Healthcare leaders from points around the globe are pretty much “all in” on generative AI. A full 85% of 2,800 surveyed across 14 countries say they’re investing in the technology now or planning to do so within three years.

That’s one finding from Philips’s Future Health Index report for 2024. The Amsterdam-based health-tech giant commissioned the market research firm GemSeek to survey and interview a worldwide sampling of healthcare leaders. The researchers drew responses from around 200 participants each in Australia, Brazil, China, India, Indonesia, Italy, Japan, the Netherlands, Poland, Saudi Arabia, Singapore, South Africa, the United Kingdom and the United States.

The resulting Philips report, released June 18, devotes a section to data and AI.

“Healthcare organizations have a wealth of data but a poverty of insights,” the authors comment. If not bridged, they suggest, this “insights gap” threatens to thwart efforts to provide evidence-based care equal to the present moment in healthcare history.

Some of the report’s key takeaways are contained in the following five passages.

1. Staff lose precious time pulling patient data together.

Effective patient care relies on accurate and timely access to data, but healthcare professionals often face significant challenges in accessing disparate data and then integrating it into a cohesive patient story.

The vast majority of healthcare leaders (94%) say their organization experiences data integration challenges that impact its ability to provide timely, high-quality care. As a result, almost 4 in 10 healthcare leaders (38%) say that staff lose precious time pulling patient data together, leaving less time to care for patients.

They also see increased operational costs due to data inefficiencies and limited coordination between care providers or departments, as well as unnecessary repeat scans and increased risk of errors.

2. Healthcare leaders see potential to improve patient care through data.

Healthcare leaders see a wide range of opportunities to improve patient care by bringing data from disparate sources together in a meaningful way. For example, they believe data-driven insights could help optimize treatment plans and care pathways (43%), identify evidence-based practices (37%), and reduce waiting lists for diagnostic and elective procedures (36%).

But to deliver on these possibilities, healthcare leaders recognize they first need to get the foundations right. When asked what needs to change in how healthcare data is handled, they highlight the importance of improved accuracy of patient data (40%), interoperability among different platforms and healthcare settings (39%), and data security and privacy (38%).

All these items represent long-standing challenges in healthcare that must be effectively addressed to harness the full potential of data-driven insights for better patient care. Echoing last year’s Future Health Index findings, this is where many healthcare leaders see an important role for external partnerships.

3. Healthcare leaders are implementing AI from the hospital to the home.

This year’s Future Health Index findings show how healthcare leaders have already implemented AI for clinical decision support across different areas of the hospital, with in-hospital patient monitoring, medication management, treatment planning, radiology and preventative care leading the pack.

As healthcare leaders increasingly focus on expanding care beyond hospital walls, implementing AI in remote patient monitoring is an area of focus for the next three years.

4. Generative AI adoption in healthcare set to rise within the next three years.

Generative AI has caught the attention of healthcare leaders in the past year, since its rapid emergence into the public domain. They recognize the benefits that generative AI could bring to patient care by unlocking new efficiencies and insights from patient data.

Our research shows that 85% of healthcare leaders across the surveyed countries are already investing (29%) or plan to invest (56%) in generative AI within the next three years. However:

There are significant cross-country differences in how quickly healthcare leaders plan to invest in generative AI, which are consistent with overall differences in speed of adoption of AI for clinical decision support.

5. Responsible use of AI requires appropriate safeguards.

While there is widespread excitement about the possibilities of AI in healthcare, there is also a shared recognition that it needs to be implemented in a responsible way to avoid unintended consequences.

Almost 9 in 10 healthcare leaders (87%) are concerned about the possibility of data bias in AI applications widening existing disparities in health outcomes. To address this risk, healthcare leaders say it is important to make AI more transparent and interpretable for healthcare professionals and offer continuous training and education in AI.

Another strategy that healthcare leaders pointed to is policies for the ethical use of data and AI. This can only be achieved through cross-sector collaboration and coalition-building.

Philips has posted the report in full for free.

Freeing Clinicians from the Documentation Burden with Nabla - Curious about real-world data from ambient AI implementation? Check out this NEJM Catalyst article to see how Nabla has reduced documentation burden for one of the largest medical groups in the US, fostering more personal and effective patient interactions.

Buzzworthy developments of the past few days.

