Healthcare Brew covers pharmaceutical developments, health startups, the latest tech, and how it impacts hospitals and providers to keep administrators and providers informed.

Healthcare Brew

Quinn Sental

Healthcare Brew delivers the latest news and trends in healthcare.

healthcare-brew.com

An image of a red cross, pill bottles, and health carts surrounding a phone. 

Digital tools may ease inequities in maternal healthcare, report finds

It’ll all boil down to equity-centered design, according to the Deloitte Center for Health Solutions.

Ransomware 101: The need-to-knows for healthcare execs

The Change Healthcare cyberattack may be just the beginning.

Amanda Eisenberg

HCA Healthcare pilots Google AI to aid handoff documentation

The large language model processes a 12-hour shift and condenses the highlights into a report that helps the next nurse pick up where the previous left off.

A computer processor icon denoting healthcare technology

Tech

Cassie McGrath

In 2017, Hurricane Maria devastated Puerto Rico.

The storm led to a collapse of the island’s healthcare system, a 

 reported, following power outages and a lack of disaster management plans that contributed to 3,000+ deaths. 

 in 2017 that Puerto Ricans living on the island already had low access to healthcare before the hurricane hit. Years after the storm, the region faced ongoing staffing shortages and issues with federal funding that restricted care, 

That’s why, in October 2017, Emmanuel Oquendo and Israel Figueroa founded BrainHi, a health tech company with an AI receptionist called Lara, which completes administrative tasks for providers. The tool can answer missed calls and schedule appointments, and is designed to reduce clinicians’ workloads. Some of the company’s clients include NeoMed Center, Ashford Hospital, Manatí Medical Center, and Mennonite Healthcare System, all based in Puerto Rico. 

Oquendo spoke with Healthcare Brew about his goals for the company going forward.

This interview has been edited for length and clarity.

Do you solely serve Puerto Rico at the moment? Are you looking to expand?

When we launched, we were very focused on that critical moment in Puerto Rico [after Hurricane Maria], and how we can help our island and providers.

 face a unique challenge, just like Puerto Rico, where the staff is often overwhelmed by daily demand, and they’re not able to serve all the patients. So what we found is that Puerto Rico is not the only place in the US that may find these sorts of issues and may need automation. Puerto Rico remains to be the core of our company, but we work with health centers in Florida and in other places in other parts of the US.

You introduced AI for administrative tasks before its current rise in popularity. What have you learned from this experience?

We’ve been working nearly a decade on this.

We’re not only focused on the AI component but we offer a full digital front-door platform that serves as a host for all the communication that is happening, and it facilitates and streamlines a collaboration between AI and the human side. That’s what’s given us a great advantage as we expand now and accelerate growth because now that we are gaining a ton of traction and there’s 

, we’re ready for this. We’re ready—not only with the technology, but we deeply understand the workflows and the integrations that are required to really 

What separates AI receptionist Lara and BrainHi from your competitors?

In addition to the comprehensive platform, we stand out in a crowded place of 

 because we’re looking beyond transactional interactions. Our focus is on building a trusted, long-term relationship between the AI and patients because we believe that the next frontier for AI is to create human connection at scale.

Let’s say that Lara has a call with a patient. As soon as we get that first call, we start creating a profile on that patient that goes far beyond what is documented in the electronic health record.

What we found is that getting all these personal notes about the patient and understanding what they need allows us to personalize that relationship much more. We still believe that AI is not there yet in terms of being completely human, if you will. But by being very diligent on understanding the patient’s needs, understanding their preferences, actively listening to the patient, and taking notes on the conversation, we’ve seen that it greatly impacts engagement.

Emmanuel Oquendo, founder of AI administrative company BrainHi

Meet Emmanuel Oquendo, founder of AI administration company BrainHi 

This founder made an AI admin tool in response to Hurricane Maria’s impact on Puerto Rico. 

Last June, a patient in the UK died amid a ransomware attack on National Health Service systems at London hospitals.

The hackers targeted blood services and surgeries, the 

 this past June 25, and the patient reportedly died after waiting a long time for a blood test result.

Additionally, according to the BBC, 10,000 appointments were canceled and there were 600 related “incidents” that led to 170 patient care issues, resulting in various levels of harm.

Qilin, a Russian criminal cyber organization, is reportedly responsible for the hack.

This news comes as IT consultancy Omega Systems published its 2025 Healthcare IT Landscape 

, which found 52% of 250 healthcare executives believe a fatal cybersecurity incident is “inevitable” in a US healthcare facility over the next five years.

Additionally, about 20% of the healthcare leaders reported patient care has already been negatively affected by cyberattacks, and 80% reported they were targeted by a cyberattack in the past year.

