Healthcare

Technology Overrides All When Lives Are at Risk

Hospital physicians and CIOs are apt to implement new technologies before regulators approve new processes because ... it’s all about saving lives.

By Anthony Birritteri, Editor-in-Chief On Jun 6, 2025

In this installment of New Jersey Business Magazine’s continuing Business Roundtable Series, conducted in conjunction with the New Jersey Business & Industry Association, healthcare professionals discuss how new technology is being implemented, especially in today’s AI-influenced world.

Meet the Panelists:

Andy Anderson

As executive vice president, chief medical and quality officer of RWJBarnabas Health, Andy Anderson, MD, MBA, oversees 38,000 employees, including 9,000 physicians, who treat an area covering more than five million people. He is tasked with ensuring top-tier clinical outcomes and highest quality care for RWJBH patients and staff alike. Dr. Anderson joined RWJBH in 2018 as president and CEO of the Combined Medical Group of RWJBH and Rutgers Health.

Sunil Dadlani

Sunil Dadlani is executive vice president, chief information and digital transformation officer at Atlantic Health System. He joined the healthcare system in September 2020 and today oversees information technology for the entire network. Prior to joining Atlantic Health, Dadlani was the CIO for the New York State Department of Health. He successfully led the largest public-sector digital business technology transformation, delivering digital healthcare services to 19 million New Yorkers.

Sai Kandamangalam

Sai Kandamangalam is executive vice president/chief information officer at Holy Name Medical Center. He oversees all IT functions, including data infrastructure, software development and integration, creative web services/design, telecommunications, and the help desk. He joined the hospital after serving as senior vice president of software engineering at PharmaPoint; director of software development for SourceMedical Solutions; and tech lead/senior software engineer for Cerner Corporation.

David Johnson

David Johson is senior vice president and chief innovations and information officer at Inspira Health. He previously served as vice president innovation and operations improvement at Inspira (2005 – 2021); and manager of operations improvement (2004 – 2006), and manager of rapid development team (2003 – 2004), Temple University Health System. He earned his undergraduate degree from Rutgers University and his advanced graduate degree from St. Joseph University School of Business.

Tarun Kapoor

Tarun Kapoor, MD, MBA, is senior vice president and chief digital transformation officer at Virtua Health. He oversees Virtua’s Digital Transformation Office and orchestrates Virtua’s enterprise-wide master plan in support of an intuitive care journey for all consumers. Dr. Kapoor was president of VirtuaPhysicianPartners™ and senior vice president and chief medical officer for Virtua Medical Group (VMG). Dr. Kapoor joined VMG in 2008, where he was the associate director of the Virtua Hospitalist Group.

Jordan Tannenbaum

Jordan Tannenbaum, MD, MBA, MPH, PgC Clinical Informatics, is chief information officer and chief medical information officer at Saint Peter’s Healthcare System. He supervises all aspects of computing infrastructure, cybersecurity, and applications. As CMIO, he implements and optimizes EMRs, supervises upgrades, and ensures regulatory and quality compliance with EMR-related regulations. Dr. Tannenbaum practiced pediatrics for 30 years while also heading clinical informatics projects.

What have you observed about the pace of technological change in healthcare today?

Johnson: With the exception of the last five years [due to COVID], it’s been very slow. We have been lagging in some basic solutions, and are playing catch up [with] other industries, especially on the consumer/patient-facing side. It’s been frustrating, but it’s also been a catalyst for me in terms of innovation possibilities.

Kandamangalam: [The implementation of healthcare] technology assets used to be slow. After COVID, it started ramping up. Now, everybody talks about artificial intelligence (AI). It is increasing the speed of [healthcare] development and deployment. I think that in the next 10 years, things are going to be completely different.

Why was the industry lagging in implementing technology prior to COVID?

Tannenbaum: We are under a lot of constraints when adopting [patient] technologies. One issue is: Who’s going to pay for it? You can have the greatest technology, but if insurance companies aren’t going to pay for the treatment, nobody is going to adopt it.

Then there are liability issues. If we are using a new technology and it goes sour, who’s liable for that – the hospital, the doctor using it, the manufacturer? Then there is validation, and this is especially true with the AI. Does [the AI outcome] make sense? That has to be validated.

Anderson: The integrated health system is another perspective. Because most integrated health systems are basically compilations of different hospitals with different technology solutions, there was a time factor involved in [creating] a unified platform, to have singular data or singular approaches to the work.

How has telemedicine increased in usage since the COVID pandemic?

