Integrating AI Insights into Healthcare Workflows: Achieving Effective Adoption

By Rosemary Ventura, Chief Nursing Informatics Officer, University of Rochester Medical Center

Artificial Intelligence (AI) has transformative potential for healthcare, offering significant improvements in diagnostic accuracy, reduction of the cognitive burden, and enhancement of operational efficiency. However, realizing these potentials hinges on the careful integration of AI insights into established healthcare workflows, which demands meticulous planning and strategic execution. In this article, we will explore key informatics and change management strategies to effectively enable the incorporation of AI into healthcare systems, ideally tackling historical pain points electronic medical record (EMR) systems have struggled to address. 

Many of us who experienced the rollout of healthcare information technology (HIT) projects have gathered a wealth of valuable lessons. Before diving headfirst into AI, it’s crucial to evaluate what worked and what didn’t in those earlier initiatives. The core values of clinical informatics, fostered through those initiatives, can serve as a framework for an effective implementation strategy. Partnered with change management methods, organizations can leverage the following strategies that prioritize thoughtful planning and alignment with clinical needs to successfully implement AI tools.

Leveraging clinical informatics and change management strategies is essential for organizations embarking on their journey toward AI adoption in healthcare workflows.

1. Choose Wisely: There are endless possibilities for the application of AI, making it essential to select a clinical or operational workflow that resonates meaningfully with your organization and its stakeholders. Focus on areas such as enhancing quality outcomes, streamlining billing and regulatory requirements, promoting clinician wellness, and alleviating documentation burden—each of which presents opportunities for impactful AI solutions. Approach this decision with a data-driven mindset, beginning with small, manageable projects that target “low-hanging fruit.” By prioritizing these areas, you can achieve quick wins that build momentum and demonstrate the value of AI, paving the way for broader implementation in the future.
   
2. Nothing to Me Without Me: Incorporating foundational principles, such as engaging clinical informaticists and front-line clinicians early in the development process, is key to adoption. This collaborative approach not only ensures that the tools meet real-world needs, but also fosters a sense of ownership and trust among users. AI introduces new concepts and emotions—such as anxiety about job displacement or skepticism about decision-making transparency—which can impact the workforce’s reception of these tools. By involving clinicians from the onset, leadership can address concerns, gather valuable insights, and build solutions that align with clinical workflows. Emphasizing a “Nothing to Me Without Me” philosophy not only enhances usability but also empowers clinicians, creating a culture of collaboration and innovation that is vital for the future of AI. 
   
3. Incorporate Change Management: For many, the use of AI is a novel experience, and their comfort level can significantly influence the success of implementation. While departments like radiology have embraced AI for some time, other clinical areas may be more accustomed to traditional clinical decision support (CDS) systems that still require manual intervention. The shift towards AI that automates tasks—such as placing orders or making decisions without direct human intervention—can create uncertainty and resistance among clinicians who are used to being closely involved in every step of the decision-making process. Conducting a thorough change management analysis helps assess the willingness of front-line clinicians and staff to adopt new tools. By acknowledging the differences in familiarity with AI and fostering a supportive environment, healthcare organizations can enhance acceptance and ultimately improve the quality of care delivered through this innovative technology.
   
4. Include Core Concepts: Communicate, Educate and Train, and Support. Given our HIT evolution, the level of technical sophistication of our clinical workforce and their lived experiences, it may be easy to underestimate and overlook the importance of these foundational concepts when introducing AI to healthcare workflows.
   
   • Communicate- Start with transparency. It is important to share the “why” and how the project was selected, including project data and goals. Next, reinforce the “what’s in it for me.” Maintain open lines of communication to explain how AI will integrate with existing workflows, impact day-to-day responsibilities, eliminate burden, and drive improvements in patient care rather than complicate it.
     
   • Educate and Train- An effective education strategy is imperative to ensure all staff have a fundamental understanding of how AI tools function and their intended applications within the organization. This fundamental knowledge helps demystify AI and empowers clinicians to leverage them confidently in daily workflows. When it comes to training, as leaders and informaticists, we envision AI solutions that are not only intuitive but also serve as a much-needed alleviation of workflow burdens. The goal is for these tools to be user-friendly, enabling clinicians to seamlessly integrate them into their daily routines without extensive training demands.
     
   • Support- Acknowledge the novelty of AI and establish support systems, such as access to clinical informaticists or IT specialists, to assist staff during the transition period. While these go-live techniques are often utilized for large-scale implementations, consider these new ventures into AI as experiences that may uncover unique day-to-day workflow challenges, or unexpected behaviors not seen with other technologies.
     
5. Monitor, Evaluate & Share AI Performance: The integration of AI into healthcare workflows is not a one-time event, but requires continual monitoring and evaluation. This includes iterations to the technology and workflows, specifically tracking metrics such as technical accuracy, clinician satisfaction, and meeting desired outcomes. Implementing mechanisms for clinical feedback, including leveraging informatics governance groups, for workflow modifications, will ensure a holistic approach to adoption. Technically, ongoing monitoring and reporting of AI tools’ performance in the form of accuracy and validity to key stakeholders, will build trust and sustained commitment to the tools. 

Leveraging clinical informatics and change management strategies is essential for organizations embarking on their journey toward AI adoption in healthcare workflows. By focusing on core concepts such as understanding the desire for change, providing education about AI, and ensuring tangible workflow gains, organizations can set the stage for a successful integration.