe‑assisted encounters by 7,147 clinicians, one of the largest real‑world evaluations of ambient AI documentation in European outpatient care. Monthly adoption grew from 2.7% to 31% of outpatient visits while the scribe's semantic agreement with the conversation held near 89%. Consultations ran just 0.37 minutes longer on average, and clinician experience improved in 5 of 7 measured domains, including administrative burden and stress. My Take
The scribed visits didn't get shorter, but the clinicians got less burned out. That matters, and it's an important endpoint in its own right (I'm biased). A pre‑registered randomized trial assigned 977 university students to a conversational AI intervention, integrative group therapy, or a waitlist for 12 weeks. The AI arm outperformed both comparators on generalized anxiety, well‑being, and life satisfaction, with gains holding at a 3‑month follow‑up; PTSD symptoms didn't differ across groups. Participants reporting high loneliness engaged roughly twice as much with the AI, and that engagement fully explained the anxiety improvement. My Take
The biggest gains from conversational AI went to the lonely and those with low social attachments. This is important since these groups are often the toughest to reach by group interventions. One author is from the AI company itself, which makes this trial's pre‑registration an even more important signal. |
|
1
Key Insight
AI needs to make friends with our nurses
America's largest clinical workforce is struggling with AI. Nurse AI use tripled to 44% in a year, yet nearly half of users report no time saved and only 8% hear a clear AI strategy from their employer. The same week, Kaiser call‑center nurses opened contract talks with AI oversight as a central demand, moving AI governance into a union contract. The missing variable in both stories is voice. Elsevier's global clinician survey found 41% of nurses say their views are rarely or never reflected in AI decision‑making, and only 19% of physicians see that exclusion at all. Inclusion converts a deployed tool into a working one, along with training: nurses who got training were 50% more likely to save over an hour a day than those who didn't. Takeaway
AI alone doesn't fix healthcare, and definitely not without nurses. Inclusion and training are the visible fixes; mutual trust is the bigger one: nurses trusting their leadership with AI and leadership trusting nurses as stewards of AI in their own workflows. |
|
Know someone who'd find this useful? Share |
AI needs to make friends with our nurses
3 signals, 2 studies, 1 insight for health leaders
Jul 10, 2026
•
3 min read