Kapoor: We established our Digital Transformation Office at Virtua right after the second wave of COVID, during a time when we experienced a decade’s worth of tech change in just a week. I told the board, “Telemedicine is here to stay.” COVID was the catalyst. But data now shows telemedicine didn’t maintain its momentum – utilization dropped from 80% during the peak to just 10%–11% today. This taught us not to confuse forced adoption with sustained, voluntary use. We’re seeing a similar pattern with AI. After the hype, we’re likely entering the “trough of disillusionment.”

Anderson: AI is augmenting the clinician. We (RWJBarnabas) use Epic’s [AI-based] Deterioration Index, which takes 20 to 30 variables and predicts whether someone might die. It doesn’t mean it’s always right, but it’s helpful in pulling data together and prompting the physician to take a closer look. It’s important to have [the two] working together.

Dadlani: Some tasks are better done by AI due to the human mind’s data limitations. It really depends on the use case. In clinical settings, organizations have low risk tolerance, rightly so, as lives and patient safety are involved. But in non-clinical areas – like patient engagement or revenue cycle management – there’s more flexibility [with AI]. Financial errors like billing or coding mistakes, while important, aren’t life-threatening. Some tasks will use a hybrid model: AI will handle the initial workflow, and human experts will do the final validations and reviews.

Kandamangalam: I am very upbeat about AI. It is going to stay and change the way we care for patients. In the next 5 to 10 years, it’s probably going to be more in the assisted form, rather than autonomous. But most of the AI use now is not at the point of care. That’s going to change with respect to AI technology improving quickly.

Johnson: There’s been a major breakthrough in AI, and that is the ability for it to interpret and read images, just like it could read text and numerics now. That, combined with the development in processing speed, is the game changer for our industry. You might have been surprised when a lot of us said we were behind [in technology implementation]. That’s not going to be the case. I think we’re actually going to excel with innovation.

How are medical schools integrating AI today?

Dadlani: There is a huge gap between the pace at which the technology is moving [at healthcare facilities], versus the pace at which it is becoming part of the medical school curricula, or even part of regulations. For example, how do you put regulations and curriculum in place when the technology changes by days and minutes? That’s a challenge.

How do you balance implementing something that’s cutting edge while waiting for regulatory approvals?

Kapoor: It’s going to be nearly impossible for health systems to wait for federal guidance, because the Feds are the Feds. We cannot wait for them or the states to come up with regulations [on healthcare technologies]. … If we are not, as health systems, coming up with our own policies or collaborating together on policies, we can get ourselves in a lot of trouble by falling behind and not pushing ourselves forward.

Kandamangalam: With respect to AI currently used in our health systems, most of the policy and procedures already in place can be adopted for it. So, we don’t need to create special policies, except here and there, based on the use cases.

So, hospitals could implement AI technologies, and not encounter severe regulatory issues?

Kapoor: The point is that we have a responsibility. If we feel that there’s a tool that can actually improve outcomes for people, don’t we have a responsibility to take on some of that risk? … That’s why we all went into this field, right? To help reduce disease, burden, and pain.

If you don’t have that FDA approval, how do you deal with liability issues?

Anderson: There’s a lot of uncertainty there. The value, though, is speed to market and innovation. The risk is safety, because it hasn’t been scrutinized as thoroughly as it might have been otherwise. So, we’re going to probably see situations where things get to market, and then we find bad outcomes. We have to go back and stop doing things that we started doing.

Kapoor: We might also hear the opposite criticism: “Why didn’t you adopt this sooner? If a mistake was made, the machine could’ve been a backup – why didn’t you use it?”

Do all of you trust the information that AI returns to you?

Johnson: Not inherently. I think AI has to be validated. And that’s time consuming if you do it right.

Dadlani: The term we use is zero trust. You have to consistently validate, refine, and improve AI.

Tannenbaum: If AI is your source and your source is old, you have to keep feeding it and basically recalibrating your AI with new information.

Anderson: Part of the flaw with AI is that it wants to give you an answer. There may not be an answer, but AI will give you one. They call it “hallucinating.” That’s part of the complexity in how to fine tune AI.

Kandamangalam: AI, as of now, is a secondary player, but the primary players (the physicians) are using AI, which is feeding information back to the them. For example, it is feeding the physicians, and caregivers with information saying, “This is what I am thinking as a treatment plan.” Still, physicians need to verify that in order to agree with the plan. As more and more of this fine tuning happens, as AI continues to learn, it will be a different situation three years from now.

To access more business news, visit NJB News Now.