Everybody loves to get and play with new toys. When we were children the toys were things like Barbies or Play-Doh. Now the toys we’re excited to get and play with are the emerging technologies in healthcare like AI or IoT devices. The big difference though is we can’t take these toys out and play with them with abandon like we did with the Play-Doh. While new and exciting, there are serious consequences to think about and regulations that need to be put in place and that will be put in place.
To talk more about the regulatory considerations for implementing these emerging technologies, we reached out to our incredible Healthcare IT Today Community for their insights. This is what they had to say on this topic.
Matt Cohen, Director of AI at Loyal
As the buzz continues around emerging technologies in healthcare IT, like generative AI, it’s paramount for health systems to practice a level of caution when adopting these tools. While these shiny and new technologies have the opportunity to become useful assets, there’s still a lot of room for data error and misuse, and in the world of healthcare, that makes it a great challenge for now. Patient trust is paramount, and ensuring personal information remains protected while in the hands of providers is key to avoiding misdiagnosis, among other possible negative outcomes. While I believe utilizing intelligent software to aid in improving the healthcare experience is key, a proper balance must be implemented between safety and utility when adopting these new technologies.
Dr. Adrienne Boissy, Chief Medical Officer at Qualtrics
As artificial intelligence and machine learning move from buzzwords to people-facing applications in healthcare, human oversight is imperative. What keeps me up at night are the potential risks of these advanced technology tools – deception, security, monetization, and amplification of bias and discrimination. Technology – like humans – has both the power to hurt and to heal. Algorithms based on biased data can lead to inaccurate diagnoses, predictions, and actions that can further marginalize vulnerable populations, fuel racism, and drive worse health and financial outcomes. That said, healthcare stands at a critical inflection point about how to best care for caregivers and leverage new tools so that they can continue the extraordinary work of caring for others.
In addition, the regulations keep coming, which create more work for teams, and we hope also drive the quadruple aim of healthcare. To that end, regulations that lack empirical data to demonstrate impact should be revisited and uniformity about foundational standards and measurement in medicine should be clearly defined. This will dramatically impact the gaming of scores to get rankings, the sheer numbers of raters and rankers, and create alignment across healthcare as to the value that matters most.
As such, as a regulator for a day, I’d implement critical clarification and updates of HIPAA, its scope and classification when it comes to medical guidance from AI-powered tools, outline the responsibility and liability of AI, and explore the selling of data for profit without transparency and consent. I would seek to accelerate solutions that amplify the collective good through thoughtful design and demonstrate a significant impact on the well-being, burnout, equity, and health of employees and patients globally.
I’d also issue guidelines from a multidisciplinary expert panel, which includes patients, ethicists, and clinicians, about expectations for AI governance. AI and ML applications in healthcare hold incredible possibilities, but how we get there matters just as much, including adherence to shared ethical and moral principles of putting human beings – and humanity – at the center of all we do.
Mike Pattwell, Principal Solutions Consultant at Edifecs
Regulators are increasingly emphasizing innovation in healthcare IT and this fact is opening the door for nontraditional players to enter the healthcare space. Barriers for traditional payers to innovate with emerging technologies will signal the need to partner with tech-savvy vendors to stay relevant in their space.
Dave Bennett, CEO at pCare
Emerging technologies are implemented, used, and seen every day, but AI is one technology in healthcare that we really need to back up, assess, and strategically consider how we’re going to use it. We must look at it very deeply from the patient and end-user perspective to ensure we’re appropriately and responsibly implementing it. I believe this is a great consideration within the industry to best benefit everyone.
AI holds a great deal of promise. It’s amazing to see what we can do with it. Even so, stepping back and keeping it simple and user-friendly is crucial. We need to ensure people understand how it’s going to work, what the end result is, and keep a purposeful focus on the ways the technology can be optimized.
I think there is a possibility on two fronts. First, safety. I can foresee some application of guardrails to ensure the implementation of emerging technologies ‘does no harm.’ Second, competition. This may be longer coming but considering all the M&A activity coming under scrutiny for creating unfair market advantages, it is foreseeable that if new technology creates an unfair advantage regulators could step in.
Calum Yacoubian, Director of Healthcare Strategy at Linguamatics, an IQVIA company
The last ten years have posed new questions and challenges for regulators at a faster pace than ever before, as AI has become more widely adopted in healthcare settings. However – nothing has accelerated the need for regulatory change as quickly as the advent and roll out of generative AI and large language models (LLMs). With the democratization of AI that LLMs offer, there has been an explosion of potential applications in the healthcare setting – but for now, this remains a largely grey area, with limited specialist regulation in place. I foresee significant updates from the regulators in terms of how these technologies can be used, and the measures that will need to be in place for them to be applied in clinical settings.
Maxim Abramsky, AVP, Product Management at Edifecs
Regulatory bodies very rarely enforce specific technology, but in order to provide compliance in the fastest, most cost-optimal, and efficient manner, plans are looking at various technologies available and emerging: AI/NLP/ML, and robotic process automation (RPA).
Eric Prugh, Chief Product Officer at Authenticx
There’s a general concern about AI because of how quickly its popularity has accelerated. Many of the currently trending AI models are trained on billions of parameters. Nobody really knows the origin of those datasets, so people are rightfully worried. There’s not been a ton of transparency from big AI providers on what information or data has even been used to train the models, and we are just getting to a place where feedback and improvements to retrain models in tools like ChatGPT can be selective and can honor specific security and privacy requirements.