Mark Dollar, chief executive of Synnovis, the tech company that manages patient data at the London hospitals, told the BBC the company was “deeply saddened to hear that last year’s criminal cyberattack has been identified as one of the contributing factors that led to this patient’s death.”

Healthcare symbols over a red binary background

National Health Service reports fatal cybersecurity attack in London

A recent study found many executives believe a fatal cyberattack in a US healthcare facility is “inevitable” over the next five years. 

Welcome back to AI 411, a monthly roundup of artificial intelligence (AI) announcements from across the healthcare industry.

This month, the FDA announced the launch of a large language model (LLM) AI tool named Elsa. In a 

 on its website, the FDA said that Elsa is designed to “summarize adverse events to support safety profile assessments, perform faster label comparisons, and generate code to help develop databases for nonclinical applications.” It marks the agency’s first step in its AI journey.

But that wasn’t the only news this June. Here’s your roundup of health-related AI updates from around the industry.

 On June 11, the AMA adopted a new policy calling for “explainable clinical AI tools” that give users clear information about the safety and efficacy of the product. In other words, the AMA is asking AI tool companies to provide clear explanations of how their tools work so clinicians can make the best decisions when using them.

Cigna announced on June 12 the release of several new digital tools, including an AI virtual assistant that can chat with patients about insurance and care options on the company’s app. Soon, the company said in a release, the tool will include new provider matching technology that connects patients with clinicians based on insurance and needs.

Digital musculoskeletal clinic Hinge Health announced on June 17 the release of HingeSelect, a provider network powered by AI and designed to connect patients to in-person care at low prices.

Software company LeanTaas announced its new AI-powered tool iQueue for Surgical Clinics on June 10, which will use AI modules to help coordinate surgeries from the administrative side to the operating room, including case-building, patient outreach, insurance authorizations, scheduling.

AI assistant company Nabla, known for its ambient scribe, announced a $70 million Series C funding round on June 17. Fundraising was led by HV Capital, with additional investment by venture capital firms Highland Europe, DST Global, Cathay Innovation, and Tony Fadell's Build Collective. The company has so far raised $120 million, and new money will go toward product development and expansion.

On June 2, the Massachusetts Institute of Technology and biotech Recursion announced the release of Boltz-2, an open source AI that helps pharmaceuticals design small molecule drugs by predicting if they will work effectively on patients.

The University of Texas MD Anderson Cancer Center and HealthEx. 

MD Anderson and tech company HealthEx announced on June 9 a strategic collaboration to develop AI tools that can support the patient consent process and make it easier for patients to control how their data is used.

Mental health company Wysa announced on June 11 the launch of Wysa Gateway, a chatbot that uses AI to facilitate conversations between therapy providers and health plans.

Healthcare Brew monthly series on AI Startups

AI 411: June 2025

Your monthly roundup of AI announcements is here.

illustrated headshot of caroline catherman, healthcare brew reporter

Caroline Catherman

Are you curious about all the ways that multimillion-dollar medtech can go wrong?

If so, you’re in the right place!Welcome to Recall Roundup, a noncomprehensive roundup of recent FDA Class I recalls—the most severe and potentially life-threatening kind. Recalls range from correcting an issue to removing a device from the market altogether.

Blood products administration set recall.

 Fresenius Kabi has pulled an incorrectly assembled line of its Ivenix LVP Blood Products Administration Set. If used, patients could be infused with unfiltered blood.

Cook Medical wants customers to return certain Cook Beacon Tip 5.0 Fr angiographic catheters because the tip is at risk of detaching.

Medline told customers to stop using and destroy certain lots of neonatal and pediatric convenience kits that contain recalled Smiths Medical endotracheal tubes. The company said the tubes are too small and could cause underventilation.

Some Dexcom G6, G7, One, and One+ glucose receivers with certain serial numbers may have receivers that weren’t correctly assembled and should be returned.

Zyno Medical asked customers to stop using certain lots of its Z800WF Infusion System because “unreleased software versions” were installed “without verification or validation.”

Medtronic is asking customers to replace certain models of its Newport HT70 series ventilators and control boards. They have defects that can cause the ventilator to shut down during use or otherwise malfunction.

assortment of various medical devices around a teal healthcare symbol

Recall roundup: FDA medical device recalls during June 2025

Pulled devices include ventilators and infusion systems.

AI regulations in healthcare are, well, sort of nonexistent at the moment.

 from regulating AI for the next 10 years, though that was scrapped from the June 6 

 for a new proposal that restricts broadband funding in states that regulate AI.