• Platform thinking plus sector-wide collaboration. That’s the one-two punch healthcare needs to solve its most pressing problems—things like “too much disease, increasing demand and uneven quality.” So asserted at least one speaker at PlatforMed 2024, held in Phoenix earlier this month. Unsurprisingly, subject matter experts from Mayo Clinic, home of Mayo Clinic Platform, were on hand in some number. “Platforms represent our greatest opportunity to bring real, lasting and transformative change to our global healthcare system,” Gianrico Farrugia, MD, president and CEO of Mayo Clinic, told attendees, according to Mayo’s own coverage. Elsewhere Mayo has defined platform thinking in healthcare as a way to digitally support the “aggregation and harmonization of disparate clinical and non-clinical data” so as to “cure more people, connect people and data to create new knowledge, and transform healthcare by developing new, vetted end-to-end algorithms and other innovative solutions to improve care.” You don’t have to know what’s involved in that challenge to be glad the very capable people at Mayo Clinic Platform are up for it.
   
• A nonprofit policy think tank in D.C. has issued calls for two governing bodies to get more involved with AI. The think tank, the Federation of American Scientists, wants the National Institute of Standards and Technology (NIST) to lead an interagency coalition to produce standards that enable third-party research and development on healthcare data. And it wants Congress to authorize the establishment of a NIST Foundation to financially support the agency’s AI mandate. The latter, the scientists’ federation says, would “increase trust in AI technologies and lead to greater uptake of AI across various sectors.” Don’t overlook healthcare among those sectors, FAS.
   
• Consider Mark Cuban an enthusiastic adopter of Gemini for Google Workspace. And not just because that’s what Google wants you to do. In giving the endorsement on a Google blog, Cuban says his Cost Plus Drugs migrated from Microsoft to Google for collaboration and productivity support because “There are two kinds of companies—those that are great at AI and everybody else.” And if Cost Plus Drugs is going to be anything, “great at AI” is it. To be sure, the blog post is promotional. That doesn’t mean it’s not worth a read.
   
• Are Apple and Meta in talks on an AI partnership? Maybe. Maybe not. Either way, the prospect has the business press falling all over itself to get a scoop. The Wall Street Journal: “Exclusive: Apple, Meta have discussed an AI partnership.” Bloomberg: “Apple spurned idea of iPhone AI partnership with Meta months ago.” Reuters: “Apple, Meta not in talks for AI partnership.” Try to keep up.
   
• To predict the earnings potential of any given AI company, look at the innovation quotient of its key people. The U.K. research firm Zeki does something very close to that. Its latest analysis assesses the 10-year talent levels of AI-savvy engineers, scientists and researchers at more than 40,000 companies across 94 countries. Zeki is primarily interested in investment angles, separating likely money makers from probable also-rans. The new report zeroes in on four strongly positioned AI startups in healthcare—Caspar AI, Clear Guide Medical, ThinkSono and Aural Analytics. It’s downloadable here in exchange for contact info.
   
• Teachers like AI for lightening their workloads and, in the process, leaving more time to tutor struggling students. Participants in a pilot program in Indiana said as much when surveyed. Most also said the technology had a positive influence on student learning. But students beware: Teachers are no easier to fool with technology than they were when cheating tactics ranged from low-tech (answers inked onto palms and forearms) to no-tech (furtively glancing at a smart kid’s paper). In one Hoosier State school system, first-time offenders using generative AI to complete an assignment could lose credit for the assignment and have their parents notified. Do it again and you could be kicked out of the class and slapped with an F. Still, an educator tells a local news outlet, “We cannot block generative AI or other AI tools from students.” Along with the temptations, the teacher points out, the technology can bring “many potentials and benefits.”
   
• When ChatGPT-4o went up against Claude 3.5 Sonnet, a clear winner emerged. Tom’s Guide officiated the duel. Who won between OpenAI’s best-trained chatbot and its strapping challenger from Anthropic? Take a guess, then get the play-by-play along with the final results.
   
• GenAI can do a lot of things, but it sure can’t spin comedy gold. Some funny-boned researchers recently found this out the painful way when they put large language models on the spot, improv-like. Evidently the punch lines generated more groans than giggles, as MIT Technology Review recounts. Sample prompt: “Write a joke about pickpocketing.” Response: “I decided to switch careers and become a pickpocket after watching a magic show. Little did I know, the only thing disappearing would be my reputation!”
   
• Recent research roundup:
   
  • Case Western Reserve: Researchers awarded $2.78M federal grant to improve rectal cancer treatment with artificial intelligence
     
  • Michigan State University: Transforming drug discovery with AI
     
  • Nanyang Technological University: Gastroenterologists generally trust AI tools, study finds
     
  • Trinity College Dublin: Scientists devise algorithm to engineer improved enzymes
     
• Funding news of note:
   
  • Prior authorization platform Humata Health closes $25M investment
     
• From AIin.Healthcare’s news partners:
   
  • Cardiovascular Business: AI-powered platform for arrhythmia detection gains FDA approval
     
  • Health Imaging: Supplemental AI education during residency linked to key benefits
     
  • Health Exec: Medical staff must have a say in what AI is adopted for patient care, says AMA