We will need to closely examine what AI regulation might look like — and who will be responsible for creating and overseeing the regulatory processes. We can hypothesize about the potential issues resulting from healthcare organizations adopting and implementing AI into their processes, especially when it comes to privacy. As stakeholders within the healthcare ecosystem, we must be very mindful of how we use AI to help organizations gather the data and insights they need without mismanaging patients’ personal health information or sacrificing their privacy rights.
The key factors to consider as we look at regulation for AI—specifically for healthcare—are how AI companies maintain a responsible and ethical lens on recommendations and decisions, how they limit biases, the scope in which AI can be used, and how they control data being used to train and train models.
Laxmi Patel, Chief Strategy Officer at Savista
As AI and machine learning technologies advance, there may be a need for regulations to ensure their ethical and responsible use in healthcare. Regulations could focus on transparency, bias mitigation, validation of algorithms, and patient safety. Additionally, improving data interoperability and the seamless exchange of health information across different healthcare systems and providers has been a focus of policymakers. New regulations could be implemented to encourage the adoption of standardized data formats and APIs, making it easier for various healthcare IT systems (including EMRs and practice management systems) to communicate and share patient data securely.
The implementation of AI in healthcare requires careful attention to several regulatory considerations around ethical issues including:
- Safety and Transparency
- Prioritizing the safety and transparency within AI systems is paramount
- Implementing measures to validate and explain the decision-making processes of AI algorithms helps build trust among healthcare professionals and patients
- Algorithmic Fairness and Biases
- Vigilance against algorithmic biases is vital to ensure fair and unbiased AI-driven healthcare practices
- Continuous monitoring and assessment of algorithms can help identify and mitigate potential biases that might impact patient outcomes and healthcare equity
- Data Privacy
- Protecting patient data is of utmost importance
- Adherence to strict data privacy regulations and security protocols safeguards sensitive information from unauthorized access or breaches
Brian Fugere, Chief Product Officer at symplr
The potential for artificial intelligence and IoT devices to improve the efficiency of workloads, add scale in healthcare, and improve patient care is exponential. However, as the potential benefits of these expand, so do mounting concerns regarding regulations. As AI enters the healthcare industry, a top priority remains patient safety and privacy, which would include updating HIPAA (Health Insurance Portability and Accountability Act) guidelines to protect from potential data breaches and to ensure technology is safeguarding sensitive patient information.
Another significant concern with emerging AI technologies is data bias, with research showing that AI algorithms can be influenced by data bias, holding the potential to exacerbate healthcare disparities and inaccurately diagnose patients. Regulatory considerations for AI algorithms must include a diversified sample set of data, which is representative of the population as a whole. As technologies develop and transform the possibilities of healthcare operations, quality reviews, outcomes assessments, and comprehensive security measures are the first line of defense to maintain patients’ trust in healthcare organizations, and regulations to keep up with these developments to protect patients and guarantee safe and effective outcomes.
Brian Hanley, VP of Public Sector at Edifecs
Patient consent is going to be increasingly important – ensuring that patients have the ability to choose who to share their data with, for how long and for what purposes; and that the technology exists in order to parse their data and share according to the penitent’s specifications.
Colin Banas, M.D., M.H.A., Chief Medical Officer at DrFirst
With regulations, there’s always a tension between establishing guardrails and suppressing innovation. With the tremendous potential for AI to serve as a much-needed “automation co-pilot” in clinical care, it’s crucial that regulations allow for creative approaches and don’t put developers at risk of revealing their “secret sauce.” The ONC’s proposed rule for “Health Data, Technology, and Interoperability,” known as HTI-1, breaks new ground by recognizing advanced clinical decision support tools and differentiates simple from complex decision support. This is an important step forward, as clinical-grade AI has made meaningful impacts by supporting and augmenting clinicians’ decision-making while reducing repetitive and administrative tasks that can wear down healthcare providers and limit their time with patients.
Eli Ben-Joseph, Co-Founder & CEO at Regard
While it’s great to see the healthcare industry experimenting with generative AI technologies to improve healthcare delivery, we are still in the early phases of this new era. I believe this path must be taken with caution, underpinned by robust regulatory guidelines to ensure safe and effective use. This process could be akin to the stages of drug approval, involving rigorous testing, clinical trials, and post-market surveillance to monitor performance and address any unforeseen issues. Although some major companies oppose stricter oversight, I believe these regulations are essential for a safe future of AI/LLM use in our healthcare system.
Elaina McMillan, Director of Product Marketing at Edifecs
Mandates require technology infrastructure that builds on each other. For example, the CMS-0057-P proposed rule requires the use of FHIR APIs and essentially builds upon previous regulations. By having an agile, secure, and scalable infrastructure in place, like an interoperability platform, Cloud and SaaS infrastructure, and even AI to help speed up and automate processes, industry organizations anticipate and respond to mandates more quickly while also building business value. It’s incredibly important that healthcare organizations think of their mandate-driven investments as a way to build long-term value for their business rather than just a “check-the-box” as regulations will continue to compound and build on each other.
So many good ideas here! Thank you to everyone who took the time out of their day to share a quote with us and to all of you for taking the time to read this article! We couldn’t do this without your support. We’d love to hear from you as well! Leave a comment down below or share this on social media and tell us what regulatory concerns you think there need to be for implementing emerging technologies in healthcare IT.