, and according to the Deloitte Center for Health Solutions, 75% of top healthcare companies are experimenting with generative AI. The research organization also reported that 82% of healthcare organizations plan to “implement governance and oversight structures for generative AI.”

That’s where the public-private Coalition for Health AI, or CHAI, comes in. Launched in December 2021, the group of 3,000 providers, tech companies, and other healthcare organizations—including Mayo Clinic, Cleveland Clinic, and Mount Sinai—are working together to create standards for AI in healthcare.

 for responsible AI use in healthcare, released last June, which companies can reference when building out their models, including principles of trustworthy AI, system design, and monitoring.

“It started with a group of eight of us in the private sector coming together, agreeing that this is an effort building best practice frameworks,” Brian Anderson, CEO at CHAI, told Healthcare Brew. The group hopes to “provide the kind of guardrails and guidelines that build trust” across the spectrum, from payers to providers to patients.

One of the core tenants of CHAI’s guidelines is that “transparency is critical,” Anderson said.

The company has what it calls a model card, which Anderson described as a “nutrition label” with information like metric goals, warnings, and testing types so providers and health companies can see what sorts of AI practices others use. For instance, if the Mayo Clinic had a new AI note-taking tool it was using, the model card would outline how the health system uses it, metrics to show performance (such as its safety and usefulness), and potential risks, biases as well as ethical considerations to keep in mind.

Fairness is also a key tenant of CHAI, Anderson said, to ensure training models are “highly effective on people” across the country, from rural states to urban areas to tribal nations.

CHAI’s focus is on the tech side, Anderson said, and it doesn’t advocate for any specific external regulations or federal policies.Though he said many members are most interested in regulations that don’t stifle innovation or create an “uncoordinated patchwork set of state regulations,” he added.

“We want our public-sector colleagues to see CHAI as a resource to educate them on the principles of responsible AI and how the private sector is adopting, implementing, and moving forward in that space,” he said.

On June 11, CHAI announced a new partnership with the Joint Commission, a nonprofit that also sets standards for healthcare and accredits hospitals. For instance, the organization doles out a Gold Seal of Approval to healthcare organizations that hit certain quality and safety measures.

Together, the two organizations hope to “accelerate the development and adoption of AI best practices and guidance across the US healthcare system” by creating AI rules and tools as well as a new certification program “rooted in the Joint Commission’s platform for evidence-based standards and CHAI’s consensus-based best practices for health AI,” according to a release.

That guidance is coming in the fall, followed by the accreditation, the release said.

Another priority for the organization this year, Anderson said, is to connect with smaller providers, like 

, to “make an intentional, significant effort to build playbooks for those community clinics that don’t have the kinds of resources that an [academic medical center] has.”

The organization also announced Kyu Rhee, president and CEO of the National Association of Community Health Centers, as a new board member in February as a part of these efforts.

What are the benefits of CHAI for companies?

Emmanuel Oquendo, CEO of software company BrainHi, became a member of CHAI in November.

BrainHi, based in San Juan, Puerto Rico, helps automate phone calls and other operations for medical practices that need 

Oquendo joined CHAI, he said, because one of the organization’s principles is to ensure AI algorithms and models “are fine-tuned for specific communities’ needs.”

“There are some algorithms that might work great for a Caucasian community but not necessarily for 

He was also inspired to join because CHAI provides an opportunity to learn and network with other healthcare experts about using AI “responsibly and securely,” he said.

Oquendo is now part of a CHAI working group developing guidelines on deploying AI chatbots in healthcare, and said CHAI provides him with valuable guidance to bring back to his own company’s best practices. “I felt that I would get very important information by being part of the community that I could share with my customers,” he said.

AI lacks regulations in healthcare. The Coalition for Health AI set some guidelines

AI use is growing in healthcare. But regulations aren’t. 

AI lacks regulations in healthcare. This group set some guidelines

 (AI) has made its way into healthcare, from 

 are already happening, and there’s no doubt that AI is starting to change how care is delivered.

Already 75% of big healthcare companies are “experimenting with or planning to scale generative AI across the enterprise,” according to Deloitte Center for Health Solutions.

But less discussed are the potential downsides of this new technology. Just as the implementation of electronic medical records 

—like easier collaboration across departments, but more time at the computer—it’s likely AI will see a similar varied trajectory.

“There’s vast amounts of marketing hype [around AI], compared to the actual clinical proof,” James Barlow, a professor at Imperial College Business School in London, told Healthcare Brew. “How do you as a company or developer actually build up an evidence base for the claims you’re making?”

We spoke with experts about the potential challenges AI presents to healthcare, and how the industry can best guard itself against them.

To build models, most AI tech is trained using existing data, whether that's

published research or a patient’s chart. That means, in order to work well, the tools must have access to accurate and high-quality information. But experts are concerned whether this is happening.

 in general, but in healthcare, it could make the quality of care worse, especially for higher-risk patients, Barlow said. “What base are you training the AI on? Are the datasets representative? If they’re not, they might produce biased results.”

Harish Yalamanchili, a surgeon at vascular clinic Coastal Vein Care in California

said understanding what sources data is trained on is “one of the most important things.”

“How careful are you in generating and making sure that your sources are carefully vetted?” he said, adding that “we’re running out of data to build our AI.”

 publishers and other online sources are preventing their information from being used to build AI. In other words, AI needs existing information, but many sources have already been used or are getting restricted.)

There’s concern this could lead companies to “loosen the restrictions a little bit,” on the quality of data used to build platforms, Yalamanchili said.

“AI for me should be a tool that makes your ability to treat people easier,” Yalamanchili said. “The outcomes have to be equivalent to the outcomes without AI, or better.”

Though, he added, he also worries these tools could make it easier to miss errors introduced into a patient’s chart. “If you have a corrupt data point anywhere…how do you know what the reality was before this?” he said.

Bad data or errors could create mistrust between clinicians and patients, Barlow added. There’s also 

In terms of building up guardrails around these challenges, Barlow said a lot of it comes down to trust.

“[Building trust] requires being very open and transparent about the data that’s being used, the quality of the data, how you’re designing the algorithms that use that data, whether you’re using advanced datasets,” he said.

It’s also important to follow any future laws that may emerge, he said. There are currently no “absolute” regulations for AI in healthcare, Yalamanchili added, though organizations like the American Medical Association and Coalition for Health AI have 

Yalamanchili said including more clinicians in leadership could also help, as they know what would best support staff and patients.

“The way we’re going about it, and just throwing all these darts at the dartboard, and hoping that one of them hits bull’s-eye,” he said. “We’re talking about actual human lives,” said.

Experts share how to protect against the downsides of AI in healthcare 

There’s a lot of hype about AI in healthcare, experts say, but being transparent is essential.

There’s a cybersecurity problem breathing down the healthcare industry’s neck. 

On April 7, Baxter Healthcare pulled all 4,100+ Welch Allyn Life2000 ventilators due to severe cybersecurity concerns. The FDA labeled this a Class I 

, meaning these issues threatened patients’ lives, though the agency hasn’t reported any injuries or deaths.

Naomi Schwartz, a former FDA employee and VP of services for MedCrypt, a cybersecurity firm for medical device manufacturers, told Healthcare Brew this recall should be a lesson for the medical technology industry.

This is one of several device recalls in recent years prompted by 

, and in her opinion, Baxter did the right thing fast. The global medtech company first flagged these weaknesses in November, noting that there hadn’t been any hacks up to that point. An April market removal is a relatively short turnaround, she added.

But the vulnerabilities were easily avoidable with a development framework known as 

, the idea that companies—not consumers—are responsible for cybersecurity, and products should have features like multi-factor authentication.

On a scale of “you have to have a PhD” to “a teenager” could hack the system, Baxter’s ventilator security was more toward the latter, Schwartz said. Let’s run through some of the issues:

 For one, the ventilators didn’t encrypt sensitive information, like passwords, according to Baxter’s November 

. “If I’m just issuing all my data in plain text, that’d be like me sending you an email saying, ‘Hey, my front door is unlocked. Walk into my house, why don’t you?’” Schwartz said.

 All a hacker needed to do to access the device was walk into a hospital and plug a piece of hardware into a physical port on the ventilator, Schwartz explained.

 The software used to test and calibrate the ventilators didn’t require authentication from the user either, so anyone could have tweaked the ventilator settings. This flaw was severe enough to earn a score of 10/10

on a scale used by the government to measure vulnerabilities—“a nightmare scenario,” Schwartz said.

These issues should prompt other companies to double-check their own ventilators’ security—especially 

 that were made many years ago, Schwartz said. 

“These are very common problems, and they’re all things that a secure-by-design set of practices would have prevented,” she said.

The good news is Schwartz thinks Baxter responded quickly and appropriately once it discovered these flaws. 

“The people out there who are producing and selling these products are doing their due diligence. They’re going back and checking older systems. They’re making sure that things are good and secure, and when they’re not, they’re taking appropriate action,” she said.

This incident comes after the FDA and Congress have 

 medical device cybersecurity requirements in recent years. 

For instance, in March 2023, the Protecting and Transforming Cyber Healthcare Act started requiring medical device manufacturers to address cybersecurity requirements in their submissions for market approval.